http://www.stanford.edu/~rajatr/jmlr_index.html JMLR Vols 1-8 http://www.jmlr.org/papers/volume2/cristianini01a/cristianini01a.pdf 321 Latent Dirichlet Allocation abramowitz abuse adding advances aitr aking aldous algorithm along alternating american analysis annotated annual appendix applications apply approach appropriate approximate arbitrary arguments arises artificial ascent aspect association assumption attias avery baeza based bayes bayesian begin below berkeley berlin between bibliography blei bound bounding bounds caenorrhabditis cambridge carlin categorical censored center cgcbib chapman chichester class classification clustering cognitive collaborative comparison complexity component computation compute computed computer computing conditional conditionally conference consider considered constrained contain containing content converges coordinate corpus cubic data deerwester defined denote denotes dependence depends derivative derivatives derive deriving describe described diaconis diag diagonal dickey difference digamma dirichlet discussed discussion distribution distributions divergence division document documents dover dropped dumais each earlier easily ecole editor editors eighteenth elegans elements empirical entropy environments equal equation equivalent estimate estimates estimating estimation exactly exchangeability exchangeable expand expands expansions expectation expected experimental exponential exponentiating expression fact factor factorizations family filtering final finally find finding finds finetti first five fixed flour following form found framework from full function functions furnas furthermore gamma gelman general generated generative genetic ghahramani given gives gradient graphical griffiths hall hand handbook harman harshman have heckerman here hessian hierarchical higher hill hofmann http hypergeometric ijcai immediately improving independent indexing individual inequality inference information initialization intelligence international interpretations intractable introduction inversion invoke involves isolate isolating iterating iterative jaakkola jelinek jensen jiang joachims john jordan journal kadane kappen kass kernel labeled lafferty lagrange lagrangian landauer large latent lawrence learning left leisink lemma letting likelihood likelihoods linear lines london lower machine made making marginal mathematical matrix maximal maximization maximize maximizing maximum mccallum mcgill mcgraw meila method methods minimizing minka mitchell model modeling models modern moreover morris multi multinomial multiple multipliers multiplying must naive natural need neto neural newton newtonraphson next nigam normalization notation note notions observe obtain obtaining omitting only optimization order over overall overview oxford pages papadimitriou parameter parameterization parameters parametric particular pennock point popescul posterior practical presented press probabilistic probabilites probability problem procedure procedures proceedings processing progress propagation property provides raghavan raphson recall recent recognition recommendation references related remainder rennie repeatedly report representation reprint requires respect respectively restored retained retrieval ribeiro right ronning rubin saint salton same saul scalar scale scales science second section sections select semantic setting seventeenth several show shows side sigir simplicity simplifies simulation since single size smoothed society solve sons sparse special specific speech springer stationary statistcal statistic statistical statistics steffey stegun step stern steyvers structure subscript such sufficient summing support surrogate swmed systems take taking tamaki technical technique terms text that then theory therefore this those thrun thus time topic topics total translation trec true twenty uncertainty under ungar unified unique univ unlabeled until used useful uses using usually valencia value values variational vector vempala verified viewed where which wiley with word workshop write written xiii yates yields york zero http://www.jmlr.org/papers/volume4/herbrich03a/herbrich03a.pdf 255 One-Class SVMs for Do cument Classification aaai ablex advances agents algorithm algorithms analysis approach artificial automated back balabanovic berkeley billsus browsing butterworths california canada carnegie categorization characteristic cikm classification cognitive collection compression computer concept conference cottrell datta department detection development dimensional distributed distribution document dokimente dortmund dumais ecml edition editor editors environments estimating european examination example examples exclusively experiments extensional features filter fourteenth fredric from gathering gluck haifa hearst heckerman heterogeneous high http identification identifying image inductive information intel interesting international irvine japkowicz joachims know lake lang learning ledge lewis ligence london machine machines management manevitz many marti mellon methods microsoft montreal munro muramatsu myers networks neural news newsweeder notes novelty pages papers pazzani pittsburgh platt positive press probabilistic proceeding proceedings profiles programming propagation references relevant report representation representations research retrieval reuters revising richard rijsbergen rocchio sahami scholkopf school science second series seventh sharkey shawe shoham sigir sites smola spring springer support symposium syskill tahoe taylor technical test text thesis tong trained twelfth university user vector verlag volume webert williamson with working yang yiming york yousef zipser http://www.jmlr.org/papers/volume6/goldsmith05a/goldsmith05a.pdf 94 New Horn Revision Algorithms about acyclic adrian advances alberto alexander algorithmic algorithms allen altos amsterdam analysis analytical angluin annual april arimura artificial ashok attribute attributes auer august automated avrim balazs based bases belief bias blair blum boolean bound bradley bshouty chandra chapter clause clauses colt combining complexity compression comput computation computational computer computing concept conference conjunctions counterexamples dana data databases david december declarative deductive department dept derivations dirk domain doshi driven earlier editor efficient elicitation empirical entailment evelina evolui exact extracting fabrizio failure feldman finitely first formation formulae formulas foundations frazier from function gelder general generalizations generation geoffrey goldsmith good greiner gyorgy hammer harel hellerstein hints hiroki horn howard ieee infinitely inform information intelligence international irrelevant italy jackson japan jeffrey jignesh journal jude judy kaufmann kentucky kernel kluwer knowl knowledge kogan koppel krzysztof lamarckian lamma learning lecture leonard line lisa littlestone logic logical long machine machines many maria martin master methods michael minker mistake models moniz monotone mooney morgan moshe nader natural negation networks neural nick notes november october once optimal order ourston padova pages parity pereira peter petsche philip pitt preference preliminary presence press proceedings programming programs projective propositional quasi queries query raedt raymond read references refined refinement regular report research results revising revision richards riguzzi robert ronen rules russell sandholm science segre selecting sendai sentences sets shavlik sloan springer stefan structural syst systems szorenyi technical their theories theory thesis thomas tight tion towards towell trans tuomas turan umesh university using version volume walker washington with without workshop wrobel yale yaleu zinkevich http://www.jmlr.org/papers/volume7/barber06a/barber06a.pdf 192 Expectation Correction for Smoothed Inference in Switching Linear Dynamical Systems about acoustics advances aeronautics aerospace aiaaj aiii algorithms alspach american analysis andrews annals applications approach approximate approximations artech artificial aspects association astronautics attias audio automatic barber bayesian belief biometrika biswas blackwellised boyen carlo carter cemgil chain chib complex computation computational computer computing conditional conditionally conference control controls darwiche defense deng dept detection diagnosis different distributions dooren doucet dueker dynamic dynamical econometric econometrics editors endogenous estimates estimation expectation extension extensions family fault fieguth filter filtering finite forecasting formula freeman freitas friedman gaussian generative ghahramani gordon graphical grewal hall harrison heskes hidden hinton house human hybrid icassp idiap ieee ihler implementation implementations inference information institute intelligence international jaakkola jensen jordan journal julier kailath kalman kappen kaufmann kitagawa kohn koller language lauritzen learning lerner likelihood linear local logistic maccormick makov markov markovian mathematics maximum means media meetings mesot methods minka mixed mixture model models monitoring monte morgan motion multimodal murphy music national neal nelson networks neural nijmegen nips noise nonlinear nonparametric north norwood numerical orlando oxford pages park parr particle pattern pavlovic practice prentice press principles probabilistic probabilities problems proc proceedings processes processing propagation rabiner radboud rauch reasoning recognition references regime regression rehg review robust rong root roweis russell science selected sensing sequential series seventeenth shalom shumway signal simul smith smoother smoothing society software sorenson space speech springer square stable stanford state states statistical statistics stochastic stoffer strengths striebel sudderth summer switching symp systems techniques their theory thesis time titterington tong toronto tracking tractable transactions transcription tung tutorial uhlmann uncertainty unifying univ university using variances variational varying verhaegen vision volume west wiley willsky with xiao zoeter http://www.jmlr.org/papers/volume8/elbaz07a/elbaz07a.pdf 213 Separating Models of Learning from Correlated and Uncorrelated Data aldous algorithm algorithms analysis annual association bartlett basic beyond blum bshouty cambridge chains characterizing circuits colt communications computation computational computer computing conference constant construct countable cryptographic cryptography depth distribution donnell efficient elbaz exploiting extension fifth finite first fischer foundations fourier framework from function functions furst gamarnik generator goldreich goldwasser hardness hastad hoffgen ieee impagliazzo information jackson journal kearns kharitonov klivans learnability learnable learning levin linial luby machinery mansour markov markovian membership micali model models monotone more mossel nisan open pages parities polynomial press problem proc proceedings pseudorandom quasi queries query random rectangles references relevant respect roch rudich science sciences servedio siam specific spectrum statistical structures symposium system tamon theory third thirty thresholds time tools transform twelfth twenty twentysixth under uniform unions university using valiant variables vazirani verbeurgt volume walk walks weakly with workshop york http://www.jmlr.org/papers/volume6/ong05a/ong05a.pdf 66 Learning the Kernel with Hyperkernels adaboost advances akaho albert algorithm algorithmic algorithms alignment american anal analysis annals another applic applied approximation aronszajn asai ashrae asymptotic auxiliary bach barber bartlett bayesian becker bennett between bishop boosting bounds bousquet boyd brailovsky burges cambridge cbms chapelle characters choosing christopher clarendon classification classifiers competition completion complexity component computation computer conditions cones conf conference connection control convolutional cortes crammer cristianini cruz data david davis definite definiteness department design dietterich dimensional discovery discrete discrimination distribution division duan edinburgh editors efficient elisseeff energy estimating estimation estimators evaluation experiments field fine francisco freund functions gaussian ghahramani ghaoui graphical hanl hasselmo haussler herbrich herrmann high hornik html http hyperkernels hyperparameters ieee independent informatics information inseparable intelligence international jmlr jordan journal kandola kaufmann keerthi kernel kernels keshet kibernetica kimeldorf knowledge lagrangian lanckriet large lauritzen learing learning leave leisch likelihood linear linearly lofberg luckiness luntz machine machines mackay mangasarian margin margins master math mathematical mathematics matlab matrix mean measures methods meyer mining modelling models morgan mozer mukherjee muller multiple musicant nature neal networks neural neurocomputing nonnegative observational obtained onoda operators opper optimization optimizing over oxford pages pami parameters parser pattern penalized performance philadelphia platt positive prediction press problems proc procedure proceedings processes processing programming pseudoinverses publishers rank rasmussen ratsch recognition references regional regression regularization related report representations reproducing research results review robust russian santa schapire scheinberg scholkopf schuurmans science sedumi seeger semi semidefinite series sets shawe siam simple singer smola society soft software some sparse spline springer statistical statistics structures sturm sullivan support symmetric systems target taylor tchebycheffian technical technicheskaya terms test theory thesis toolbox touretzky training transactions transcations tsuda tuning tutorial ucsc under university using vandenberghe vapnik vector volume wahba williams williamson winther with yalmip york zhang http://www.jmlr.org/papers/volume2/gentile01a/gentile01a.pdf 330 Extensions to Metric-Based Model Selection aaai accuracy adaptive advances algorithm algorithms annals approach artificial autocorrelation based bengio berlin boosting bottou business campbell chapelle combination comparing complexity conference consistent covariance criterion data definite density dependences diebold dietterich econometrica econometrics economic economics editors empirical equivalent estimation experiments financial foster freund george girosi heteroscedasticity inflation information intelligence international issue journal learning lecture leen machine mackinlay mariano markets mathematical matrix methods metric minimization model modeling multilayer multiple multivariate national networks neural newey notes pages poggio positive practice predictive press princeton proceedings processing references regression regularization risk rissanen sample schapire schuurmans science scott selection semi simple small southey special springer statistical statistics stochastic systems that theory thirteenth tresp university vapnik verlag vicinal visualization volume west weston wiley with york http://www.jmlr.org/papers/volume3/globerson03a/globerson03a.pdf 280 Shallow Parsing using Noisy and Non-Stationary Training Material adwait alamitos american angluin anlp annotated annual antal applied article aslam association available avoidance backpropagation bagging banko base based bergen berkeley bias boitet boosting bosch bounds brants brill buchholz building california cardie chapter christian chunk chunking citeseer claire classification classifier combination combining computational computer conference conll corpora corpus cullen daelemans dana daniel david decatur dietterich disambiguation disfluencies driven editors elizabeth empirical english ensemble entropy eric erik error examples exceptions first forgetting foundations francisco free from general generalization gildea grammars guide harmful henderson html http hypothesis identificatio identification ieee improved international introduction jakub javed john jorn july june kaufmann kittler kudoh laird lance language large lcgwww learner learning lecture lexical linguistics lisbon lunch machine macready madrid marcinkiewicz marcus mardis matsumoto maximum meeting memory methods michele miles mitchell morgan multiple naacl natural noise noisy north norway notes noun osborne overfitting pages parser parsing part partial penn pennsylvania performance pete philip phrase pierce pittsburgh preliminaries press proc proceedings processing pruning publishers query ramshaw ratnaparkhi reference references report roli rule sabine sang santorini scaling schaffer science scott search seattle seventeenth shallow shared shriberg similarity simple single sixth skut sloat smoothing society software speech springer statistical support syllabification symposium systems tagger tagging taku task technical technology text theorems theory thesis thirty tilburg timbl tjong toulouse transformation treebank trento twente university unrestricted upenn using variation vector veenstra verlag version very walter whitelock william with wolpert workshop york yuji zavrel http://www.jmlr.org/papers/volume7/kitzelmann06a/kitzelmann06a.pdf 118 Inductive Synthesis of Functional Programs: An Explanation Based Generalization Approach aaai abstraction achievements aips algorithmic analogical annual applying approach artificial autmatic autom automatic avoiding bamberg berardi berlin biermann caching camacho cambridge cameron cogsys collier computer computing conference construction control data definitions dept dershowitz design detection dialogs documents edinburgh editor editors effect efficient elsevier entailment examples feng finite first flener folding from functional functions fundamenta generalization germany gold guiho handbook http identification ijcai incremental induction inductive inference informaticae information intel international issue jones jouanaud journal kaufmann king kitzelmann kodratoff krishna language learnability learnable learning leeuwen ligence limit lisp local logic logical lowry machine macmillan macro malerba mark mass master mccarthy methodology methods mike morgan muggleton normal note ohmsha olsson optimal ordered pages partridge pitfalls planning plotkin polynomial positive press problem proc proceedings program programm programming programs prospects quinlan raedt reasoning recursive references relations report rewrite rewriting ross sattar scheduling schema schmid schmidhuber science setting sixteenth smith softw software solver special springer srinivasan stoc summers survery symposium synthesis systems technical techniques term theoretical theories theory thesis time tokyo transformation universal university using valiant variables varlaro volume when wiai with wysotzki years yilmaz york http://www.jmlr.org/papers/volume6/winn05a/winn05a.pdf 53 Variational Message Passing adaptive advances algorithm algorithms american annual applications applied approach approximate approximations artificial association attias bayesian beal belief berger bernardo best between bishop bugs cambridge clarendon components computer computing concepts conference cowell dawid dietterich distributions editor editors engine engineering european expectation expert experts extensibility extensions field florida framework fusion ghahramani gibbs gilks graphical hierarchical hinton http icann image incremental inference information intelligence international introduction jaakkola jordan journal junction justifies kauffmann kaufmann kjaerulff kluwer lauritzen learning leen linear logistic lunn mean means meek message methods minka mixed mixtures modelling models morgan muller neal networks neural nineteenth ninth october other oxford pages passing pearl perform press principal probabilistic probabilities problems proceedings processing program propagation references regression rejection sampling saul science sixth smith society solla spiegelhalter springer statistical statistics structure structured structuring svensen systems that their theory thesis thomas tree tresp uncertainty university using variances variants variational verlag vibes view vision volume west wiegerinck wild winbugs winn workshop http://www.jmlr.org/papers/volume3/crammer03b/crammer03b.pdf 289 Algorithmic Luckiness advances algorithm algorithmic algorithms alignment alon american analysis annual anthony application applications applied approximants approximate approximation artificial association bartlett based bayes bayesian becker berlin blumer boser boucheron bound bounded bounds bousquet brain brussels california cambridge campbell chervonenkis class classifier classifiers clustering code combinatorial compression computational computer computing concentration conditions conference consistency convergence cortes cristianini cruz data delsarte density dependences dependent devroye dietterich dimension discrete discriminant douglas editor editors ehrenfeucht elisseeff empirical erdos error esann estimation events facto families feder feller fisher floyd foundations frequencies functions generalisation generalization generalized ghahramani graepel greene guyon haussler herbrich hierarchies hoeffding ieee image inequalities inequality information introduction isbn john journal july kahane kandola kearns kernel kernels large larsen learnability learning leen linear littlestone lkopf luckiness lugosi machine machines macwilliams makovoz marchand margin massachusetts massart mathematics mceliece method methods mika minimization model more motion muller multi nature necessary networks neural optimal organization over pages pattern peled perceptron performance pittsburgh point press probabilistic probabilities probability proceedings processing psychological random rate ratsch recognition references relating relative relevance report representer research result review risk rodemich rosenblatt rumsey sample santa sauer scholkopf schuurmans science series sets sharp shawe signal smola some sons sparse spencer springer stability statistical storage structural structures sufficient sums support svms symposium systems target taylor technical than their theorem theory tipping training transactions tresp uniform university upper vapnik vapnikchervonenkis variables vazirani vector verleysen volume warmuth watkins welch weston wiley williamson wilson with work york http://www.jmlr.org/papers/volume7/niculescu06a/niculescu06a.pdf 153 Bayesian Network Learning with Parameter Constraints analysis annals annual applications applied artificial auai available bayesian belief berkeley bilinear bilmes bishop boutilier brain california cambridge carnegie carolina carpenter characterization clinical clustering cognitive comprehension computation computing conference constructive content context course cscald dale data decode dependent design dirichlet dirtree discovery distribution domain during dynamic eddy editor estimators event existing experimental exploiting february filter financial fmri fmriactivation freeman friedman from geiger germond getoor global goldszmidt graphical heckerman hidden hooper hospital http human hutchinson identify ieee ijcai images independence induction inference information intelligence international introduction jordan june just kalman keller knowledge koller kuhn lafferty language larranaga learning linear local lozano machine mapping markov mathematical means mellon methods microsoft mining minka mitchell models module multinets murphy networks neural neuroimage newman niculescu ninth nonlinear north numerical object online optimal oriented outcomes overlapping pages paper papers parameter parameters patient pena pereira pfeffer press priors probabilistic probability proceedings process processes programming rabiner recipes recognition records recursive references regev related relational report representation research retrieval rustandi sandilya scientific second segal selected sentence separating siam sigir sigkdd similarity smoothing spatial specific speech states statistics study style symposium technical tenenbaum teukolsky thesis this through thulborn time tree tucker tutorial uncertainty university unpublished vetterling wang welch with zhai http://www.jmlr.org/papers/volume8/landwehr07a/landwehr07a.pdf 221 Integrating Našve Bayes and F O I L i aaai accuracy activity advances against aggrega aggregation alexandrin algorithmic algorithms applied approach artificial ashvin ashwin assessment assisted attributes automatica banff based bayes bayesian better blair blockeel branham brazdil brian bristol bruce building burnside camacho canada carrie case castro celine chemical city classification classifier classifiers classii claudia clausal clustering combine comparing compounds computer computing condensed conditional conference conjunction considerations constraints costa craven daniel daphne data databases david davis definitions dehaspe description dial dietrich discovery discretization distribution diterpene diverse document domingos douglas dutra dzeroski ecml editors efficient eibe eleventh elizabeth ellis elucidation entailment entropy environmental eran estimation estimators estrogens european examination experiment explorations fang fawcett feature feng fiers fifteenth fifth finding first flach florida foil foster fourth frank frequent friedman from gallagher gama generation getoor goldszmidt graphs grossman hannu hass heidtke hendrik hong horwood hypertext icdm icml identifier ieee igor ijcai implemenations induction inductive ines integrated integrating intelligence internatinal international invention inverse irene issue japan java jennifer jensen jesse joao joint jorge jorma journal karsten kaufmann kersting king knowledge knowlege kohavi koller kononenko kremer kristian krogel lachiche landwehr large lavra lavrac lawrence learning lecture leming likelihood lise logic logical lookahead lyle machine madison mark maruoka maximizing maximum melbourne methods michael michele mining modeling models moises moland morgan mrdm muggleton multi multirelational mutagenicity nada naive national natural network networks neville nfoil nicholas niels ninth notes order oregon page pages pavel pennock pennsylvania perkins perlich peter pfeffer pittsburgh pompe popescul portland practical predicate press principles probabilistic proceeding proceedings progol programming programs provost qian quinlan raedt ramon reasoning references relating relational relations relationships representation representations research researchers rissanen robert roger ross rouveirol rules santos saso schulze science sean seattle sebag segal seventeenth seventh sheehan shortest siems sigkdd simple sixteenth slattery society special spectra springer srinivasan stacy statistical stefan steffen stephen sternberg steve stochastic structure structured study substructures successes synthetic taskar techniques tenth theories theory third thirteenth tion toivonen tokyo tong tools torgo toxicology transformation twentieth twenty ungar uros using volume washington weida wettschereck whistler william wisconsin with within witten working workshop wrobel york http://www.jmlr.org/papers/volume7/abbeel06a/abbeel06a.pdf 167 Learning Factor Graphs in Polynomial Time and Sample Complexity abbeel algorithm algorithms analysis annals applications applied approach approximating approximation artificial asymptotic automata bach barahona based bayes bayesian belief bell berlin besag bilmes biometrika bounded carlo causation cheng chickering chow class classifiers clifford colt comets comparison complete complexity comput computational computer concepts conditional congress consistency consistent constrained control cover cowell cybernetics dasgupta data dawid decomposable della dependence dependent deshpande differential discrete discriminative distributions divergence edition editors efficiency efficient efficiently electric elements estimates estimation estimators expert exponential factor families features field fields finding finite fisher fixed fleming freeman frey friedman from garofalakis gaussian geman generalizations generalized generative geyer gibbsian gidas glass glymour graffigne graphical graphs greiner guyon hammersley hard harder heckerman hoffgen huang ieee image inducing information institute intelligence interaction international isbn ising jerrum john jordan journal junction karger kaufmann kelly koller kschischang kullback kunsch laboratories lafferty large lattice lattices lauritzen learnability learning lecture leibler lenz likelihood lions loeliger logistic machine malvestuto markov mathematicians mathematics maximum mccallum meek methods mitsubishi model models monte morgan naive narasimhan networks nips notes ogata optimal oxford pages path pattern penalized phys pietra polynomial polytrees prediction press probabilistic probability proc procedure product propagation pseudo pseudolikelihood publishers random references regression report reseach research respect robust royal sample scheines search second selection series seymour siam simple sinclair society sons spatial spiegelhalter spin spirtes springer springerverlag srebro statistical statistics stepwise stochastic structure symposium systems takeuchi technical than their theory thin thomas thompson time transactions tree trees understanding university unpublished vapnik vision warmuth weiss width wiley with yakhini yedidia york http://www.jmlr.org/papers/volume3/perkins02a/perkins02a.pdf 302 On Bo osting with Polynomially Bounded Distributions academic achieved adaboost adaptive advances agnostic algorithm algorithmic annual application automata avoid been boolean boost boosting bshouty communications complexity computation computational computer computing conference core cryptographic david decision determinism distribution domingo efficient extended feldman finite formulae foundations fourier freund functions gavinsky generalization hard implications improved information international jackson jority journal kearns klivans kluwe known kushilevitz learnability learnable learning levin limitations line logic long machine madaboost malicious mansour membership modification more most neural noise optimally proceedings publishers queries query randomness references respect result schapire science sciences servedio sets siam smooth spectrum statistical strength symbolic symposium system tamon theoretic theoretical theory transform trees under uniform using valiant version watanabe weak with http://www.jmlr.org/papers/volume7/porta06a/porta06a.pdf 186 Point-Based Value Iteration for Continuous POMDPs aberdeen acting active adapting adaptive advances alberta algorithm algorithms american amherst analysis analytic annual anytime appearance applications approach approximate artificial assistance assistants association astrom athena australia austria automation autonomous auxiliary bakker banff bartlett barto based baxter bayes bayesian belief bellman belmont bertsekas blake boger bounded boutilier british brooks building burgard cambridge carlo carnegie cassandra challenges cheng clustering columbia compact complexity compression compressions computational computer computing concurrent condensation conditional condon conference continuous control controlled controllers convergence cued darrell david decision dellaert dementia density department dialog dialogue directed discrete domains doucet duff durrant dynamic dynamical dynkin edinburgh edition edmonton elderly elicitation engineering environments estimation exact experiences exponential factored family fast fernie filter filtering filters finding finite formulation freitas from gaussian gesture goldberger goldsmith gordon gradient guide hanks heuristic hierarchical highdimensional hoey homes hong horizon hypotheses ieee incomplete incremental infeg infinite infinitehorizon information intelligence intelligent internal international introduction isard iteration japan john joint jordan journal kaelbling koenig kong krose kurien large learning linear linguistics littman localization long lusena machine madani mahadevan makarenko management markov massassachusetts mathematical mathematics matt mdps meeting mellon memory method methods meuleau mihailidis mixture mobile model models monahan monte montemerlo multi mundhenk national navigation neural neuro nips nonapproximability nursing observable observation observations operations optimal orlando pages papadimitriou parametric partially particle pattern pegasus pentland perseus persons peshkin pineau pitt planning point policies policy pollack pomdp pomdps poole porta portland poupart practice predictive preference press princenton probabilistic probability problems procedures proceedings processes processing programming propagation pruning puterman random randomized reasoning recognition references regan reinforcement report representations research results robot robotic robotics robots roweis rudary sample sampling satinder scalable scaling science scientific scotland search seattle sendai sensor sequences sequential series shephard short simmons simple simulation singh size smith solution solutions solving sondik sons spaan spaces speeding spoken springer stanford state statistics statististical steven stochastic stockholm survey sutton sydney systems task technical term terwijn theocharous theoretic theory thesis through thrun towards tracker tracking tsitsiklis uncertainty undecidability under university using value vdcbpi verlag verma vienna vision visual vlassis whitehead whyte wiley williams wingate with workshop york young zhang http://www.jmlr.org/papers/volume6/gretton05a/gretton05a.pdf 100 Kernel Methods for Measuring Independence abed acad academic achard acta adaptive advances aistats akaho algorithm algorithms amari american amin analysis annals applications applied approach arias artificial associated association astakhov asymptotically atoires augath bach baker bakir banach based behaviour belitski bell belouchrani berkeley berlin bernoulli bickel biological bischof blanchard blind block bousquet brabanter bracewell breiman cambridge canonical cardoso chapman chen cichocki cichoki classification clustering code colt comon component comprehensive computation computational computers concept conference consistency consistent constrained constraints context contrast convergence correlated correlation correspondence covariance cover cristianini cross curves cybernetics dans data dauxois deconvolution demos density department dependence dependent diagonalisation dimensionality duda edelman edition efficient eigenvalue elements engineering entropy essentials estimates estimating estimation existence fast fbach fine fisher fmri foundation fourier framework franz frequency friedman friman fukumizu functions fyfe garat gaussian gbauer generalisation generalizations geometry gestel girolami gradient grassberger greenacre gretton group groups hall hardoon harmeling hart haykin hein herbrich hilbert hill horn hosseni html http hungar hypothesis icann icassp ieee illustrate image imps independence independent index influence information ingster inst institute intelligence international into iscas jacod jmlr john johnson joint jordan journal jutten kandola karhunen kawanabe kcca kernel kernels kraskov kuss learned learning least letters leurgans likelihood lkopf ller logothetis lomi london machine machines macmillan master math mathematical mathematics matlab matrix maximization mcgraw measure measurement measures measuring medical meeting melzer ments meraim method methodological methods miller minimax mixture mixtures modified moor moulines mourier moyeed multidimensional multiple multivariate murayama music mutual natural nauchn networks neural nips nkiet nonlinear nonparametric norms note operators optimal optimization order orthogonality pages pajunen paninski papoulis partial pattern pauls pearlmutter pham phys planck plumbley poincar post press prewhitening principal principles prip probability problem proc proceedings processes processing properties protter psychometric quadratic quasimaximum random rank recognition reduction references regression reiter report representations reproducing research results rinen rosipal royal samarov samples scheinberg schmidt schoelkopf second sect sejnowski seminar sensitive separability separation series shawe siam signal signals silverman smith smola society sons source sources soviet space spaces spacings spectral springer squares statistical statistics steinwart stiefel stochastic stork structures submitted supervised support suykens systems taleb taylor technical technique temporally testing tests theis theory thesis thomas through time training transactions transform transformations translation tree trejo tsybakov uncertainty understanding unifying uniqueness unit university unpublished using vanderwalle variables vector verlag version volume when wien wiley with workshop york zhang ziehe zwald http://www.jmlr.org/papers/volume7/passerini06a/passerini06a.pdf 114 Kernels on Prolog Proof Trees: Statistical Learning in the ILP Setting aaai acquiring advances aggregate aggregation akutsu alberta algorithm algorithmic algorithms american amherst amini analysis annals annual applications applied approach approaches april artifical artificial association austria background banerji banff based bayesian becker bergadano between bias biermann biocomputing biodegradability blockeel bohnebeck bongard books bratko bruynooghe bulletin bundy cabelli california cambridge canada chambery classes classification classifiers collins combining comprehensible computational computations computing concept conference constructing continuous convolution correspondence cortes costa cristianini cruz cucker cumby cyclic czech data databases dealing debugging decision declarative decomposition defining description diego dimensional dipartimento discovery discrete discriminative distances down duffy dzeroski edinburgh edition editor editors effect eighth engineering entailment eskin estimation european evaluation example experimentation experiments explanation exploiting explorations extensions fast feature firenze first flach foil fold fortieth foundations framework france frasconi freund from frontiers functions gartner generalization generation generative gramming graph graphs greiner ground haffner hall harmelen haussler hawaii heuristics horvath ieee inducing induction inductive informatica information inokuchi instance intelligence international invention inverse investigations involving issue jaakkola joint journal karalic kashima kaufmann kedar keefe keller kernel kernels kersting kimeldorf king knobbe knowledge kompare kramer krishnaswamy labeled laer lakshman large lavrac learning leslie lihue linguistics lists lloyd lnai lncs lodhi logic logical logics machine mahe margin marginalized marseille massachusetts matching mathematical mathematics matwin meeting menchetti menlo mereotopology methods mining mitchell model models modern mohri mooney morgan morik muggleton multi multiplexer mutagenicity mycroft national nature nedellec networks neural newsletter nineteenth ninth noble norvig notices obermayer objects october order over pacific page pages park parsing part partial parts passerini pattern perceptron perlich perret pfahringer philadelphia pitmnann poggio practice prague predicting predictive prentice press principles problem proceedings processes processing progol program programming programs prolog proofbanks propositional propositionalization prospects protein provost quinlan raedt ramon ranking rational recognition reconstruction reddy references refining regression regularization relational relations report republic research roth russell sammut santa saraswat saunders schapire scholkopf schuurmans science scotland search second semantic series session seventh shapiro shawe siebes sigkdd signatures sistemi sixth smale smola smoothing society software solving spartan special spectrum speed splines springer springerverlag srinivasan statistical statistics sternberg stochastic string structural structured structures study support survey symposium system systems tagging taming tausend taylor technical technischen tenth terms text theories theory thesis third thirteenth three thrun torge towards transactions tree trees tsuda turcotte twelfth twentieth twenty type typed ucsc ueda unifying universita universitat university using utgoff vapnik varzi vector velde verlag vert view viswanathan volume voted wahba wang warmuth washington watkins weighted whole wholes wien with witten working wrobel york zelle http://www.jmlr.org/papers/volume6/cuturi05a/cuturi05a.pdf 70 Semigroup Kernels on Measures academic advances agnan akaho alain alexander algorithm algorithms amari amnon analysis andrew angles annals annual applicability applications artificial ation bach barndorff based basic bayesian becker bent berg berlinet bernhard between beyond biocomputing bischof boser bottou bousquet calcul cambridge canonical christensen christian christina christine class classific classification classifiers classify component computational conference context correlation covariance cuturi definite dietterich dieudonne differential diffusion dimensionality distributions divergence divergences dominik dordrecht editors eleazar embedding endres eskin extraction feature ferdinand finite flemming fourier francis from fuglede fukumizu generalized generative geometry ghahramani groups guyon harmonic hayward hein hermann hilbert hilbertian hiroshi horst howard icann ichi ieee igor imps independent inference infinitesimal information institute internat international isbn january jason jean jebara jens jensen joachims johannes john jordan journal kass kenji kernel kernels klaus kluwer kondor kullback lafferty lauritzen lawrence learning lebanon leibler leon leslie lior machine machines marco margin mathematical mathematics matthias measures meeting melzer method methods metric metrics michael mismatch models moreno multimedia nagaoka networks neural nielsen noble nonlinear nuno obermayer optimal optimization osterreicher over pacific page pages paris paul pedro peter philippe positive press principal probability proc proceedings processing product properties protein psychometric publishers purdy reduction references regularization reiter reproducing research ressel reus risi rudin saul schindelin scholkopf scientific sebastian seeger semigroup semigroups sets shannon shashua shotaro shtarkov shun smola society sons space spaces spectrum springer stafford statistical statistics string strings supervised support suzanna symposium systems text theory thomas thorsten thrun tjalkens tony tops training transactions tree using vajda vapnik vasconcelos vector vectors verlag vert walter weighting weiss weston wiley willems william with wolf workshop world yair http://www.jmlr.org/papers/volume8/mease07a/mease07a.pdf 219 Boosted Classification Trees and Class Probability/Quantile Estimation aaai adaboost adacost additive advances agarwal algorithm algorithms american annals applications artificial association asymptotic bankruptcy bayes bayesian bias bickel binary bioinformatics blanchard boosted boosters boosting bowyer bregman buhlmann building buja calibrated carbonell card case chan chawla class classes classification classifier classifiers classify cohen collins comparative comparison computational conference consistency contrary convergence correct cost credit data databases decision department detection dettling discovery discussion distances distributions does duffy early effective eighteenth elements eleventh elicitation elkan estabrooks estimates estimation european evaluating evidence exact expectations experiments exponential expression fast first foster foundations fourth fraud freund friedman from functions gene generalized hall hastie helmbold icdm ieee ijcai imbalanced improvements improving inductive information intelligence international japkowicz jiang joint joshi journal kegelmeyer knowledge kumar lafferty lazarevic learing learner learning lebanon likelihood logistic loss lugosi machine management margin maximum mease method methods mining minority misclassification model models multiple naive national neural northwestern obtaining overfit oversampling pages papers path personal potential practice prediction predictive principles probabilities probability proceedings process processing rare rate references regarding regression regularized report resampling research risk ritov rosset rule savage scalable schapire selection sensitive sets seventeenth shen simple singer smote smoteboost solution some springer statistical statistics stine stolfo stopping structure study stuetzle synthetic systems technical technique text theory thirteenth three tibshirani ting toward trees tumor uniform university variable vayatis view views with wyner yang zadrozny zakai zhang http://www.jmlr.org/papers/volume5/zhang04a/zhang04a.pdf 0 Hierarchical Latent Class Models for Cluster Analysis aaai absence acids activity advances advancesin aids akaike algorithm algorithms also alvord american amherst amsterdam analysis analyze annals annual applications approach approximating approximations area arnold arthritis arthur artificial assess assoc association asymptotic australia autoclass automatic banks bartholomew based bayesian belief belmont between biggar biological biology biometrics biometrika blattner bohrnstedt boston both british building cambridge case catchment center challenge characterize cheeseman chickering chloroplasts chow class classification clinical cluster clustering coil coleman community company compuserve computational computer computing conceptual condep conference connolly constructing control correlation cover cowell data dawid dempster denver department dependence depressive diagnosis dimension dimensionality direct directed discovering discovery discrete disorder dissertation distributions drummond dryman durbin eaton eddy edition edward effects efficient electrical elements elidan encyclopedia engineering epidemiologic eputten error espeland estimating estimation everitt evol expert exploratory exponential factor families fayyad finding formann fraley free friedman from garrett gaussian geiger gibson goedert gold goodman graphical green groups guide hagenaars handelman hannover hanson hasegawa heckerman henry hidden hierarchical homepages houghton house housing http icml identifiable identification ieee ijcai illinois incomplete indicators individual infection inference information inheritance innovations institute insurance intelligence international intertwining introduction itasca john joint journal jsuebersax kaufman kendall kindel king kishino knoke knott knowledge kohlmann koller kotz krogh laird langeheine lanterman latent lauritzen lazarsfeld learning levine liacs library likelihood local london look lotner lrgc machine magidson mail major marginal martin mass massachusetts mathematical mattussek maximum mccutcheon measurements meek meila menlo method methods mifflin mining missing mitchison mixtures miyata model models modified muenster multivariate murphy nashville networks neural nimh ninio nips nucleic order oregon origin other ourworld park parsimony part patterns paup peacock penalized phylogenetic phylogeny piatesky pittsburgh portland practical predicting predoviciu presence press probabilistic probability proceedings processing profile programme protein proteins psychiatry psychometrika publishers pupko putten qualitative raspe read references referral regional relative report research response results retroviruses review rheumatoid rheumatology rissanen rost rousseeuw royal rubin scandinavian schwarz science sciences scientific seattle selection sentient sequence sets severity shaoiro siam sinauer smyth social society sociological sociology some someren sons sorenson south spiegelhalter springer statistical statistics status stratified structural structure stutz sunderland suppl surveys swofford sydney systems table technical technology themes theories theory thomas three trait transactions trees uebersax uncertainty unidentifiable university unobservable update user using uthurusamy values vanlehn variable variables verlag vermunt volume wales wallace washington wasmus waxman weiss when wiley with york zeger zhang http://www.jmlr.org/papers/volume7/crammer06a/crammer06a.pdf 122 Online Passive-Aggressive Algorithms additive advances agmon algorithm algorithmic algorithms align altun annual appear approaches approximate april artificial automata automatic based batch beyond boostexter boosting bounded bounds boyd brain buckley cambridge canadian categorization category class classification cliffs collins computation computational conference confidence convergence convex crammer cristianini cruz dekel descent discriminative document editor eighth eleventh elisseeff englewood european evaluating experiments exponentiated family fast feedback first fourteenth free freund from gentile gerard gradient guestrin hall helmbold herbster hidden hierarchical hofmann http huji ieee implementation improved inequalities information interdependent international introduction january joachims jornal journal kernel kernels keshet kivinen klasner koller labeled language large learning line linear littlestone logarithmic long loss machine machines management march margin markov mathematical mathematics maximal maximum memory method metrics mistake model models multi multiclass music natural networks neural neurocomputing neurons noise norm novikoff online optimization organization output pages parsing pattern perceptron perceptrons polyphonic power prediction predictions predictors prentice press probabilistic problems proceedings processing proofs pseudo psychological ranking rated recognition references regularization relative relaxation relaxed relevance reprinted reranking research retrieval review robustness rocchio rosenblatt salton santa schapire scholkopf selective self seventeenth seventh shais shalev shawe shwartz signal simon singer single smart smola spaces statistical storage structured suffix support symposium system systems taskar taylor term text theory thesis threshold tolerant transactions trees tsochantaridis twentieth twenty ultraconservative university using vandenberghe vapnik vector versus volume warmuth watkins weighting weston wiley williamson with http://www.jmlr.org/papers/volume2/megyesi02a/megyesi02a.pdf 318 Learning to Construct Fast Signal Processing Implementations academic acoustics adapted adaptive algorithm algorithms appear applications april architecture auslander automated automatic based beauchamp boston brewer bryan burrus calculation carnegie complex computation computer computing conference cooley cosine data david decompositions department design discrete drexel dynamic electrical engineering eric faculty fftw fifth formula fourier francisco frigo from functions gavin generator hadamard haentjens high howard http ieee implementation implementations inductive international investigation jeremy jianxin john johnson jose journal june kaufmann lagoudakis learning level libraries library littman machine manuela markus master mathematical mathematics mellon methods michael michail modeling models morgan moura optimal optimization optimized padua pages parallel performance philadelphia pittsburgh platform portable porto practice prasanna predict press principles proceedings processing programming programs puschel quinlan references regression reinforcement related report robert ross science sciences search selection sepiashvili series seventeenth sidney signal sigplan singer software speech spiral statistical stochastic symposium technical thesis torgo training transactions transform tree tukey university using veloso volume walsh xiong http://www.jmlr.org/papers/volume7/angluin06a/angluin06a.pdf 182 Learning a Hidden Hypergraph adaptive additive algorithmica alon angluin annual apaydin application applied asodi attribute beigel benny chen closing computing conference cycle damaschke dana discrete edge efficient fortnow genome graph graphs grebinski gregory hamiltonian hidden hypergraph jiang journal kasif kucherov lance learning mapping matching mathematics mehmet model noga nonadaptive optimal pages peter physical press procedure proceedings queries querying recomb reconstructing reconstruction references richard rudich sequencing serkan shotgun siam simon springer steven subgraph sudakov symposium theory thirtieth under using vera versus vladimir whole http://www.jmlr.org/papers/volume8/list07a/list07a.pdf 215 General Polynomial Time Decomposition Algorithms aaccuracy active adaptive advances alexander algorithm algorithmic algorithms analysis annual appear application approach association assumptions asymptotic available bernhard bhattacharyya boser bottou boyd burges cambridge chang chen chih chiranjib christopher christos chung cjlin classification classifier classifiers clint combinatorial communications complexity computation computational computer computing conditional conference control convergence convex craig criteria csie david decomposition design detection dimension direction dunn edgar editors egham extremals face fast feasible federico fixed formal freund from full general generalized gilbert girosi gradient guaranteed guyon hall hans holloway hsuan hsuen http hush ieee improvements information ingo international isabelle jason joachims john journal keerthi kelly kenneth kernel krishnaj lagrangian large laskov learning leon liao library libsvm lieven linear linearconv list london machine machinery machines making mangasarian manual margin mark megiddo method methods minimal murthy murty musicant narasimha near networks neural nikolas nimrod note olkopf olvi optimal optimization osuna overrelaxation pages papadimitriou paper papers patrick pattern pavel peng platt polynomial practical prentice press proceedings processing programming rates recognition reference references regression report research robert rong royal sathiya saunders scale scholkopf scovel selection sequential series shevade shirish shuo siam signal simon simple simplesvm singular smola sons statistical steiglitz steinwart steven stitson stopping study successive support systems technical theorem theory thorsten tien time training transactions type ulrich university using vandenberghe vapnik vector version vishwanthan vision vladimir weston when wiley will with without working workshop http://www.jmlr.org/papers/volume7/centeno06a/centeno06a.pdf 119 Optimising Kernel Parameters and Regularisation Coefficients for Non-linear Discriminant Analysis academic adaboost advances agarwal alexander algorithm algorithms american analysis andrea annals annemie anouar anthony apart applied approach aronszajn baltimore bart baudat bayesian becker behavioural berlin bernhard beyond bishop boston bottou brabanter brian brodley california cambridge carla carroll charles christopher class classification classifiers college component computation computations conference connectionist cruz curve danyluk data david deepak dempster department design detection dietterich discovery discriminant discriminants donald dordrecht douglas duda edition editor editors eighth ellis esat estimating estimator eugenics fast fatiha fisher fitting framework from fukunaga fung gaston gaussian gene generalizations generalized geoff germany gert gestel glenn golub gramacy graphical guide gunnar hagan hamers hart herbert herbrich hinton holloway hopkins horwood ieee incomplete information informative international introduction jason johan john johns jolliffe joos jordan journal july katholieke kauffman keinosuke kernel kernels kevin klaus kluwer knowledge kristiaan label laird lambrechts lanckriet lang larsen lawrence learning least leen letters leuven likelihood limiting linear lker lkopf ller loan london lukas machine machines mackay mangasarian margins massachusetts mathematical mathematics matrix matthew matthias maximum measurements methodological methods michael michie mika mining models moor morgan multiple nabney nature neal neil netherlands netlab network networks neural noise obermayer olvi onoda optimal optimization optimizing outlier oxford pages pattern pelckmans peter plausible practical prediction predictions presence press principal probabilistic probable problems proceedings process processes processing proximal radford ralf raymond recognition references regression regularization report reproducing review richard ripley robert ronald roth royal rubin ruppert santa saul scene school sciences scientific sebastian seeger sejnowski semiparametric series seventh shrinkage sigkdd signal sista smola social society soft sparse spiegelhalter spirals springer squares statistical statistics summer supervised support suykens suzanna svmlab systems takashi taxonomic taylor technical technischen tell terrence theory thesis thomas thrun todd tony toolbox touretzky transactions tresp tsch universit universiteit university user using vandewalle vapnik vector verlag vladimir volker wand weiss weston wiley williams williamstown wilson witbrock with world yair york http://www.jmlr.org/papers/volume4/basalyga03a/basalyga03a.pdf 250 On the Influence of the Kernel on the Consistency of Supp ort Vector Machines ability advances analysis annual anthony applications approximation areas artificial associative available bartlett based berlin biological bottou bound burges cambridge carl central chapman classifiers comp compactness compint complexity computational computer conference consistency convergence crisp cristianini cross cybernetics data department dependent deutsch devroye dudley editors egham electronically empirical entropy fakultat frankfurt friedrich generalization generalized germany gradstein gyorfi hall harri herbrich hierarchies holloway http ieee influence informatik information integrals integraltafeln intel introduction jena jordan kernel learn learning leave lecture leen ligence limit lkopf london lugosi machine machines main manual math mathematik measures methods minet minimization muller neural nonlinear notes operators optimal orem other over pages pattern pedersen performance poggio port press probab probabilistic probability proceedings processes processing products produkt real recall recognition reference references report reports representer research rhbnc risk royal ryshik saunders schiller scholkopf science series shawe smola solution sources springer statistical steinwart stephani stitson structural submitted summen supo support tables taylor technical theo theorems theory tong trans uniqueness universit university vaart validation vapnik vector verlag weak wellner weston wiley williamson with york zhang http://www.jmlr.org/papers/volume7/debie06a/debie06a.pdf 155 Fast SDP Relaxations of Graph Cut Clustering, Transduction,and Other Combinatorial Problems advances algorithm algorithms analysis anjos applied approximation august balanced becker berkeley berlin blum boyd bundle burer california cambridge chapelle chawla cluster clustering combinatorial computer computing cones conf conference connections constraints convergence convex cristianini cuts data demonstration dietterich discrete division editors equivalence exploratory factorization finding from gaussian general ghahramani goemans good graph graphs habilitationsschrift helmberg hertz hillel iapr icml ieee ijcai image improved information intel interior international issue joachims jordan journal kamvar kandola kernel kernels klein konrad label labeled lang learning lifting ligence lisbon lkopf local machine malik manning mathematical mathematics matlab maximum method methods mincuts minima mixture models monteiro moor nearly neural nonlinear normalized obermayer optimization over partitioning pattern point portugal power press problem problems proc proceedings processing programming programs rank recognition references relaxation relaxations rendl report research review satisfiability science second sedumi segmentation semi semidefinite series shawe shental siam software solving special spectral statistical strengthened sturm supervised supplement suykens symmetric systems taylor technical text thrun toolbox transactions transduction transductive university unlabeled using vandenberghe weinshall weiss weston williamson with wolkowicz workshop xing yahoo zentrum zuse http://www.jmlr.org/papers/volume6/marchand05a/marchand05a.pdf 45 Learning with Decision Lists of Data-Dependent Features abound aditi advances alamitos algorithm andrzej annual anselm anthony attributes averages avrim blum blumer boolean bounds california cambridge cantelli class classifiers colt compression computation computational computer concepts conference covering cruz data david decision define dependent dhagat dimension ehrenfeucht eighteenth eiter error esko examples fifth floyd foundations from functions generalisation generalization geoffrey glivenko golea graepel half halfspace haussler heikki hellerstein herbrich hierarchical hinton ibaraki icml ieee information international intersections irrelevant isbn japkowicz john journal jyrki kaufmann kazuhiso kernel kivinen learnability learning letters linear lisa list lists littlestone machine makino manfred mannila marchand margin marina mario martin massachusetts mendelson michael mohak mona morgan mostefa nathalie network neural nevo nick nips occam online pages pairs phase points press proc proceedings processing publishers quickly rademacher ralf random razor reasearch references related relating report research revow rivest robert ronald rule sally sample santa science sets shah shahar shawe simple singh society sokolova spaces sparse sparsity splits statistical symposium systems taylor technical terms theoretical theory third thirteenth thomas thore threshold toshihide transactions transitions trees twentieth ukkonen university using vapnik vapnikchervonenkis vladimir volume warmuth when wiley williamson with workshop yaniv york http://www.jmlr.org/papers/volume2/tax01a/errata.pdf 327 A Neural Probabilistic Language Model acoustics advances aistats american amherst amsterdam analysis annual approach artificial association assp automatically backing backprop baker based beijing bellegarda bengio berger berlin biology book bottou brown category chen china class classification clustering code cognitive college colorado columbus communication como comparison component compression computational computer concepts conference connectionist continuous contrastive cowan curse data database deerwester della denver dept derived desouza dimensional dimensionality discovery discrete distributed distributional distributions divergence dongarra dumais dyer editors efficient eighth electronic elman empirical english entropy erlbaum estimation european eurospeech experts extensible extracting fellbaum finding florida forum framework from furnas gatsby gauvain gcnu gelsema giles goodman gram greece hanson harshman hidden high hillsdale hinton holland http icslp ieee ijcnn importance improved improving indexing information interface international interpolated issue italy jelinek jensen joint journal kanal katz kaufmann kneser knowledge krogh landauer language large latent lawrence layer learn learning lecun leen letter lexical lexicon linguistics ller london march markov mateo maximum mccallum meeting mercer message microsoft miikkulainen minimizing mining model modeling modelling models modular montr morgan multi multiple natural negative nets network networks neural niesler north ohio organizing orlando paccanaro pages parameters part passing pattern pereira phoneme pietra positive practice prediction press probabilistic probabilities proceedings processing products profiles progress propositions protein quick recognition recognizer references relations report representations research rhodes riis rudnicky sampling schmidhuber schutze schwenk science secondary self semantic sequence sequential sigir signal smoothing society solla source space span sparse special speech springer srilm standard statistical stolcke structure structured study systems taking technical techniques technology tenessee text time tishby toolkit trade training transactions tricks unit universit university unix using vocabulary volume walker whittaker with woodland word wordnet words york http://www.jmlr.org/papers/volume8/owen07a/owen07a.pdf 230 Infinitely Imbalanced Logistic Regression aaai active analysis applications approach artificial belmont binomial bolton breiman chawla chipman classification cohn computationally data detection editorial efficient estimates events existence explorations fraud friedman from ghahramani hand html http icml imbalanced intelligence issue japkowicz jordan journal king kolcz lago learning likelihood logistic maximum models newsletter olshen papers political proceedings rare references regression report research response review royal science series sets sigkdd silvapulle society special stanford statistical stone stong technical technometrics theory thesis tong trees university wadsworth with workshop zeng http://www.jmlr.org/papers/volume2/fine01a/fine01a.pdf 331 Dimensionality Reduction via Sparse Support Vector Machines academic accuracy advances aeronautics aiaa algorithm algorithms america american analysis anion annual applications applied approach approximation arcing arciniegas artificial association astronautics atomic automatic bagging barnhill bartlett based basis basque bennek bennett berlin boser bradley bredensteiner breiman breneman building burges cambridge campbell cancer chapelle chapman chemical chen chromatography classification classifier classifiers clean combinatorial combining comparison computation computational computer conf conference control country cplex cramer data decomposition demand dennis department derivative descriptor design detection development dimensionality discovery division donoho dreyfus economics editor editors embrechts environ exchange experimental extraction fayyad feature fifth free games gene generalized generation genetic geometry girala gorini grinstein guyon hall handbook harrison hart haussler hedonic high honavar icann ieee ilog image incline incremental information inns institute intelligence international inza issmo issue john joint journal july kaufmann kernel kernels kittler knowledge kohavi koza kudo kzaa larranaga lasso learning least linear london machine machines management mangasarian manual margin materials mathematical mccullagh merino method methods meyer miller mining model models modern momma monographs morgan mukherjee multidisciplinary nasa nature nelder network networks neural norm novelty optimal optimization ozdemir page pages parsimonious pattern pennsylvania philadelphia phillips pittsburgh plus poggio pontil population prediction predictors press prices probability problems proceedings processing programming programs property protein pudil pursuit qsar quantitative quiroga random ranking ratsch recognition reduction reference references regres regression report research reston retention ripley rosen royal rubinfeld saunders scholkopf science search second selection sensitivity series shimbo shrinkage siam sierra sion sklansky smola society somol song spain special specific springer sspr stanford statistical statistics stoppiglia structure subset sukumar support svms symposium systems technical technische theory thesis this throughput tibshirani times torczon training tugcu universitat university usaf using vapnik variable vector venables village visualization volume washington weston wierse williamson with work workshop wrappers yang york young http://www.jmlr.org/papers/volume7/kok06a/kok06a.pdf 168 Collaborative Multiagent Reinforcement Learning by Payoff Propagation aaai aamas academic academy accelerating adaptive adopt advances adversarial agent agents algorithm algorithms alife alspector amherst amsterdam analysis analyzing anytime approach approximate approximating april arai arbor artificial asada aspects asynchronous athena august australia automation autonomous banff barto basis bayesian becker behavior belief bernstein bertele bertsekas bled bounds boutilier boyan brioschi burgard cambridge canada categorization causal chalkiadakis changing chapter claus collaborative communication communications communicative complex complexity computer computing concise conference consistency constraint context control cooperate cooperation cooperative coordinate coordinated coordination cowan crick crites cybernetics dale dayan dcop decentralized dechter decision department difference discrete distributed durfee dutta dynamic dynamically dynamics editor editorial editors edmonton elevator elhorst empirical entertainment environments exchange exploiting exploration exploring factor factored faculty february flight formal freeman frey functions games gammon generalizations gerald goldman graphs greiner guarantees guestrin hague hale hansen helicopter icml ieee ijcai immerman improve improving independence independent inference informatics information initiative institute intelligence intelligent international introduction jaakkola jack january japan jennings joint jordan jose joshua journal july kaelbling kaufman kaufmann keepaway kitano kluwer knowledge koenig koller kschischang kuhlmann kuniyoshi lagoudakis learning lesser lianwen littman loeliger loopy machine madison magazine making march markov mateo mdps melbourne meuleau michigan millennium moallemi model models modern modi moore moors moreau morgan moving muhammad multi multiagent multiple murphy national netherlands nets network networks neural neuro nips noda nonserial note november observable observation october optimization optimizing ortiz osaka osawa packet pagello pages parker parr partially payo pearl performance perspective peshkin pfeffer planning plus policy poole preprocessing press probabilistic problem proceedings processes processing product programming propagation publishers puterman pynadath quality rationality reasoning references reinforcement report research riedmiller rish robocup robot robotics robots routing russ sastry saul scaling scheme schneider scholkopf schuurmans science sciences scientific search seattle sekaran sensor september shapley sharing shen signal simmons slovenia soccer spaan sparse specific stanford statistics stochastic stockholm stone strategies structured study sutton sweden sycara sydney systems tambe targets team teamwork technical techniques temporal tesauro theoretical theories theory thesis thrun transactions transition tree tsitsiklis uncertainty under understanding university using utrecht value venkataraman vlassis wainwright watkins weiss wiley willsky with without wong workshop world yedidia yokoo york zhang zilberstein http://www.jmlr.org/papers/volume8/grauman07a/grauman07a.pdf 229 The Pyramid Match Kernel: Efficient Learning with Sets of Features adaptable adelson advances affinely agarwal algebraic algorithm algorithms america american amherst amsterdam analysis anandan angles appearance application applications approach approximate april arregui august automated bags ballard barla based bayesian beijing belongie berg berkeley between beyond bileschi binary biological boughhorbel boughorbel boujemaa bray british building burt cambridge cameras canada caputo categories categorization categorizing category cateories challenge chang chapelle china city class classes classification classifier classifiers code collections color combining communications compact computational computer conditions conference content context contexts contour cormen correspondences cortex cristianini csail csurka cuturi cybernetics czech dance darrell data dealing december deerwester denver dependent descriptors diagonals diego diffusion dimensional dimensions direct discovering discovery discrimination discriminative distance distances distinctive distortion divergence division dumais early earth editors efficient efros eichhorn eklundh embeddings eskin estimation european examples exploiting fast feature features february fergus finite fisher fleuret framework france freeman fritz from frome functions furnas gartner gaussian generative google gool graf grauman greyvalue grossberg guibas hadjidemetriou haffner harshman hashing haussler hayman hazan hierarchical high histogram histograms hofmann holub hulle human hybrid icpr identification ieee image incremental indexing indyk inference inferring information intelligence intelligent inter interest intermediate international intl introduction invariant invariants jaakkola january jebara john joint journal july june kernel kernels keypoints knowledge kondor kullback kumar lafferty landauer laplacian large latent lazebnik learning lebanon leibler leiserson leslie leung library libsvm like local localized lodhi london lowe machine machines madison maire malik march massachusetts match matching material materials matrices measuring mechanisms mercer methods metric meyers mikolajczyk mining model models mohr montreal more moreno motion mover movies multi multiclass multimedia multiresolution mutch natural nayar nearest neighbor netherlands networks neural nice nine nips noble object october odone optical over overlapping parameter pattern perception perona platt poggio points ponce porter pose prague preattentive press principal principles proceedings processes processing program puzicha pyramid rasmussen real recipe recognition recognizing references regions regression relational relevance report representation representations representing republic research retrieval riesenhuber rivest roobaert rubner russell sawhney scale scene schmid scholkopf science semantic semigroup sensitive september sequences serre sets shaffalitzky shakhnarovich shan shape shashua shawe sift signal significance silhouettes singer sivic slncs society sons sparse spatial statistical strategies strings stripping structure structured suffix sukthankar support survey svms swain systems tarel taylor technical tested text textons texture thaper their theories theory thesis three tomasi training transactions transfer triggs tuytelaars university unlabelled unsupervised using vancouver vapnik vasconcelos vector vectors vehicle verri vert video videos view viola vision visual volume wallraven wang washington welling weston wiley willamowski williams with without wolf workshop world york zhang zisserman http://www.jmlr.org/papers/volume6/nakamura05a/nakamura05a.pdf 77 Inner Product Spaces for Bayesian Networks aaai adaptive advances akaho akira alamitos alexander alexei algorithm algorithmic altun analysis andras annals annual anthony approach approaches arrangements arriaga artificial aspects asymptotic atsuyoshi auer automata avrim balcan bartlett based bayes bayesian becker belief berlin bernhard between bianchi blum boolean boser bound bounded brendan cambridge carlos case central cesa chapman chickering christopher circuits classes classification classifiers clinical cocoon cohn colt combinatorial combinatorics communication communications complexity computation computational computer computing concept concepts conference connectionist contemporary control craig cryptographic dale daphne david decision depth devroye diffie digital dimension discriminative distributed dorronsoro duda dudley editor editors eike eiron embeddings empirical engines error estimating euclidean exploiting extensions field finite fisher florina forster foundations francisco frankl frey from function functions gabor generalized generative given graphical graphs guestrin gunnar guyon gyorfi gyorgi habib hajnal half halfspaces hall hans hariharan hart haussler heckerman hellman hidden hierarchical hilbert hofmann icann ieee inference information inner intelligence international into ioannis isabelle jaakkola john johnson jones jordan jose journal judea jurgen kaufmann kawanabe kearns kernel kernels kiltz klaus knill knowledge koji koller krause large largest laszlo laurence lawrence learning leave leaving lecture lemma limit limitations lindenstrauss linear lipshitz local lokam london lowdimensional lower lugosi maass machine machines madhavan maehara manfred mapping mappings margin margins maria mario markov martin maruoka masayuki mathematics matthias maxwell mccullagh mean measures meek meir menlo michael models morgan motoaki mubarakzjanov mukund muller multiplicative nadav nakamura nathan national natural neal nelder network networks neural nicolo niels norm notes numao nuria obermayer oliver optimal pages park pattern pavel pearl peter possi possible predicates press probabilistic probability proceedings processing product projection properties pudlak radford ramesh random rank ratsch recognition references regression regularization reischuk relations representation representing research reverend richard robert robust rosa royal rudiger rustam santosh sara satyanarayana saul saunders scene schmitt scholkopf schuurmans science sciences sebastian series shai shawe shotaro shraibman sigmoid signal simon singer smallest smola society software solla some sonnenburg sons soren spaces span sphericity spiegelhalter springer springerverlag srebro state statistical statistics string structure support suttorp symposium system systems szegedy taskar taylor technology theorems theoretic theoretical theory thirteenth thomas thorsten threshold thrun tommi trace training tsochantaridis tsuda turan ulrich unbounded uncertainty university vapnik vector vempala verlag vinay vinokourov vishwanathan vladimir volume warmuth warnow whittaker wiley with wolfgang workshop yasemin yoram york http://www.jmlr.org/papers/volume1/boyan00a/boyan00a.pdf 337 Grafting: Fast, Incremental Feature Selection by Gradient Descent in Function Space additive advances algorithm algorithms american annals annual approach architectures artificial association available bartlett based baxter biased blake boosting boser california cambridge chang chapelle cjlin classification classifiers combining computation computational computer conf confidence csie data databases dept dictionaries discrimination dualities edition editors edwards elements entropy estimation experiments feature fifth filters flannery frean frequency freund friedman function functional fung girosi gradient grafting guyon hastie hoerl html http hybrid hypotheses icml ieee improved information institute intelligence irvine jaakkola jebara john jones journal kauffmann kaufmann kennard kernels kira kohavi large lasso learning library libsvm logistic machine machines madison mallat mangasarian margin mason matching maximum mead merz method minimization mining mlearn mlrepository morgan mukherjee nelder networks neural newton nonorthogonal numerical olkopf optimal pages perkins pittsburgh poggio pontil practical predictions press problems proc processing projection pursuit rated recipes references regression regularization rendell report repository ridge schapire scholkopf schuurmans science sciences selection sept shrinkage signal simplex singer sleeman smola software springer statistical statistics stuetzle subset support svms systems technical techniques technometrics teukolsky theiler theory tibshirani time toronto training transactions uncertainity university using vapnik vector vetterling view weston wisconsin with workshop wrappers zhang http://www.jmlr.org/papers/volume7/evendar06a/evendar06a.pdf 142 Action Elimination and Stopping Conditions for the Multi-Armed Bandit and Reinforcement Learning Problems action adaptive advances adversarial aiello algorithms allocation amer american anal angluin annual appear appeared appl applied approximate approximately approximation armed aspects association asymptotically athena auer bandit barto based belmont berry bertsekas bianchi borkar bound bounded bounds buffer bull cambridge casino cesa chapman circuits competitive complexity computation computer computing conference control convergence decision delayed design differentiated dynamic efficient even experiments exploration fast finite fitted foundations fourteenth freund fristedt from functions gambling hall hamiltonian hoeffding howard icml ieee indirect inequalities infocom information international interscience iteration journal kakade kaufmann kearns kernel kesselman langford learning loss lotker machine macqueen management mannor mansour markov matchings math mathematics method meyn modified morgan multi munos near neural neuro nineteenth optim optimal ormoneit overflow page pages patt policies polynomial preliminary press probabilistic probability problem problems proc proceedings processes processing programming puterman queue rajagopolan random rates references reinforcement research rewards rigged robbins rosen rules sample sampling schapire schieber science sciences scientific sequential services shamir siam singh sixteenth society some statistical stochastic sums sutton sviridenko switches symposium system systems szepesvri theory thesis time tsitsiklis university upper valiant value variables version watkins wiley http://www.jmlr.org/papers/volume6/drineas05a/drineas05a.pdf 102 š On the Nystrom Method for Approximating a Gram Matrix for Improved Kernel-Based Learning acad academic accepted achlioptas advances algebra algorithm algorithms analysis annual applications approximate approximating approximation approximations april aspects azar baltimore based belkin belongie bengio bhatia biological burges cambridge carlo chung classifiers clustering component compressed computation computational computations computer computing concept conference contemporary cristianini cybernetics data decision decomposition delalleau delves density department diffusion dimensionality discrete distribution donoho drineas edinburgh editor effect efficient eigenmaps eigenvalue embedding equations extensions fast february fiat finding fine foundations fowlkes framework freund frieze gaussian generalized geometric girosi global golub goreinov gram graphs greedy greville grimes grouping harmonics haven hessian heterogeneous highdimensional hopkins horn ieee improved information input institute integral intelligence international introduction inverses isomap israel iterative johns johnson journal july kannan karlin kernel kernelbased lafon langford laplacian large learning linear loan local locally lowrank machine machines mahoney malik manifolds maps march massachusetts mathematics matrices matrix maximum mcsherry method methods mika mohamed monte muller multiplication nashed natl networks neural niyogi nonlinear nystrom observations osuna other ouimet pages paiement pass pattern perturbation planck prediction press problem problems proc proceedings process processing projective pseudoskeleton publication rademacher rank rasmussen reduction references report representation representations research roux roweis rules saia sample sampling saul scheinberg scholkopf schwaighofer science seeger shawe siam signal silva simplified smola sparse spectral speed springer springerverlag stewart support symposium systems taylor technical techniques technology tenenbaum theoretical theory thesis training transactions tresp tyrtyshnikov university using vector vempala verlag view vinay vincent volume wang weinberger williams workshop yale yaleu york zamarashkin http://www.jmlr.org/papers/volume7/zhao06a/zhao06a.pdf 193 On Model Selection Consistency of Lasso analysis angle annals appear asymptotics berkeley biased boosted buhlmann computational computing consistent curved degrees department dimensional donoho dual efron elad estimation estimators freedom graphical graphs hastie high hoerl journal kennard knight knot large lasso least leng matrices meinshausen methodologies model neighbourhood nips noise nonorthogonal note optimization osborne overcomplete paths preprint presence presnell problems procedures random recovery references regression regularized related relaxation report representations review ridge rosset royal science selection series shrinkage sinica society solution sparse spectral splines stable statistica statistical statistics submitted technical technometrics temlyakov tibshirani tracking turlach type wahba with zhao http://www.jmlr.org/papers/volume3/caruana03a/caruana03a.pdf 277 Text Chunking based on a Generalization of Winnow aaai abney abound additive advances algorithm annual association attributes average based berwick brill buchholz categorization chunk chunking chunks classifiers combination computation computational concepts conference conll construction dagan damerau david dietterich dordrecht driven dual editors elements empirical eric erik exponentiated extraction formulation fred gentile gradient grammar grammars grove halteren hans hidden hinge identification ijcai inference information introduction irrelevant johnson journal karov khardon kivinen kluwer kudo kudoh language learning leen linear linguistics littlestone logic loss machine machines margin matsumoto mccord meeting methods michael mistake models muslea naacl natural neural nips pages parsing part patterns practical prediction press principle proc proceedings processing psycholinguistics punyakanok quickly references regularized relational roth rule sabine sang second sequential shared simple slot some speech support survey system systems tagging taku tasks tenny text thomas threshold tjong todd tong tresp updates using valiant variables vasin vector versus volker warmuth when winnow with workshop wpdv yuji zhang http://www.jmlr.org/papers/volume7/chen06a/chen06a.pdf 189 Consistency of Multiclass Empirical Risk Minimization Methods Based on Convex Loss advances american analysis annals association bartlett based bayes behavior boosting boucheron bounded bounds bousquet chen classification classifier combinatorics combridge complexities comput computational conference consistency convex convexity devroye differences error esaim gaussian gyorfi international john jordan journal large learning lugosi machine machines margin math mcauliffe mcdiarmid mendelson metheds method methods minimization multicategory multiclass nonparametric pages pattern press principles probabilistic probability proc rademacher recent recognition references regularized reseach research results risk soft some sons springer statist statistical statistics structural support survey surveys tewari theory university vapnik vayatis vector verlag wien wiley xiang ying york zhang zhou http://www.jmlr.org/papers/volume6/dekel05a/dekel05a.pdf 55 Smo oth -Insensitive Regression by Loss Symmetrization adaboost additive advances algorithms analysis annals annual approach approximation april based bennett bianchi body boosting bregman brownian cesa collins composition computation computational computer conference descent distances drifting duffy elements equation exercise exponential exponentiated fisher freund friedman function games generalized geometric girosi gradient greedy gupta hastie helmbold huber ieee incorporating information international into issue january john journal kivinen knowledge lafferty learning lebanon leveraging likelihood line linear lkopf logistic machine machines maximum measurement medicine models motion nelson networks neural neurocolt neurocomputing nineteenth opper pages penrose poggio prediction predictors prior proceedings processing rahim references regression report robust rochery schapire science sciences september simple singer smola sons special sports springer statistical statistics support system systems technical techniques theory thirteenth tibshirani tutorial twelfth using vapnik vector versus view volume warmuth wiley york http://www.jmlr.org/papers/volume5/stracuzzi04a/stracuzzi04a.pdf 22 Randomized Variable Elimination aaai abound abstract accelerating addison additive advances algorithm algorithms almuallim amherst anaheim analysis angle application applied approach approximation architecture artificial attribute attributes backpropagation bankert based bayesian blake bond brain brunswick california cambridge capital cardie caruana cascade case cases chapelle cherkauer chimerge classification cloud communications comparative comparision comparison components computation computer computers conference context cooley corporate correlation damage data databases decision denker department derivatives devijver dietterich discovering discovery discretization distributed domingos dunteman editor editors edwards efron eleventh empirical engineering enhancement etxeberria european exploratory exponentiated facilitate fahlman feature features feng filter finnoff flotzinger fourteenth frean freitag friedman function future gallant genetic goldberg gradient graphs greedy growing hall hamilton hassibi hastie hergert hoeffding ieee improve improving induction information intelligence international introduction inza irrelevant john johnstone jong jose kaufmann kerber king kira kittler kivinen knowledge kohavi koller kubat langley large larranaga layered lazy learnable learners learning least lebiere lecun linear littlestone lohnes machine many markets maron mateo mcclelland merz methods mining mitchell model moody moore morgan mukherjee multivariate national network networkbased networks neural newbury ninth nonconvergent notes numeric oblivious optimal optimization order orginization pages parallel park past pattern patterns perceptron performance pfleger pfurtscheller poggio pontil practical prediction prentice press principal probabilistic problem problems proceedings processing programs projection pruning psychological publications pursuit quickly quinlan races randomized rating reading real reasoning recognition refenes references regression rendell report repository representations restructuring review rosenblatt rule rumelhart sage sahami science search seattle second selection sensitive setino setiono shavlik shutherland sierra signals simpler sleeman society solla solution sons space stanford statistical statistics statlog storage stork stracuzzi study subset surgeon svms systems technical tenth theory thermal thesis threshold thurstone tibsharini tool tools toward transaction transactions trees tukey types university updates using utans utgoff vafaie valiant vapnik versus volume volumes waikato warmuth werbos wesley weston when wiley with working workshop world wrappers york zealand zimmermann http://www.jmlr.org/papers/volume3/seeger02a/seeger02a.pdf 370 PAC-Bayesian Generalisation Error Bounds for Gaussian Pro cess Classification academic adaptive advances alex alexander algorithmic algorithms although american analysis andr annals anthony appear applied approximate artificial association assumed autoreason averaging barber bartlett based bayes bayesian becker bernhard beyond bibby bound bounding bounds bousquet brain brian burges california cambridge carlo carnegie caruana cbms chapman charles christopher citeseer classification classifiers cohn committee communications complexity compression computation computational computer conference continuous convex cover cristianini cruz csat cumulative curves data david department dependent dietterich dimension discovery division edinburgh edition editor editors elements elisseeff entropy error factors family fast field flannery from gaussian generalization generalized ghahramani gibbs grace graepel graphical greedy green hall haussler heckerman herbrich hierarchies horn html hybrid ieee ihara implementation index industrial inference informatics information informative institute intel introduction jaakkola january john johnson jordan journal kaas kaufmann kearns kent kernel kernels kluwer know known kullback langford lawrence learnability learnable learning ledge leen lehel ligence line linear littlestone luckiness machine machines manfred mardia margin mark martin massachusetts mathematics matrix matthias mcallester mccullach mean meir mellon method methods metric michael minimal minimization mining minka mixture model models monographs monte morgan muller multivariate mutual neal neil nelder neural nick nonparametric notation number numerical nystrom observational olivier opper optimization ordered organization other over pages papers pattern perceptron peter platt points prediction press princeton probabilistic probability process processes processing psychological radford raftery ralf recipes recognition references regional regression relating relative relevance report research review rich risk robert rockafellar roger rosenblatt sample santa saul schapire scholkopf science scientific seeger selection sequential series sets shawe shef sheffield shunsuke siam silverman smola society solla sollich some sons sparse sparsity speed spline splines stability statistical statistics stochastic storage structural support svms systems taylor technical technology telecommunications teukolsky theorems theory thesis thomas thore tipping tommi tong toronto training transactions tresp true tutorial uncertainty university unordered using valiant vapnik vector vetterling vladimir volker wahba warmuth whittaker wiley william williams williamson winther with work workshop world zhang http://www.jmlr.org/papers/volume2/zhang02a/zhang02a.pdf 350 Recommender Systems Using Linear Classifiers aaai advances algorithm algorithms analysis andrew applications approach approaches apte architecture artificial asia august automating baltimore based basic bayes bergstrom biased blake bottom breese california cambridge carl categorization chickering christopher chute classification classifiers clustering collaborative combination commerce compaq comparison compression computations computer conald conference context continuous cortes cscw damerau data databases david decision dependency dept discovery dumais eachmovie ecml edition editors emperical empirical estimation european evaluation event example fast features fifteenth fifth filtering formal foster fourteenth frank generalization glassman goetz golub grouplens hampp heckerman hoerl hofmann hopkins html http iacovou icml ieee inductive inference inform information intel international irvine iyengar joachims john johns johnson jordan journal july kadie kamal karypis kaufmann kearns kennard keogh know konstan learing learning ledge ligence ligent linear loan machine machines maes management mansour many mapping markov matrix maximizing maxwell mccallum meek merz method methods mining mlearn mlrespository model models morgan mouth mozer naive national near netnews networks neural nigam nonlinear nonorthogonal oles open optimal pacific pages pattern performance petsche platt predictive press problems proceedings processing programs properties pruning publishers quinlan recognition recommendation recommender references regression regularized relevant repository representations research resnick retrieval ridge riedl ripley robert rounthwaite sahami sarwar science shardanand shtarkov social sons statistical study suchak support systems technometrics text their theory third tjalkens tong trans transactions tree tresp uncertainty ungar university using vapnik variables vector visualization weighting weiss wiley willems with word workshop yang yiming york zhang http://www.jmlr.org/papers/volume2/zhang02b/zhang02b.pdf 358 Covering Numb er Bounds of Certain Regularized Linear Function Classes abound additive advances agnostic akad alexander algorithm algorithmic alon amer american analysis annals anthony application applications approximation association asymptotic attributes average averaging banach barron bartlett based bayesian berlin bernhard bianchi boosting bounded bounds burges cambridge cantelli capacity carnegie cesa chapter characteristics chervonenkis christopher city class classes classification classifiers colt combinatorial completely complexity comput computation computer convergence course covering cristianini data david decision density dependent devroye dimension dimensionbased dimensions discriminant discrimination dokl dual dudley editors effectiveness efficient empirical entropy events explanation exponentiated families formulation foundations frequencies freund from function functional functionals general generalization generalizing gentile geometric glivenko gradient greedy grove gurvits gyorfi haussler hierarchies hilbert hinge hoeffding ieee important inequalities information interior introduction iowa irrelevant isbn isoperimetry issn jaakkola jebara jection john jones journal kernel kernels kivinen kolmogorov langford learnability learning lecture ledoux leen lemma line linear littlestone loss lugosi machine machines march margin marina martin math mathematical mathematics maximum mcallester meila mellon mendelson methods metric michael michel minimization model more muller nauk nello network networks neural nips note notes number numbers operators other over pages pattern performance peter point pollard pontriagin potential practice prediction press probabilistic probabilities probability proc proceedings processes processing programming propri pursuit quickly random randomized rates recognition reduction references regression regularization regularized relative report results risk robert sample sauer scale schapire schnirelmann scholkopf schuurmans science seeger sensitive series sets shahar shawe sigmoidal simple size smola solla some sons space spaces springer sssr statist statistical stochastic structural sums superpositions support symposium syst systems talagrand taylor technical than their theoretic theoretical theory threshold tihomirov todd tommi tong tony training trans transactions transl trique uniform universal university updates vapnik variables vector verlag versus volume voting warmuth weights when wiley williamson with ying yoav york zhang http://www.jmlr.org/papers/volume3/auer02a/auer02a.pdf 309 Multiple-Instance Learning of Real-Valued Data acta aided aligned ambiguity annals annual appear applications approach approximating aromatic artificial auer axis based bauer berry blum chapman chapter chemistry chimica classification compass computer computing conference critchlow decision department design dietterich dissertation distributions doctoral drug empirical evaluation examples fehr forces framework francisco from galindo geometric goldman haubrichs helvetica hyper information instance intelligence intermolecular international italy jain john kalai kaufmann lathrop lazy learning long lozano machine maron mathematics molecular morgan mult multiple musk natural neural ninth note odorants page parallel patterns perez perret physical press problem proc proceedings processing product pseudo random ratan real rectangles references regression report respect rice ross ruffo scene science scott security sets shape single solving sons srinivasan symposium synthesis systems technical theoretical theory tool torino trees turin twenty university valued wang webster wiley with zucker http://www.jmlr.org/papers/volume3/bousquet02b/bousquet02b.pdf 308 Efficient Algorithms for Decision Tree Cross-validation agrawal algorithm algorithmic algorithms artificial attributes automatic bagging based belmont berlin blockeel boosting breiman california ceedings clark classification classifier clustering combining communication computer conference constraint continuous cross data database databases decision dehaspe demoen department discovery discretization down dzeroski efficiency efficient eleventh error estimated european experiments extending fast fayyad feature fifteenth fifth first foil fourth francisco freund friedman html http improving induction inductive information intelligence international interval irani irvine janssens john joint journal kaufmann king knowledge kohavi laer large learning lecture logic logical machine mehta merz meta methods midterm minimizing mining mlearn mlrepository models moore morgan muggleton multi multiple murphy mutagenicity niblett notes olshen order packs pages parallel parameter personal pkdd power practice predictors principles proceedings programming programs query quinlan raedt ramon references regression report repository research restructuring rissanen scalable schapire science selection series shafer sixth sliq springer springerverlag sprint srinivasan sternberg stone struyf study tables technology theories theory thirteenth through todorovski tree trees twelfth university utgoff validation valued vandecasteele verlag very volume wadsworth with workshop http://www.jmlr.org/papers/volume7/mukherjee06b/mukherjee06b.pdf 191 Estimation of Gradients and Coordinate Covariation in Classification acad accepted advances algorithms amer american analysis annals annual applications applied aronszajn asian association atomic barnhill bartlett based basis behavior beyond biology bloomfield bounded bounding bounds bousquet buhmann caligiuri cambridge cancer chapelle chen choosing class classification classifiers coller combinatorics component computation computational computer computing conference consistency convergence convex convexity coordinate cortes covariances data decomposition detection differences dimensional discovery donoho downing editor elements empirical equivalent evgeniou expression feature friedman function functions gaasenbeek gene gine girosi golub gradients guyon hastie hermes high huard image international ionizing jordan journal kernel kernels koltchinskii lander lasso learn learning lecture liang linear machine machines manifolds mason math mathematics mcauliffe mcdiarmid mesirov method methods micchelli microarray microarrays minimization modeling models molecular monitoring mukherjee multicategory multilayer multiple natl networks neural notes object observational optimization pages panchenko papageorgiou parameters pattern philadelphia poggio pontil prediction predictive press probability proc proceedings processes programming pursuit quadratic rademacher radiance radiation recognition recomb references regression regularization representations reproducing research response risk royal satellite saunders schoelkopf science scientific selection series shrinkage siam siemons significance slonim smola smoothing spline splines springer stat statis statistical statistics submitted support surveys tamayo that theory tibshirani trans tusher understanding unlabeled using vaart valued vapnik variance vector verlag versus vision volume wahba weak wellner west weston wiley with york zhang zhou http://www.jmlr.org/papers/volume8/khardon07a/khardon07a.pdf 211 Noise Tolerant Variants of the Perceptron Algorithm ability adatron advances aggressive agnostic algorithm algorithms analysis annual applying approximate association automata bagging based batch bianchi blake boolean boosting boser brain burges cambridge campbell cesa classification classifiers cohen collins communications comparison computational computing concept conconi conference constructing convergence conversions crammer cristianini data databases decision dekel dietterich distribution driven editors effective efficient ensembles entity european experimental extraction fast fifth fourth framework freund friess from functions gallant garg generalization generic gentile graepel guyon herbrich hettich hill html http icml ieee improve induction information international introduction joachims journal kandola kearns kernel kernels khardon kivinen kowalczyk krauth large learnable learning like line linguistics littlestone long machine machines making mansour margin margins mathematical maximal maximum mcgraw meeting merz methods mezard mitchell mlearn mlrepository model named networks neural newman nips noise novikoff online optimal order organization pages passive perception perceptrom perceptron perceptrons perspective physics pitt practical press probabilistic proc procedure proceedings processing proofs psychological randomization ranking recent references relaxed repository research results review rosenblatt roth rule scale schapire scholkopf second selective sellie servedio shalev shawe shwartz siam signal simple singer sixteenth smola sparsity stability storage support symposium systems taylor theory three toward training transactions trees tsampouka twelfth understand uneven university using valiant vapnik vector voted voting weak williamson winnow with workshop zaragoza http://www.jmlr.org/papers/volume6/lafferty05a/lafferty05a.pdf 35 Diffusion Kernels on Statistical Manifolds accuracy advances aizerman alan alexander algorithm algorithms amari american anal analysis andrew annals appear applic applications applied approach april artificial asymptotic attainable automation barndorff bartlett based bases becker belkin berger bernhard boser bounds bousquet bradley braverman bruce bull bulletin calcutta calyampudi cambridge carnegie categorisation categorization cencov chengxiang classes classification classifiers comments comparison complexities composite computational conference construct control covering craven cristianini croft david dawid dayne decision dietterich differential diffusion dipasquo discrete discriminative divergence document dortmund editors edmond edward efron emmanuel equivalent estimation exploiting exponential foundations frank frederico freitag from function functions further gauduchon gaussian generative geometrical geometry george girosi gous grace graphs grigor guyon haussler hayward heat hiroshi hoffmann hyperbolic hypertext ichi icml ieee imre inference information input institute intelligence international isabelle isri jaakkola january joachims john journal kamal kass kaufmann kernel kernels kimeldorf knowledge kondor kullback labeled lafferty language laplacian lauritzen learing learning lebanon lecture lectures leibler lewis linear local london machine machines manifold manifolds marc marcel margin mark masakazu math mathematical mathematics maximum mazet mccallum mellon mendelson method methods michael mikhail milnor mitchell modeling models monographs moreno morgan morse multilayer multimedia nagaoka nello nelson networks neural nevada nielsen nigam nikolaevich nikolai niyogi noguchi notes numbers nuno obermayer oles olivier optimal other pages paper parameters partha partially pattern paul pedro performance perish peter philip poggio ponte potential press princeton probability proceedings processing publish purdy rademacher recognition references regularization regularized remote research results retrieval richard riemannian riemannienne ringuette risi risk robert rosenberg rozonoer rules sammut schoen science sean seminar sept series shahar shawe shing shun sigir slattery smoothing society some sons space spaces spectre spherical spivak spline springer stanford statistical statistics stefan steven structural structure study subfamily support symposium systems taylor tchebychean tensor text that theoretical theory thesis thorsten thrun tomaso tommi tong topology training trans translation translations tung univ university using vapnik variete vasconcelos vector vegas verlag vladimir volume wahba wide wiley williamson world ying zhai zhang http://www.jmlr.org/papers/volume7/watanabe06a/watanabe06a.pdf 125 Stochastic Complexities of Gaussian Mixtures in Variational Bayesian Approximation advanced akaike algebraic algorithms alzer analysis analytic annals aoyagi application approaches approximate artificial asymptotics attias based bayes bayesian beal berkeley bernald bounds castelle college complexities complexity computation conference conic cybernetics dacunha dimension estimating failure field function functions gamma gassiat gaussian generalization ghahramani graphical hartigan honor identifiable ieee ieice inequalities inference inferring information intelligence intelligent international interpolation japanese journal kiefer kullback latent layered learning levin likelihood locally london lower machines mackay mathematics mean methods mixture mixtures model modeling models networks neural neyman normal online opper pages parameters practice press probability proc procedure proceedings references report rissanen saad sato schwarz selection singapore singularities solla some spain statistical statistics stochastic structure systems technical testing theory thesis tishby uncertainty university upper valencia variable variational watanabe yamazaki http://www.jmlr.org/papers/volume8/xue07a/xue07a.pdf 205 Multi-Task Learning for Classification with Dirichlet Process Priors academic across active advances agricultural algorithm algorithms american analysis ando annals annual antoniak application applications applied approach artificial aspects association bakker based baxter bayes bayesian beal bias biological biometrika blei boltzmann boston breaking burr cambridge carlo caruana casella chain chapter classification clustering cognition collaborative colt combining component computation computational conference constructive content data definition density department dept development dietterich dirichlet discovering distributed dominici doss edition editor editors educatinal effects employing ensemble environmental escobar estimation evgeniou evidence exchangeable experiments explorations extensions ferguson filtering framework from gating gaussian gelfand generalized ghahramani gibbs glass good graphical group heskes hierarchical hierarchically hinton hoff independent inductive inference information informative intelligence internal international introduction ishwaran jaakkola james jordan journal kernel kluwer latent lawrence learn learning leen lindley linear literature logistic machine machines mackay mallick markov mccallum mcclelland meta method methods micchelli microstructure mixed mixture mixtures model modeling models monte mukhopadhyay muller multi multiple multitask neal networks neural nigam nonparametric page pages parallel parmigiani platt pontil pool pratt predictive press primary priors problems proceedings process processes processing propagation publishers quintana random reckhow references regression regressions related relearning report representations research researcher retrieval review robert rosner royal rumelhart sampling saul scholkopf schwaighofer science second secondary section sejnowski semiparametric series sethuraman several sigir sinica sixth society some springer statistica statistical statistics stick structure structures sullivan systems task tasks technical text their theory thrun toronto tresp tribute uncertainty university unlabeled using variational vector verlag volume walker washington weiss west with wolpert workshop yang york zhang http://www.jmlr.org/papers/volume7/lecue06a/lecue06a.pdf 137 Lower Bounds and Aggregation in Density Estimation adaptation adaptive aggregating aggregation akaike aleatoires algorithm algorithms alternative american analysis annales appear approach arbitrary artificial association augustin available barron bayesian bernoulli biometrics birge buckland bunea burnham catoni ccsd center classification cnrs combinatorial combining complexity computational conditional convex density descent devroye different distribution ecole edition entropy epsilon estimate estimation estimators flour from functionnal hartigan hasminskii haven heidelberg html http ibragimov ieee impose index inference information integral intelligence introduction journal juditsky jussieu kernel laboratoire learning least lecture lecue leung likelihood linear lmens lomi lower lugosi machines manuscript marge mathdoc mathematical mathematics maximum mean methods minimum mirror mixing mixture model modele multivariate nauchn nazin nemirovski newyork nobel nonparametric notes online optimal optimization parametric parametrique paris part penalized performance prediction preprint preprints priors proba probabilites procedures rates references regression regressions report research rigollet saint scholkopf selection semin simultaneous springer square stat statist statistical statistics stochastic strategies studies submitted tech testing textes theory topics transactions tsybakov universal universites university using vayatis warmuth watson with yale yang zhang http://www.jmlr.org/papers/volume3/barnard03a/barnard03a.pdf 292 PAC-Bayesian Generalisation Error Bounds for Gaussian Pro cess Classification academic adaptive advances alex alexander algorithmic algorithms although american analysis andr annals anthony appear applied approximate artificial association assumed autoreason averaging barber bartlett based bayes bayesian becker bernhard beyond bibby bound bounding bounds bousquet brain brian burges california cambridge carlo carnegie caruana cbms chapman charles christopher citeseer classification classifiers cohn committee communications complexity compression computation computational computer conference continuous convex cover cristianini cruz csat cumulative curves data david department dependent dietterich dimension discovery division edinburgh edition editor editors elements elisseeff entropy error factors family fast field flannery from gaussian generalization generalized ghahramani gibbs grace graepel graphical greedy green hall haussler heckerman herbrich hierarchies horn html hybrid ieee ihara implementation index industrial inference informatics information informative institute intel introduction jaakkola january john johnson jordan journal kaas kaufmann kearns kent kernel kernels kluwer know known kullback langford lawrence learnability learnable learning ledge leen lehel ligence line linear littlestone luckiness machine machines manfred mardia margin mark martin massachusetts mathematics matrix matthias mcallester mccullach mean meir mellon method methods metric michael minimal minimization mining minka mixture model models monographs monte morgan muller multivariate mutual neal neil nelder neural nick nonparametric notation number numerical nystrom observational olivier opper optimization ordered organization other over pages papers pattern perceptron peter platt points prediction press princeton probabilistic probability process processes processing psychological radford raftery ralf recipes recognition references regional regression relating relative relevance report research review rich risk robert rockafellar roger rosenblatt sample santa saul schapire scholkopf science scientific seeger selection sequential series sets shawe shef sheffield shunsuke siam silverman smola society solla sollich some sons sparse sparsity speed spline splines stability statistical statistics stochastic storage structural support svms systems taylor technical technology telecommunications teukolsky theorems theory thesis thomas thore tipping tommi tong toronto training transactions tresp true tutorial uncertainty university unordered using valiant vapnik vector vetterling vladimir volker wahba warmuth whittaker wiley william williams williamson winther with work workshop world zhang http://www.jmlr.org/papers/volume7/sonnenburg06a/sonnenburg06a.pdf 160 Large Scale Multiple Kernel Learning accurate adaptive advances algorithm algorithms also alternatively altman alto analysis annual applications bach basic becker bennett berlin between biocomputing bioinformatics biological biology boosting bousquet boyd brodley burges caenorhabditis cambridge canberra canu categorization chapelle choosing classification classifiers column combination communications computational computer conference conic considerations convex corrective cortes cristianini curves data davis dellec demiriz discovery duality dumouchel dunker ecml editor editors efficient eighth elegans embrechts ensembles eskin european exons fast fawcett feature features finite first framework fredkin fusion gehrke generation genomic germany goadrich grandvalet graphs hawaii heidelberg heterogeneous hettich hunter hyperkernels hypothesis icml inexact infinite information international interpretable introduction issue istrail january joachims jordan journal kasif kaua kernel kernels klein knowledge kohavi kortanek kuang laboratories lanckriet large lauderdale learning lecture leslie leveraging liao lkopf lncs machine machines madison making many margin mark matching maximization maximize medicine meir memory mendelson mesirov methods metz minimal mining mixture miyano model models molecular momma mukherjee multiple networks neural neurocolt noble notes nuclear obermayer october optimization pacific pages palo parameters pevzner platt potsdam practical precision prediction press principles proc proceedings processing programming protein raedt rase recall recognition recomb references regression relationship relevant report research researchers review robust rouveirol scale scaling school science selection semi seminars sequence sequential series siam sigkdd site smola sonnenburg spaces sparse special spectrum splice spliced springer statistical string summer support svms symposium systems technical tenth text that theory thesis thrun totally training trie tsch twenty university using vandenberghe vapnik vector verlag viii volume warmuth waterman williamson wisconsin with wrobel york zhang http://www.jmlr.org/papers/volume7/ekdahl06a/ekdahl06a.pdf 190 Bounds for the Loss in Probability of Correct Classification Under Model Based Approximation affect affinity after aided aktuarietidskrift almost american analysis annals annual another anoulova applied approximating approximations areas arimoto artificial association bacterial bahadur barbosa based bases basic bayes bayesian belief berkeley bernoulli bhattacharyya bias binary binomial biomedical biometrics biometrika block bound bounded bounds boutilier brown brunk california cambridge causality channel chapman chapter characteristics chickering chow class classes classification classifications classifier classifiers clustering clusters coefficients communications comparison complex complexity computation computational computer computers computing conference considerations consistency context control cooper cover cowell cramer cross curse data dawid decentralized decisions decomposition densities density dependence dependency dept derived detection devroye diagnosis dichotomous differential dimensionality discovery discrete discrimination distance distinguishability distribution distributions divergence divergences domingos dover ecml edition editor ekdahl elements elliptically engelen equalization estimation estimators european evaluation expert explain exponential extensions factorisation families feature features finite fischer five formula foundations friedman from functions funktionaler geiger gelpke general geometric ghahramani glick goldszmidt goodman graphical greedy gruyter gyllenberg gyorfi habbema hall hand head hellerstein hermite hoeffding http huang icann identification idiot ieee improve independence independent inequalities inference information injured institute intelligence intellignce international introduction item items jaakkola joint jordan journal kailath kaleh king knowledge koller korb koski kronecker kronmal kruskal kullback lauritzen learning lecture lewis licentiate likelihood linkopings loss lugosi machine mannila mathematical mathematics matrix matusita maximum mcgraw means measure measures mehler methods mining misclassification mixed mixture model models moore multinomial multivariate murray naive nets network networks neural nguyen nicholson note notes obner optimal optimality orthogonal oxford page pages patients pattern pazzani performance philosophical pierce pistone pitman polt polynomial polynomials power prediction press princeton principles printing probabilistic probabilities probability probably problems procedures proceedings propagation publications recognition reduce references related removal report representation representations requirements research responses review riccomagno ripley rish risk royal ryzin sais sample sankhya saul scheinck schervish science search second selected selection semi sequence series sets siam signal silvey simon simple skandinavisk skene slepian smyth society solomon some specific spiegelhalter springer stanford statistical statistics stochastische storage strasser structure studies stupid surrogate swedish symmetric symmetrized symptom systems taxonomy technical techniques technology terms teugels that thathachar theoretical theory thesis thomas titterington topsoe toussaint transactions transmission trees unabhangigkeit uncertainty under universitet university upper using valued variables variance variances variational view vilata vilmansen volume wagner wainwright walter wiley williamson with wolfowitz workshop wynn zero zusammenhang http://www.jmlr.org/papers/volume7/takeuchi06a/takeuchi06a.pdf 148 Nonparametric Quantile Estimation aaai acquisition adaptive advanced advances algorithm algorithmica algorithms american analysis annals annual applied approximation asian association august averages bachrach baltimore bartlett based bassett bayesian berlin bernoulli biased bickel biometrika bivariate black bone bosch bounded bounds bousquet buhmann cambridge carroll caucasian cbms class classifiers clinical cohn compact complexities computation computational conf conference constrained convergent crossing curves dagstuhl data dependence dependences dimension distribution distributions econometrica editors efficient elements empirical endocrinal entropy estimation estimators extreme fawcett forthcoming framework frie friedman function fung furuhashi gartner gaussian general generalization geometric golowich hall hastie healthy high hispanic hoeffding hoerl hopkins huber ieee inequalities information international intl inversion john joint jordan journal kearns kennard kernel kernels klaassen knowledge koenker learning lectures linear lnai localized longitudinal maass machine machines mammen mangasarian marcus marron mathematics mendelson metabolism method mineral minimal mishra models mozer murty narashimhan nature neeman networks neural nina nonorthogonal notes number numbers observational operator operators pages pattern petsche philadelphia plate portnoy press probability problems proc processing product programming projection quantile quantiles rademacher random recognition references regional regression regressions regularization relevance report representation research results ridge risk ritov ritter robust royal ruppert sample scholkopf science seminar semiparametric series shawe siam signal simpler simplesvm single size smola smoothing society solla sons sparse spline splines springer statistical statistician statistics structural study sums support systems tajvidi takeuchi taylor technometrics tensor theory thin tibshirani tipping tishby transaction turlach university unsupervised value vapnik variables variance vector vishwanathan volume wahba wand wang washington wellner wiley williamson with without woodworth york youth http://www.jmlr.org/papers/volume7/laskov06a/laskov06a.pdf 172 Incremental Support Vector Learning: Analysis, Implementation and Applications academic accurate active adaptive advances alche algorithm algorithms amari analysis angluin anomaly aosvr appear applications approach approximation arnold asymptotic back backprop baltimore based becker benchmarks bengio bishop blankertz bordes bottou brussels burges california cambridge catalunya cauwenberghs chakrabarti chang changing chapter chemical cheung class classification classifiers collobert commands computation computations computer concept core cortes cruz curio data decomposition decremental departament department description detecting detection dietterich diettrich dimensional dimva direction discovering discovery distribution domain dornhege dortmund drug duin edition editor editors efficient engineering environments error ertekin esann eskin estimating facto fast feasible feature framework from function functions geometric ghahramani golub good heidelberg high hopkins http hypertext ieee ijcai ijcnn incremental inequalities informatics information input introduction intrusion intrusions isbn japanese joachims john jordan journal kaufmann kawanabe kernel kernels kivinen klee kluwer knirsch knowledge kohlmorgen kotenko krepki kunzmann kwok lanl large laskov learning lecture lecun leen lemmem liao libsvm line linear littlestone llengatges lncs loan local london long losch machine machines making martin math mathieson matrix method methods mika minimal minimization mining minty mohri molina morgan motor mozer muller munro murata national networks neural nips nnsp notes novelty october online optimization oxford pace page pages papers pattern perkins petsche platt poggio politecnica portnoy practical prerau press problems proc process processing putta quarter queries ralaivola rate rates ratsch recognition references regression rehabilitation report research robbins ruping saad santa scale schafer scholkopf science sciences security september sequential series sets shawe simplex single sisha sistemes smola space sphere springer stat statistical stochastic stolfo study sung supervised support svmtorch syed systems taipei taiwan taylor technical theiler theory thesis time tokyo trade training transactions tresp trial tricks tsang universitat university unlabeled unsupervised using vapnik vector vectors verleysen versitat very volume warmuth wesdon wiley williamson with workshop york ziehe http://www.jmlr.org/papers/volume4/mendelson03a/mendelson03a.pdf 256 Uniform Ob ject Generation for Optimizing One-class Classifiers aaai advances akaike algorithm applications applied approach automatic bartlett bishop burges bution cconference class classification classifier cohn computers conference congress control data davidt december delft dempster description discovery distri distribution domain doubly duin editors efficient elementary estimation extracting fayyad first from generating generation given handwritten hostetler http hypersphere identification ieee incomplete information international june kearns know koch laird learning ledge letters likelihood look luban machines maximum menlo method mining model moya network networks neural november numerals object oxford pages park pattern physics points portland press proc proceedings processing recognition references regression rubin scholkopf selforganizing shawe shrinking smola society solla sons statist statistical statistics staunton street sung support systems target task taylor technology theory thesis transactions tube tudelft tutorial ullman unconstrained uniform university uthurusamy vapnik vector walton wiley williamson within world http://www.jmlr.org/papers/volume7/kaempke06a/kaempke06a.pdf 177 Distance Patterns in Structural Similarity adel ahuja algorithm algorithmics algorithms amsterdam analysis anand antoine apostolos appearance application applications approach approximately arthur assignment baltimore barone based basque bengoetxea berlin biology blueprint bruce bruno bulletin bunch bunke caelli caetano cambridge carl case champin charles chemical chemistry christine christoph classification cliffs combinatorial common computations computer computers computing conference country currey dale dante data database david decision dennis desper detection dexa diego digest dissertation distance distribution edge editors elastic elsevier endika engineering englewood erhard error estimates estimation expert face fellous fifth finding fitting flooding flows forthcoming foster francisco freeman from fujishige functions garcia garey gene geometric george giugno godefrey gold golub gopalakrishnan graduated graph graphical graphs grid hall hardy hector high histograms hlaoui holland hopkins horst icde icpr ieee image inequalities inexact infrastructure intelligence international intractability isomorphism james jason jean john johnson journal kaizhong kampke kathleen kesselman kruger kurt labeled largest laurenz learning leastsquares leda length littlewood lncs loan lowe machine magnanti malsburg manolopoulos mapping marc martin matching mathematical matrix measuring mehlhorn melnik messmer michael models molina naher network norbert north object optimization ordinary orlin pages pami papadopoulos partial patent pattern pierre platform point polya pope prentice press principles printing proceedings quebec rahm rangarajan ravindra reasoning recognition reconstruction references representation restriction richard rosalba satoru scalable schema schuurmanns screening search searching sebastian sergey shapiro shasha shengrui similarity site skiena solnon springer stephan steven structure structures subgraph submodular substructures sundaram symposium system systems terry thomas throughput tiberio tolerant topological transactions tree trees university using versatile vingron vision wang wiskott with workshop yannis zhang http://www.jmlr.org/papers/volume1/mangasarian01a/mangasarian01a.pdf 340 MLPs (Mono-Layer Polynomials and Multi-Layer Perceptrons) for Nonlinear Modeling addendum advanced advances aire algorithm algorithms analysis anders antoniadis appear application applications applied approach architecture baltimore bancaires based bates bayesian berlin berruyer billings black brain carmona chen cherkassy comparisons computation computations conditioning confidence conjugate connectionnist construction constructive control cross damage data denker dependences determining doctorat draper economica elements empirical error estimation expansion extensions feedforward financi friedman from functions germany golub gression hassibi hastie hidden hopkins hypotheses iconip icsc identification ieee information international intervals jacobian john journal july kaufman kaufmann korn kwok learning least lection leontaritis leuwen linear loan mateo matrix methods mlps model modeling moody morgan nato network networks neural neuronaux neurons neveu nonlinear number objective optimal orthogonal outputaffine paris pattern performance personnaz prediction press prior problems procedure proceedings processing progress pruning recognition references regression risk rivals seber second selection series smith solla sons springer squares statistical statistics statistiques stoppiglia stork structure surgeon symposium system systems taylor techniques tests their theory thodes tibshirani touretzki training transactions units universit university using validation vapnik verlag vila wagner watanabe watts wechsler wild wiley wolff workshop yeung york http://www.jmlr.org/papers/volume5/luxburg04b/luxburg04b.pdf 239 Reducing Multiclass to Binary: A Unifying Approach for Margin Classifiers ability adaboost adaptive additive advances algorithm algorithms allwein alternating american analysis annals annual appear application arcing artificial association bakiri bartlett baxter berkeley boost boosting bounded bregman breiman brooks california chap classification classifiers codes cognition collins combining complexity computational computer computing conference confidence correcting cortes coupling crammer csiszar databases decision decisions della department design dietterich distances distributed divergences down edge effectiveness eighth error explanation explorations features fields fifth frean freund friedman functional games generalization generalized geometry gradient guruswami hastie hinton hoeffding hoffgen html hypotheses ieee important improved inducing inequalities inference information intelligence internal issue journal kaufmann kearns lafferty large learnability learning line logistic machine majority mansour margin margins mason mcclelland merz methods microstructure minimization mlearn mlrepository models more morgan muller multiclass murphy nature network networks neural neurocolt neurons olshen onoda output pairwise parallel pattern pietra prediction predictions press probability problems procedures proceedings processing programs propagation quinlan random rated ratsch reducing references regression repository representations research robust rumelhart sahai sample schapire scholkopf schuurmans sciences simon singer single size smola soft solving springer statistical statistics stone sums supplement support symposium system tech techniques than theoretic theory thirteenth tibshirani trainability transactions tree trees tusnady twelfth twenty university using vapnik variables vector version view voting wadsworth weights williams williamson with workshop http://www.jmlr.org/papers/volume3/getoor02a/getoor02a.pdf 297 Tracking a Small Set of Exp erts by Mixing Past Posteriors abstract adaptive additive advice aggregating algorithm algorithmica algorithms also annu appear appeared artificial auer august averaging bayes berlin best bianchi binary boolean bousquet bylander cesa coding communication completeness complexity comput computation computational computer computers concept conf conference confident derandomizing direct disjunction disjunctions distributed drift editors eurocolt european expert exponentiated extended fischer francisco freeman freund functions game garey general gentile gradient guide hans haussler helmbold herbster http identically ieee independent indirect individual info information intel intractability issue january johnson jority journal kaufmann kivinen learning ligence line linear littlestone loss lossless machine march memoryless merhav methods morgan neurocolt pages panizza paul piecewise portfolios posted prediction predictions predictor private proc proceedings references report research schapire science sciences self september sequences sequential shamir simon singer source sources special springer stationary stochastic strategies switching system technical theoretical theory tracking trans transactions ulrich uncertainty under updates versus vovk warmuth weighted willems with workshop http://www.jmlr.org/papers/volume3/reunanen03a/reunanen03a.pdf 283 Learning Monotone DNF from a Teacher that Almost Do es Not Answer Membership Queries algorithm algorithmic algorithms almost amsterdam analogical angluin annals annu annual answer april artificial auer berlin bianchi boolean bounds bshouty california castle cesa chen closure colt comput computational concept conference cruz dana david does donna eiron environments equivalence eurocolt european fallible formulae formulas from germany goldman incomplete inductive inference inform intel international issue jority july kearns learnability learning lecture leonard leslie ligence line linear littlestone logarithmic machine malicious manfred mathematics mathias membership michael ming mistake monotone nadav nader netherlands nick nicol noise noisy note notes october oracle owshanko pages peter pitt press proc proceedings queries randomly rectangles references regular reinhardsbrunn report sally santa sets sixth slonim special springer sympos teacher teachers technical term that theory thesis threshold ucsc univ using valiant verlag volume warmuth weighted with workshop york zhixiang http://www.jmlr.org/papers/volume7/lippert06a/lippert06a.pdf 133 Infinite- Limits For Tikhonov Regularization advances alex amer analysis annual appl applied architectures aronszajn asymptotic asymptotics behaviors berlin bernhard between bottou burges cambridge case cbms changjiang chih citeseer comp complexity computation computational conference covering crisp data davis department duraiswami editors efficient evgeniou extensions fast federico gacv gauss gaussian generalized girosi grace herbrich html industrial info information joachims jones keerthi kernel kernels larry lawrence learning least leon lippert machines massimiliano math mathematics method michael models networks neural nonstandard number observational pages platt poggio pontil press proc processing proved ralf ramani references regional regularization regularized relation report representer reproducing rifkin rockafellar roger ross ryan sathiya saul scholkopf series smola solution spline springer squares statistics support systems technical theodoros theorem theory tomaso trans transform tuning tyrrell uniqueness using variational vector volume wahba weiss wets wisconsin with yair yang yoonkyung zhang zhou http://www.jmlr.org/papers/volume4/mcallester03a/mcallester03a.pdf 261 Efficient SVM Training Using Low-Rank Kernel Representations acoustics adaraton advances algorithm algorithms andersen annual application approach approximation argonne arising baltimore barrier bartlett bennet blake boser burges campbell chapter choi cholesky classifiers cohn college columns compact complement computation computational computations computing conference cristianini databases dense dept development dietterich discriminative dual edition editor editors efficient entropy exploiting factorization factorizations factors fast fine fletcher form freund friess from further generalization generative gill girosi goldfarb golub gopinath greedy guyon handling haussler holloway hopkins html http hubrid icassp iden ieee implementation improved information interior international jaakkola joachims johns kaufman kaufmann kearns kernel laboratory large learning leen linear loan london machine machines making manuscript margin marxen mathematical mathematics mathematik matrices matrix mehrotra merz method methods mininal mlearn mlrepository models modification modified modifies modifying monma morgan murray national natural nature navratil networks neural neurocolt note numbers numerisches nystrom oliver operations operators optimal optimization orsa osuna pages performance philadelphia platt point powell practical preprint press primal problems proc procedure proceedings processing product programming publishers quadratic rank references regulariza regularization report repository representation representations research royal saunders scale scheinberg scholkopf schur schuurmans seeger sequential shanno siam signal simple smola software solla sparse speaker speech speed springer stanford statistical submitted support systems technical theory thomas tification tion todd training transactions tresp triangular trining university using vapnik vector verlag volker volume williams williamson workshop wright http://www.jmlr.org/papers/volume3/singer02a/singer02a.pdf 305 A Robust Minimax Approach to Classification academic advances alarc algebra algorithm analysis andersen anderson annals applications approach arcing asymptotics athena available bahadur bayesian becker belmont bennett bernardo bertsekas bertsimas bhattacharyya bottou boyd bredensteiner breiman burges cambridge chapelle chebyshev class classification classifiers computation cone cones conf convex course covariance crisp cruz detection diana diego dietterich different discriminant distributions douglas duality edition editors fast fisher francisco frenk fukunaga geometric geometry ghahramani ghaoui groot guez high homogeneous http ieee implementation inequalities information insead interior international interpretation into introduction issue jordan kadane kass kaufmann kernels kluwer lanckriet larsen learning least lebret leen letters lindley linear lobo machine marshall mathematical matlab matrices methods mika minimax minimization morgan mosek muller multivariate nature navia nemirovsky nesterov networks neural nonlinear normal notes novelty olkin optimal optimization optimizer order over oxford pages pattern performance philadelphia point polynomial popescu press probability proc processing programming publishers ratsch recognition references relevance report risk robust rodr roos scholkopf scientific second sedumi siam signal single smith smola software solla special springer squares stanford statistical statistics sturm supplement support suykens symmetric systems technical terlaky theory tierney tipping toolbox training tresp university using vandenberghe vandewalle vapnik vazquez vector vicinal weston wilson with zhang http://www.jmlr.org/papers/volume6/eibl05a/eibl05a.pdf 37 Multiclass Boosting for Weak Classifiers adaboost advances alex algorithm algorithms allwein amit analysis annals annual application approach artificial bagging bakiri bartlett bauer baxter bernhard binary blake boosting california catherine changing christopher classification classifiers code codes combined combining comparison computational computer conference confidence correcting databases decision department dietterrich digit direct effectiveness eibl empirical ensemble eric erin error european example experiments explanation frean freund functional generalization ghulum gradient gunnar gunther guruswami html http hypotheses improved improves information intelligence international irvine jonathan journal karl klaus kohavi large learning line llew machine make marcus margin margins mason merz methods mika mlearn mlrepository muller multiclass national nips online only onoda optimization output performance peter pfeiffer predictions problems proceedings quinlan rated ratsch recognition reducing references repository research robert robust ross sahai schapire scholkopf science sciences sebastian simulated singer smola soft solving statistics system takashi techniques theoretic theory thirteenth thomas twelfth unifying university using variants venkatesan voting weak with work yoav yoram http://www.jmlr.org/papers/volume2/hammerton02a/hammerton02a.pdf 314 Some Greedy Learning Algorithms for Sparse Regression and Classification with Mercer Kernels adaptive additive adler advances akaike algo algorithm algorithms american analysis annals applications approximate approximation asilomar assisted association atomic backward bartlett basis bayesian bengio between block boosting bounds bresler buhlmann burges cambridge chen chow classification classifiers communications comparing comparison computation computational computers computing conference corrections couvreur cowan criterion daniel data decomposition delgado description dictionary discussion donoho equivalence evgeniou examples experiments factorization finite fortran francisco free frequency freund friedman from function functions further gaussian girosi gradient gragg gram grant greedy grote grove hansen hastie huckle ieee information international inverse inverses iterations journal judd kauffman kaufmann kernel kernels kreutz large learn learning least length linear local logistic loss machine machines mallat matching mathematical mathematics matrices matrix mees method methods minimum model models morgan natarajan nature networks neural niyogi noise noisy nonlinear numerical nystrom occam optimality orthogonal pacific parallel partitioned pattern physica poggio pontil preconditioners preconditioning press principle problem problems proceedings process processing pursuit radial razor recognition references regression regularization reichel relevance reorthogonalization report research ripley rithms saad saunders schaback schapire schmidt scholkopf scientific search seeger selecting selection seminar series siam signal signals smola software solution solutions some sparse speed springer squares stable statistical statistics statistik stewart submitted subroutines subset sugiura sung support systems techniques tibshirani time tipping transactions university updating using vapnik vector verlag view vincent wendland williams williamson with york zhang zurich http://www.jmlr.org/papers/volume5/chen04a/chen04a.pdf 7 Image Categorization by Learning and Reasoning with Regions access advances algorithm analysis andrews application applied approach archiving artificial asymptotically attribute auer automatic axis barnard based bayes bayesian belongie berlin blake blei blobworld block blum burges cambridge capital carson chapelle characterization chen city classification clustering competition composite computational computer computing conf content cristianini curves cuts data database databases daubechies depth dietterich digital discovery duygulu edited edition empirical engineering evaluation examples expectation feature features field figueiredo flannery flexible forest forsyth frames framework freeman freitas fritts from fuzzy gdalyahu gersho glance goldman gorkani graphical gray greenspan growing haffner hall hartigan hierarchical histogram hofmann html http huang human icml ieee image images improved indexing indoor information instance integrated intelligence interpretation into introduction investigation jain joachims jordan journal kalai kernel kernelbased knowledge kumar large lathrop learning lectures libraries linguistic lippman lozano machine machines making malik manjunath markov maron marr matching maximization means merz methods mining mlearn mlrepository model modeling modern modestino mult multi multiband multimedia multiple multipleinstance multiresolution murphy natural navigating netra networks neural normalized note numerical object objects olkopf optimum orientation other outdoor pages parallel pattern perez photos picard picture pictures ponce port practical prentice press problem proc processing quantization querying raedt ramon random ratan recipes recognition rectangles references region relating relational repository representation research retrieval scale scene scenes scenic scheme scientific second segmentation semantic semantics sensitive shawe silhouettes simplicity smith smola snakes solving sorting spie springer statistical statistics storage strat support syntactic systems szummer taylor technique techniques templates teukolsky texture theoretical theory toolbox torralba transactions trees tsochantaridis tutorial understanding unifying university unser unsupervised using vailaya value vapnik vasconcelos vector verlag vetterling video vision visual wang wavelet wavelets weinshall wiederhold with wolf wong words workshop yang york yuille zabih zhang http://www.jmlr.org/papers/volume7/ye06a/ye06a.pdf 146 Computational and Theoretical Analysis of Null Space and Orthogonal Linear Discriminant Analysis academic adaptive algebra algorithm algorithms american analysis applications applied approach association authentication baltimore based belhumeour bellmanna berry beyond bioinformatics biology biometric brie california characterization chen class classification classifier classifiers comparative comparison complete component computational computations computer computers conference control covariance data deerwester diego dimension dimensional direct discriminant discrimination duchene duda dudoit duin dumais edition eigenfaces eigenfeatures elements error european expected expression extraction face family feature fisher fisherfaces foley framework frangi fridlyand friedman fukunaga furnas gene generalized geometric golub guided hall harshman hart hastie hespanha high hopkins huang ieee image indexing inference information intelligence intelligent international introduction inverse janardan johns jolliffe jonathan journal kernel kernels kpca kriegman krzanowski landauer latent learning leclerq letters liao linear loan machine machines marron matrices matrix mccarthy methods mining multiclass neeman networks neural null ogihara optimal optimization pages park pattern pentland plataniotis plus prediction press princeton principal problem problems proc processes projection pseudo raudys recognition references regularization regularized representation research retrieval review royal sammon sample schokopf scienc selection semantic series siam singular size skurichina small smola society solve solving space specific spectroscopic speed springer stabilizing statistical statistics stork study support swets system theory third thomas tibshirani tissue tour trans transformation tumors turk uncorrelated undersampled university using vapnik vector vectors venetsanopoulos verlag very vision wang weng which wiley with workshop xiong yang york zhang http://www.jmlr.org/papers/volume7/shimizu06a/shimizu06a.pdf 175 A Linear Non-Gaussian Acyclic Model for Causal Discovery adaptive algorithm algorithms american analysis annual application approach artificial assignment association asymptotic asymptotically beamforming benjamini bentler biometrika blind bollen bonferroni bootstrap bounds british brockwell browne bulletin burkard buseco california cambridge cardoso causal causality causation cela chapman charleston clustering coefficient coefficients combinatorial comon comparison component components computational concept conference confounding contrast control controlling correlation covariance cramer data davis demixing dependencies direction discovery distribution dodge edition editors efron equation equations equivariant esposito estimating estimation european extensions false fast fastica finding fixed forecasting free functions garrat gaussian geiger glymour godambe graphical hall handbook have heckerman himberg hochberg holden holm hoyer http hyndman hyvarinen ieee improved improvements independent inference information intelligence international interscience introduction jenkins john journal june kano karhunen kawanabe kerminen kluwer koldovsky laheld latent learning library likelihood linear machine mathematical mean methods mixing mixture modeling models monash muller multiple national networks neural neuroimage neuroimaging normality oakland optimization ordering oxford pages pardalos pearl performance permutation personal pham planning point powerful practical prediction presence press probabilistic problems proc procedure procedures proceedings processing properties psychological psychology quasimaximum rate reasoning references regression rejective research revised robust rousson royal scandinavian scheines search separation sequentially series setting sharper shimizu signal signals significance simes simple society software sons souloumiac source sources spirtes springer statdata statistical statistician statistics structural structure structures supplement tamhane test testing tests theoretical theory third through tibshirani tichavsky time trans trusted tsdl umass uncertainty unit university unix using validating variables variance verlag visualization volume when wiley with workshop york yuan http://www.jmlr.org/papers/volume5/kaariainen04a/kaariainen04a.pdf 14 Selective Rademacher Penalization and Reduced Error Pruning of Decision Trees academic accuracy achim agnostic alexey algorithm almuallim analysis andrzej annales annals annual anselm appear application applications armand artificial auer available averaging bartlett based berlin bernhard best birkhauser blake blockeel blumer bohanec boston bottom bounded bounding bounds bousquet bratko breiman california cambridge cathrine cesabianchi charles chervonenkis christopher christos classification classifiers claude colin colt combinatorics combining comparative comparison complexities complexity computable computational computer computing concentration conference convergence corrected cristianini data databases david decision decomposition department dependencies designing differences difficulty dimension dimensional dmitry donato douglas dragan ecml editor editors efficient ehrenfeucht eibe eighteenth elomaa empirical error esposito estimation european evarist events extension faculte fast fernando fifteenth fisher floriana fourteenth francesco francisco frank frequencies friedman func gabor gamberger gaussian generalization gine giovanni goldman good grigni grove hand haussler heidelberg helmbold hendrik high hoffmann holte html http hunch hussein icml icsc ieee implementations induction inequalities information intelligence international introduction irvine ivan java jerome john jonathan journal jude jyrki kaariainen kaufmann kearns kernel kivinen koltchinskii langford lavrac learnability learning lecture letters ljupco local localized london lozano lugosi maass machine machines malerba malinen manfred mansour marco marten mason massart mateo mathematical matti mcallester mcdiarmid mendelson merz method methods michael michelangelo mingers minimization mining mirelli mixture mlearn mlrepository model morgan nada naiso national nearly nearoptimal nello networks neural nicolo nineteenth note notes occam oliver olivier olshen optimal over pacific pages panchenko papadimitriou pascal pattern penalization penalties pereira pessimistic peter petra philips practical predicting prediction presented press prieditis printing probabilities probability proceedings processes processing programs progressive projects pruning prunings quinlan rademacher random razor reduced references regression regularization relative repository research results richard risk robert ross russell sally sammut sampling schapire scholkopf science sciences second selection selective semeraro series seventh shahar shavlik shawe shell siam siemons simplicity simplifying singer sixteenth size sloan small society some springer statistics stone structural stuart studies subclass support surveys symposium tapio taylor techniques test their theory thirteenth tion todorovski tools toulouse trading training transactions tree trees tutorial twelfth uniform university using vaart vapnik vector vincent vladimir volume wadsworth warmuth weak wegkamp well wellner with witten wolfgang workshop yishay yoram york http://www.jmlr.org/papers/volume3/szita02a/szita02a.pdf 366 ­MDPs: Learning in Varying Environments aamo aamodt abstract adaptive advances agent ahuja algorithms alsp amsterdam annals applications approximate approximation architecture article artifical artificial aspects asynchronous attila august automatic autonomous bari barto basc based behavior bellman between bolyai bounded bourgine boyan cambridge cesses chastic cimmer citeseer cognitive college comp computation computer computing conf conference connell consideration context continuous control controller convergence cowan dayan dean decision decomp delayed dels department dietterich difference discrete domains double doya dular dule dynamic dynamica dynamics dzielinski ector editor editors eecg ehavior eleventh elte endent endulum engineering ensatory eria erience european event examples experimental exploration feedback feudal files first fomin forthcoming framework fransisco from function furuta game games general generalized givan grid gross group gullapalli hagen harmelen hasselmo heger heuristics hierarchical hinton html http hunt hwang indirect information institute integrated intel intelligent international introduction inverse iteration iterative iwashiro jaakkola jersey john jordan journal kaelbling kalkkuhl kalm kaufmann king knowledge leach learning life ligence linear littman lorincz lrincz lyon machime machine maes mahadevan making markov markovian master mataric mateo mathematical mathematics maxq mdps ments metho monro morgan motion move mozer multi multiple national navigation networks neural neuro ninth nipg novemb olicy optimal oral orally oratory organizing osition page pages parameter patha people planning practice precup preliminary press princeton proceedings processes processing programming puterman real references reinforcement representing research resolution results rewards risk robbins robotic robust rozgonyi scales scaling scientific seattle self semi septemb sequential singap singh sons space state static statistics stochastic strategies sugahara survey surveys sutton swing systems szep szita takacs temp tenth tesauro theoretical theory thesis time toronto touretzky toward tsitsiklis twelfth university using utoronto value varela variable volume watkins when wiley with world yamakita york zbikowski zsef http://www.jmlr.org/papers/volume2/fuernkranz02a/fuernkranz02a.pdf 359 Round Robin Classification aaai acoustics advances algorithm algorithms alicante allwein alpaydin amsterdam anaheim analysis anand andr angulo another appear application applications approach approximate architectures artifi artificial assistant atlanta australia back bagged bagging bakiri based bauer belgium berthold between binary blake boost boosting boswell breiman brodley brooks bruges brunswick building burges california cambridge case catala chapter chemnitz cial clark class classification classifier classifiers clouse codes cohen combination comparative comparing comparison computation computer conditions conference confidence connectionist conquer constructing correcting correlated cortes coupling cristianini current dags danyluk data databases decision decomposition dellec department design dichotomies dietterich difference digit directions dreyfus ecml effective efficient empirical engineering ensemble ensembles error esann european evaluation ewsl experimental expert exploiting explorations fast feature feelders first fisher florida fogelman fort four francisco frank freund friedman from function furnkranz fusion generalization germany gish grove hand handwritten hastie hirsh html http hyperlink hypertext icassp icml identification identifying ieee implementation implementations improved improvements incremental induction information initial inspired intel intelligent interface international into irvine issue issues iwann java jordan journal kalousis kaufmann kearns kernel kinds kittler knerr kohavi kolen krieger lake large lauderdale layer learner learning learns leen ligence ligent line lists long lopez machine machines maclin magazine mantaras margin mateo mayoraz mcnemar mehrotra menlo merz method methodological methods mining mira mlearn mlrepository modular module mohan moreira morgan most muller multi multiclass multiple nashville national nato natural nearest neighbor netherlands network networks neural neurocomputing niblett nips noemon noisy note olshen opitz output outputs pacific pairwise park pattern percentages performance personnaz pfahringer platt plaza pollack polychotomies polychotomous popular porto portugal practical predictions predictors preferences preparation press price prieditis probabilistic problems procedure proceedings processing programs propagation proportions pruning psychometrika publishers pyle quinlan randomization ranka rated rault recent recognition reduced reducing references regression relations report repository research results review revisited rivest robin roli round rouveirol rule russell sampling sanchez schapire schmidt scholkopf science sciences selection sensitive separate series session shawe sigkdd signal simple singer single smola solla solving some souli spain speaker speakers special speech springer stanford statistical statistics stepwise stone structural studies study subsets supervised support svcr sydney symposium system systems tahoe task taylor technical techniques tesauro tests text theoharis theoretic three tibshirani tools touretzky training transactions trees unifying university using utgoff vapnik variants vector verkooijen verlag verleysen volume voting wadsworth watkins weston which widmer williamstown winning winter with witten work working workshop wyner http://www.jmlr.org/papers/volume6/lawrence05a/lawrence05a.pdf 90 Probabilistic Non-linear Principal Component Analysis with Gaussian Process Latent Variable Models aaron academic adrian advances alexander algorithm algorithms alternative american analysis annals anthony antti applications approximate approximations association aston bartholomew based basilevsky bayesian becker behavioural berger bernardo bernhard between beyond bibby binary birmingham bishop bouttou brendan calculus cambridge charles chichester choice christopher cohn component computation computers computing conjugate connection control csat cuts data david dawid dayan densitometry density dietterich differential dimension dimensional dimensionality dordrecht dual econometrics edition editor editors eigenvalue embedding energy engineering estimation exploratory factor family fast feed flows forward framework frey from gamma gaussian generative geoffrey geometric global goodness gradient graphical graphics griffin grochow gwilym hagan harri hasselmo heinz herbrich hertzmann high hinton honkela hypothesis ieee image inference infinite information informative institute instruments intelligence inverse isbn iterative james jianbo jitendra john jordan joseph joshua journal kantilal kearns keith kent kernel kernels kinematics klaus kluwer kohonen kruskal kullback langford latent lawrence learning leen lehel leibler linear lkopf ller locally london lowe machine mackay magnus malik mapping mappings marcus mardia marketing martin massachusetts mathematical mathematics matrix matthias maybeck method methods metric michael minka model models mozer multidimensional multiphase multivariate neal neighbor neil netherlands networks neudecker neural nonlinear nonmetric normalized nuclear numerical obermayer optimizing organizing oxford pages pattern peter petsche phillip physics popovic prediction preference press principal principled probabilistic problem proceedings process processes processing proximity psychometrika radford raftery ralf reduction references regression related research richard robert roweis royal sammon sara saul scaled scaling science sciences sebastian seeger segmentation self series siggraph silva sleep smith smola social society solla solomon some sons sparse statistical statistics steven stochastic street structure style sufficiency supervised sussex svens systems technology tenenbaum teuvo theory thesis thomas thrun tipping todd topographic torgerson touretzky transactions tresp university unsupervised using valencia valpola variable variational vector virginia visualisation volker volume wake warren weiss west wiley williams with yair york zinnes zoran http://www.jmlr.org/papers/volume2/downs01a/downs01a.pdf 333 A Divisive Information-Theoretic Feature Clustering Algorithm for Text Classification aaai about agrawal algorithm algorithms allerton american amsterdam analysis annual ansi appear applications approach arithmetic athens augmented august automated baker based bayes bayesian becher bekkerman bell berkhin bethesda bialek bias binary bled book boser bottleneck bradley carnegie categorization center chakrabarti chen cikm classification classifier classifiers classifying clustering clusters colloquium colt communication company comparative comparison complexity computation computer computing concept conditions conf conference content control cover curse data databases decompositions deerwester development dhillon dimensional dimensionality discovery discriminants distributional divergence document documents domingos dumais ecir ecml edition elements english entropy equivalence european event every examination feature features filter finding floating fourteenth francisco friedman furnas garey generalized global goldberg gray greece groups guan guyon harshman heckerman heidelberg hierarchical hierarchically high hill hofmann http icml ieee illc indexing inductive inform information inquiry institute international introduction iterative january joachims john johnson journal june kaufman kaufmann know knowledge kogan koller kullback landauer lang language large latent learning lecture leibler lloyd local logic loss machine machines management mangasarian many margin master math mathematical mccallum mcgill mcgraw means measures meeting mellon method methods mining mitchell model modeling models modern modha morgan naive nature navigating netherlands netnews neuhoff news nigam notes number optimal optimality optimization pages pazzani pedersen pereira plane platt point power press probabilistic problem proc proceedings quantization raghavan references relations relevant report representations research retrieval rissanen sahami salton science scientific scientist search second selection semantic september series shannon should siam sigir signatures simple singapore sixteenth slonim slovenia society sons spangler sparse springer standard stat statistical stochastic study sufficiency support surveys system taxonomy technical text theoretic theory thesis thomas tishby toolkit training trans under university unsupervised using vaithyanathan vapnik variance vector verbeek verlag very vldb weeder weighting what wiley winter with witsenhausen word words workshop world yang yaniv york zero http://www.jmlr.org/papers/volume6/elidan05a/elidan05a.pdf 34 Learning Hidden Variable Networks: The Information Bottleneck Approach aaai adachi advances agglomerative algorithm allerton analysis annealing annual applying approach approximations artificial association autoclass bacchus based bayes bayesian becker belief bialek biology bishop blackwell bonissone botstein bottleneck boyen breese brown cambridge carmel causation cell cells chain chang changes cheeseman chickering class classes classification cluster clustering coarsening combination combinatorial communication complex compression computation computational computer conference continuation control convergent cooper corduneanu cover cowan darwich data decisions degroot dempster department dependency deterministic dietterich dimensionality discovering discrete distributional dordrecht dynamic editor editors efficient eighteenth eisen eisner elements elidan england english environmental equivalence escaping estimating expectation exponential expression expressive factor fifteenth fifth fisher fourteenth francisco freeman friedman from fung gasch geiger gelatt genomic ghahramani giles glasgow globally glover glymour graphical graphs hajek hanson harel hasegawa heckerman henrion heterogeneous heuristic hidden hierarchical hill hinton homotopies horvitz ieee illinois incomplete incorporating incremental induction information institute intelligence intelligent international introduction jaakkola jack japan jaynes jensen john jordan jose journal justifies kanal kaufmann kearns kelly kingdom kirkpatrick kluwer knowledge koller labeled laguna laird language laskey latent lauritzen learning lecture leen lemmer likelihood linear linguistics local lopez lotner machine mantaras marginal markov markovian martin mass mateo mathematics maxima maximization maximum mcgraw mechanics meek meeting meila menlo merging method methods missing mixing mixtures model models modern molecular molphy moral morgan mosenzon muller multivariate naive nakano national neal netherlands network networks neural ninio notes number obermayer omohundro optimal optimization other oxford pages park partially pattern pearl pereira perturbation phylogenetics physical pittsburgh poole prade prediction presence press principle probabilistic probability problems proc processing program programming programs publishing reasoning recognition recursive reeves references refinement regression regularization related relaxation report research response review rose royal rubin saul scheines schuurmans science scientific score search self seventeenth shanoy simulated sixth slonim smith society solla sons sources spellman spirtes springer sreenivas statistical statistics stephen stolcke storz structure structures stutz system systems szummer tabu taylor tech technical techniques tenth that theory thiesson thirteenth thomas thrun tishby titterington tokyo tresp tutorial twelfth ueda uncertainty united university unsupervised values vanlehn variable variables variants variational vecchi verlag version view virginia volume watson weiss whiley wiley with words workshop yeast york zhang http://www.jmlr.org/papers/volume4/dybowski03a/dybowski03a.pdf 267 On Using Extended Statistical Queries to Avoid Memb ership Queries academic adam advances agnostic algorithm algorithmica algorithms amplification analysis angluin annual application aslam attribute attributes avoid avrim blum boolean boosting boston bounded bounds bshouty characterizing christino circuits clara classification colt commmunications computation computational computer computing concept conference constant core corrupted dana data decatur decision dennis depth distribution distributions dmitry efficient eighth eurocolt european exact examples extended eyal feldman fifth first fixed forth foundations fourier freund from functions furst gavinsky general generalization gennaro george goldman goldreich good hard help hypothesis identification israel jackson javed jeffrey jority journal kalai kearns kharitonov klivans kluwer kushilevitz laird learnability learnable learning leonid leslie levin linda line linial machine mansour membership merrick michael model monotone nader nathan neural nisan noam noise noisy oded optimal orlitsky pages parity philip phillip points poly predicate press problem proceedings publishers queries query random references relation respect robert rocco rosario roychodhury rudich sally schapire science sciences scott second sellie servedio sets shakelford shamir shapire shsh shwartzman siam sloan spectrum statistical steven strength symposion symposium system systems tamon theoretic theoretical theory third tolerant tolerate toward transform trees twenty types uniform using valiant vitaly volper wasserman weak weakly when with workshop yishay yoav http://www.jmlr.org/papers/volume4/druzdzel03a/druzdzel03a.pdf 268 The Learning-Curve Sampling Metho d Applied to Mo del-Based Clustering aaai advances algorithm algorithms among analysis application applications applied august autoclass basford bayesian benefit cambridge cheeseman chickering clarendon classification clustering computationally computer conference cooper cost curves data decision dekker dempster department discovery discussion distributions domingos dynamic editors efficient eighteenth fayyad federation fifth finite fourth from general graphical heckerman herskovits hertz howard hulten illinois incomplete induction inference international jensen john journal kadie kaufmann know laird langley learning ledge likelihood machine macmillan makov marcel mateo maximum mclachlan meek meeting melese menlo method methods mining mixture mixtures models morgan networks oates operational oxford pages park pearce piatesky press probabilistic proceedings progressive provost references regression research results royal rubin sampling scaling science second seer selecting shapiro sixth smith smyth societies society sons speed static statistical statistics stutz theory thesis thiesson titterington university urbana uthurusamy valencia versus wiley with york http://www.jmlr.org/papers/volume7/collobert06a/collobert06a.pdf 165 Large Scale Transductive SVMs academic active advances algorithm algorithmes algorithms american analyse analysis annual application applications approches archive artificial association asymptotic backprop bartlett based baxter bayesian becker behaviors belkin benchmark bennett beyond bordes boser boston bottou bottoueffvc bousquet burges cambridge carefully categorization cccp chapelle classification classifiers cloud cluster clustering codes cohn coil collection collobert columbia comp component compression computational computer concave conference consistency convex convexity cortes cristianini csail curves data decoste demiriz density dependences derbeko dietterich dimension edition editors effective efficient elisseeff empirical ertekin estimation explicit fast febuary fields finite from functions fung gaussian generalization ghahramani global globale graepel grande guyon harmonic herbrich hofmann html http icann icml image improved inference information insa intelligence international jaakkola jmlr joachims jordan journal kearns keerthi kernel kernels kluwer krause labeled labelled lafferty large larson lawrence learning lecture lecun leslie leveraging lewis libary linear local locales machine machines making mangasarian manifold margin margins markov mason maximum meir method methods minimal missing modified more muller murase nature nayar nene networks neufeld neural neuroprose newton nips niyogi noble notes numerique numeriques obermayer object ohio online optimal optimisation optimization page pages papers partially pittsburgh platt point practical press principal procedure proceedings processes processing protein publishers random rangarajan references report research rose rouen saul scalability scale scholkopf schuurmans science second semi semisupervised separation sequential series shen simulations sindhwani singer sinz smola software solla solution springer state statistical statistics structure supervised support svms systems szummer technical tenth text theory thesis through thrun trade trading training transduction transductive tricks tseng twentieth univ unlabeled using vapnik variables vector verlag vishwanathan volume walks weiss weston with wong workshop yang yaniv york yuille zhang zhou zien http://www.jmlr.org/papers/volume3/bendavid02a/bendavid02a.pdf 311 The Representational Power of Discrete Bayesian Networks analysis applications artificial assumption augmented bayes bayesian beyond boolean brodley cambridge chevonenkis classification classifier classifiers complexity conditions conference convergence danyluk domains domingos duda editors eighteenth european events flach frequencies friedman function geiger geometric goldszmidt hart implications independence inference intelligence intelligent international interscience judea kaufmann kauhmann learnability learning ling machine morgan naive network networks nominal optimality pages paterson pattern pazzani pearl peot peter plausible press probabilistic probabilities probability proceedings properties publication raedt reasoning references relative richard scene simple springer systems their theory uncertainty uniform university vapnik wiley zhang http://www.jmlr.org/papers/volume4/haddawy03a/haddawy03a.pdf 269 Text Classification using String Kernels acquaintance adatron adavances advances algorithm alignment annual appear approximation automatic available bartlett based berlin boser burges california cambridge campbell categorization cavnar claire classification classifiers communications computational computer conference connection convolution cristianini cruz dellec department dietterich discrete document doktorarbeit dynamic editor editors elisseef equations euroe fast feature features friess from functions gaithersburg gram grams greedy guyon harman haussler holloway html http huffman ieee independent indexing information infromation input institute integral international introduction joachims july kauffman kaufmann kernel kernels knirsch langley language large learning leen line lodhi london machine machines making many margin maryland matrix mercer method methods mika model morgan muller munchen national nature negative networks neural nips nist nystrom oldenbourg optimal optimizing pages pean philosophical pittsburgh positive practical press procedure proceedings processing pubs ratsch references relevant report representation retrieval rouveirol royal salton santa scale scholkopf schuurmans science seeger seventeenth shavlik shawe simple smola society space sparse speed springer standards statistical string structures support system systems target taylor technical technische technology text their theory training transactions trec tresp tuebingen type ucsc universitat university using vapnik vector verlag voorhees watkins williams with wong workshop yang york http://www.jmlr.org/papers/volume4/page03a/page03a.pdf 263 Exact Simplification of Support Vector Solutions accuracy advances algebra applied artificial bari bengio burges collobert conference daniel decision edition editors fast hall improving incremental information intelligence international italy joint journal kernel large learning linear lkopf machine machines methods minimal neural noble optimization pages platt prentice press problems proceedings processing references regression research rules scale schoelkopf sequential simplified smola speed statistical stockholm sung support svmtorch syed systems theory training using vapnik vector wiley with workshop york http://www.jmlr.org/papers/volume6/ernst05a/ernst05a.pdf 48 Tree-Based Batch Mode Reinforcement Learning aaai acrobot adaptive advances aggregation algorithm algorithms allocation alstr appear application approximate approximately approximating approximation armand artificial asynchronous atkeson automatic automation average bagging bagnell baird balancing barto based belgium bellman bicycle boyan break breiman california cambridge carnegie classification classifiers closed coarse coding computation computational computing conference context continuous control cost croatia curse cybernetics data decision delayed department diego diettrich difference differences dimensionality discrete discretization drive dubrovnik dynamic econometrica editors efficient electric electrical england ensembles ernst european examples extending extremely feature fifteenth fitted forests francisco friedman from function gamberger game generalization geurts glavic glynn gordon heidelberg hernandez hidden horizon icml ieee ijcai infinity information intelligence intelligent international introduction ishii iteration iteratively jaakkola jeon jordan journal july june kaelbling kakade kalaba kaufman kaufmann kernel kotkin lagoudakis langford large lasserre lavra learning least leen lerma less leveraging liege littman loop luenberger machine march markov mathematical mccallum mdps mellon method methods modern moore morgan multidimensional near nearest neighbors networks neural newyork nineteenth olsen online optimal optimization ormoneit pages parr parti pegasus perception pittsburgh policy polynomial pomdps power practical predict predictor predictors press prieditis princeton prioritized problems proceedings processes processing programming publishers randl random randomization randomized real reducing references regression reinforcement report research residual resolution rewards robotics robots rochester russell rust safely sato scale schneider search selective september shaping singh sixteenth smart soft solutions some space spaces sparse spong springer squares stable state statistical statistics step stochastic stockholm stone stuart submitted successful survey sutton sweden sweeping swing system systems technical technique temporal tesauro theory thesis time todorovski touretzky transactions tree trees tsitsiklis twelfth uncertainty university update using uther value variable vector veloso verlag wadsworth wang watkins wehenkel wiley wisconsin with workshop york yoshimoto zadrozny http://www.jmlr.org/papers/volume5/greensmith04a/greensmith04a.pdf 28 Variance Reduction Techniques for Gradient Estimates in Reinforcement Learning aberdeen acting action actor adaptive advances algorithm algorithmic algorithms analysis anderson annals applications applied approximate approximating approximation artificial ascent australia australian automatic autonomous baird bartlett barto based baxter bounds bradtke cambridge carlo cassandra chains chapman climbing cohn communications comparison complex computer concepts conference connectionist continuous control critic cybernetics dayan decision descent deterministic difference differences difficult discounted discrete discrimination divergence domains doob dynamic ecuyer editor editors elements eligibility engineering estimation evans experiments february fisher fishman following from function functions general glynn gradient graduate greensmith grimmett hall hammersley handscomb hill horizon ieee imperfect inequalities infinite information integrals intelligence intelligent international introduction jaakkola john jordan journal kaelbling kaufmann kearns kimura kobayashi konda learning least leen likelihood linear littman lovejoy machine mansour marbach markov markovian mathematical mathematics matrices mcallester measure measures method methods miyazaki models monte morgan muller national negative neural neuronlike number observable observed operations optimization optimize oxford pages partially path planning policy pomdps predict press probability problems proceedings processes processing programming puterman random ratio ratios recursions references regenerative reiman reinforcement related report research reward rubinstein sample school science sciences score seneta sensitivity series siam simple simulation simulations singh single solla solve solving some sons springer squares state statistical statistics stirzaker stochastic summer survey sutton swartz system systems technical temporal tesauro texts that theory tops touretzky traces transactions tsitsiklis university using value variance verlag volume weaver weiss wiley williams with without yamamura york http://www.jmlr.org/papers/volume7/buehlmann06a/buehlmann06a.pdf 139 Sparse Boosting acad adaboost adaptive addison additive akaike algorithm algorithms amer analysis angle appear approach approximation arcing assoc bagging bayes bennett better blanchette boosting breast breiman bridging buhlmann cancer chapman classification classifiers clinical computation conf consistency criteria criterion data demiriz description detection dictionaries dimensional disc discussion dressman efron ensembles estimation experiments exploratory expression finite frequency freund friedman function games garrote gene generalized gradient greedy green hall hansen hastie high huang human hypothesis identification ieee imaging infinite information inst intern ishida johnstone kauffman lasso learning least lecture length linear logistic loss lugosi machine mallat margins marks matching math meinshausen methods minimum model models monograph morgan muller multivariate neural nevins nonnegative nonparametric normal notes olson onoda pages penalty predicting prediction predictor predictors principle proc profiles pursuits ratsch references regression regularized relaxation report risk roughness santa schapire selection series shrinkage signal silverman soft spaces spang sparse speed splines statist statistical status subset technical technometrics theory thirteenth tibshirani time trans tukey using vayatis view wesley west with workshop zhang zuzan http://www.jmlr.org/papers/volume5/zhang04b/zhang04b.pdf 19 Statistical Analysis of Some Multi-Category Large Margin Classification Methods adaboost adaptivity additive alamos aldebaro algorithmic algorithms american analysis annals appear application applied association bartlett based bayes behavior bell berkeley blanchard blanz boosting bounds bousquet burges california cambridge category cbms certain class classes classification classifiers clint colt communication complexity conference confidence consistency consistent convergence convex convexity covering crammer data defense department dependent desyatnikov discovery discrimination discussion empirical estimation extended fast friedman function gabor geer gilles given grace greedy hastie holloway http ieee ilya implementation improved information ingo isbn jiang john jordan journal kernel kernels klautau knowledge koby laboratory labs learning linear logistic losses lugosi machine machines mammen mannor massart mathematics mcauliffe meir methods microarray minimization mining models multi multicategory multiclass national nicolas number observational olivier other pages pascal performance predictions press private process processes publications radiance rate rated rates references regional regression regularized report research rifkin risk robert royal rule ryan satellite schapire scholkopf scovel series shen shie siam singer smola smooth sons sparseness spline springer statis statistical statistics statitics steinwart support talk technical theory tibshirani tong transactions tsybakov universally university using vaart vapnik vayatis vector verlag view wahba watkins weak wellner weston wiley with xiaotong yoram york yufeng zhang http://www.jmlr.org/papers/volume2/bshouty02a/bshouty02a.pdf 352 On Using Extended Statistical Queries to Avoid Memb ership Queries academic adam advances agnostic algorithm algorithmica algorithms amplification analysis angluin annual application aslam attribute attributes avoid avrim blum boolean boosting boston bounded bounds bshouty characterizing christino circuits clara classification colt commmunications computation computational computer computing concept conference constant core corrupted dana data decatur decision dennis depth distribution distributions dmitry efficient eighth eurocolt european exact examples extended eyal feldman fifth first fixed forth foundations fourier freund from functions furst gavinsky general generalization gennaro george goldman goldreich good hard help hypothesis identification israel jackson javed jeffrey jority journal kalai kearns kharitonov klivans kluwer kushilevitz laird learnability learnable learning leonid leslie levin linda line linial machine mansour membership merrick michael model monotone nader nathan neural nisan noam noise noisy oded optimal orlitsky pages parity philip phillip points poly predicate press problem proceedings publishers queries query random references relation respect robert rocco rosario roychodhury rudich sally schapire science sciences scott second sellie servedio sets shakelford shamir shapire shsh shwartzman siam sloan spectrum statistical steven strength symposion symposium system systems tamon theoretic theoretical theory third tolerant tolerate toward transform trees twenty types uniform using valiant vitaly volper wasserman weak weakly when with workshop yishay yoav http://www.jmlr.org/papers/volume8/tibshirani07a/tibshirani07a.pdf 226 Margin Trees for High-dimensional Classification acad advances aliferis alizadeh allen amer american analysis angelo antonescu application approach armitage artificial assoc association barnhill based bennett berthold biegel bioinformatics black blue boldrick botstein breiman brown brussels byrd califano cancer cancers carcinomas cell central centroids chan characterization chemosensitivity class classification clinically clustering coller comprehensive conference curran data davis decision diagnosis diagnostic diffuse dimitrov discriminant discriminative distinct editor eisen embryonal esann evaluation expression extraction facto feature friedman frierson functions gaasenbeek gene generalized gerald golub goumnerova greiner greve guyon hampton hardin hastie hierarchical hudsom human identification information institute international journal kaufmann kern khan ladanyi ladd lander lapp large lasso latulippe learning levy lewis loda loss lossos louis lymphoma machine machines margin marti mateo math maximizing mclaughlin medicine meltzer mesirov method methods microarray model molecular molecularly moore morgan moskaluk mukherjee multi multicategory multiclass multiple multiway munagala narasimhan natl nature nervous networks neural nips norm olshen olson outcome pages park peterson pnas poggio polytechnic pomeroy powell prediction press proc proceeding proceedings processing profiling programs pwellm quinlan radiance ramaswamy references regression reich reinhold rensselaer report research rifkin ringner rosenwal rosset saal sabet sapinoso satellite scherf schultz schwab selection series sherlock shrunken signature signatures slonim splits stanford statist statistical statistics statnikov staudt staunton stolovitzky stone structured sturla substypes support system systems tamayo technical theory tibshirani tran transcriptional tree trees troy tsamardinos tumor tumour types unbiased university using vanichsetakul vapnik variable vector verleysen vural wadsworth wahba watkins weinstein weisenburger welsh westermann weston wetmore wilson with yeang zagzag http://www.jmlr.org/papers/volume1/collobert01a/collobert01a.pdf 339 Overfitting in Making Comparisons Between Variable Selection Methods adaptive agent agents alex algorithms anal anil application applications approaches architectures artificial assessment automatic autonomous barcelona based brunswick canada chen classifiers columns comparison computer computers conf coronado data dataset dermatology detecting different direct discovery douglas driver duin dynamic ebaa editor effectiveness elias emotion evaluation explanations expression feature features floating fold from george green healey herbrich icml icpr ieee ijcai information inspection intell intelligence ionosphere irrelevant jack jain jana japan jennifer jianchang john joint josef jsffs jsfs karl kaufmann kittler klaus knowledge kohavi kudo learning letters lindeberg mach machine marill measurement method methods mineichi mining model montreal morgan mushroom nagoya neural nonparametric noordhoff novovicova offline online oscillating overfitting pacl pages pattern pavel performance petr pfleger physiological picard pinz problem proc processing programs pudil quality quinlan ralf receptors recognition references results reunanen review robert rosalind ross same sample schalkoff scheffer scheme search seattle select selecting selection siedlecki signal sijthoff sklansky small smartcar society sommerfield somol sonar sons space spain spambase spectf statistical stearns steel stephen stress structural subset systems table test that theory those tobias tony topology training trans unbiased using vision vyzas waveform whitney wiley wiltschi with wojciech workshop wrapper wrappers zongker http://www.jmlr.org/papers/volume5/sallans04a/sallans04a.pdf 12 Reinforcement Learning with Factored States and Actions aaai about abstract abstraction academic ackley acquire action actions adaptive addison advances algorithm algorithmic algorithms amherst analysis anderson annals annual approach approximate approximation articial artificial asme assap assignment athena automatically bacchus back baird barto base basis bayesian behavior belief bellman belmont bertsekas binary boltzman boltzmann boutilier brafman brain brown calculations camb cambridge carlo causation ceteris chain chapter chemical coarse coding cognition cognitive cohn compact complexity computation computational computer computing conditional conference connectionist continuous contrastive control convergence cooper cooperate coupled cowell credit cued cybernetics david dawid dean dearden decision decomposition department descent dietterich differences difficult dimensional directed disorder distributed distributions divergence duane dynamic dynamical editor editors elements engineering equation errors estimation european examples experiments expert experts explorations external factored fast filtering force foundations fourth freund function geman general generalization generalized gerald ghahramani gibbs givan goldszmidt gradient graphical guestrin hanson harmony hasselmo hauskrecht haussler hertz hidden hierarchical hierarchies hierarchy high hinton hoos howard hybrid icann ieee images incremental infeng inference information intelligence internal international introduction jaakkola january john jordan journal juang justifies kaelbling kalman kanazawa kaufmann kearns kennedy klopf kluwer knowledge koller krogh laboratory large lauritzen layer learning leen letters line linear lippmann littmann lovejoy machine machines macro macros magazine march markov massachusetts matching mateo mathematical maxq mcallester mcclelland mcgovern mdps mechanics memory method methods metropolis meuleau michael microstructure minimizing model models monte moody moore morgan mozer muller multiagent nature neal networks neural neuro neuronlike nips niranjan observable operations optimal options order other pages palmer parallel paribus parr partially pattern patterson pendleton persistence peshkin phil physics planning policies policy pomdps poole potts poupart precup predict prediction preference press princeton principles probabilistic probability problems proc proceedings process processes processing products programming propagating publishers rabiner reading reasoning references reinforcement reinforcementlearning relaxation relearning report representation representations research restoration results richard richardson rodriguez rosenbluth roweth rumelhart rummery russell sabes sallans santamaria saul science sciences scientific search sejnowski series seventeenth simplification singh single sixth smolensky solla solve solving some space spaces sparse spiegelhalter springer state statements states statistics step steven stochastic successful survey sutton systems szepasvfiri technical teller temporal temporally tenth tesauro that theoretical theory thesis thrun todd toronto touretzky training trans transactions transformations tsitsiklis uncertainty university unsupervised using utilities value variants variational vector vectors verlag very view volume weakly wesley williams with workshop wright http://www.jmlr.org/papers/volume4/kisilev03a/kisilev03a.pdf 248 Graph-Based Hierarchical Conceptual Clustering acquisition activity algorithm algorithms allerman analysis analyze appear application approach approaches arlington artificial authoritative autoclass available background ball baritchi based bayesian beckwith benefits boulton bunke carbonell chakrabarti challenge chameleon cheeseman classification cluster clustering company complexity computation computer computing concept concepts conceptual conference conjunctive cook cure data database databases decision department description discovering discovery discrete distributed djoko domain domains drawing dynamic earthquake efficient electronically empirical engineering environment everitt experience expert fellbaum fifth find fisher florida formation framework freeman from galal gene gibson gonzalez graph graphviz gross guha hierarchical holder http hyperlinked ieee incremental inexact inferring information informative inquiry institute intelligence intelligent international into introduction jonyer journal karypis kaufmann kelly kleinberg knowledge koutsofios kumar langley large lattice learning length letters lexical lexicography line link machine maglothin management markov masters matching measure michalski miller minimum mining mitchell modeling morgan ninth north observation parallel partitioning pattern patterns pazzani policy potential predictive principle proceedings project publishers publishing quinlan raghavan rajapolan rastogi recognition references regulatory research rissanen rivest scalable schalkoff science scientific self sequences shim siam sigmod software sons sources stanford statistical stepp stochastic structural structure structured study stutz subdue substructure symposium system systems taylor telecommunications texas theoretical thesis thompson through tioga tompkins tools toxicology transactions trees university unsupervised used using volume wallace wide wiley wordnet workshop world york http://www.jmlr.org/papers/volume5/huang04a/huang04a.pdf 20 The Minimum Error Minimax Probability Machine academic advances aerial algebra algorithm algorithms amai analysis application applications applied approach arcing artificial athena belitski belmont berlin bertsekas bertsimas bhattacharyya biased binford birth blake boston boyd breiman busse california cambridge chan changing chapman chow citeseer classification classifiers computer computing cone conference control convex craven dagm databases deco decoste department detection diagnosis dordrecht dori edition editors eighth entropy evaluation first fourth fractional freund gaussian ghaoui girosi global globally goodwin greedy grudic grzymala hall handbook heldermann horst html http huang icml ieee image images improved increasing independence industrial inequalities information insead intelligence interchageability interior international irvine jordan journal kernels keysers king kluwer lanckriet langley large learning lebret lecture letters line linear lobo locally london machine machines maloof margin martingales massachusetts math mathematical mathematics maximum medical merz methods minimax mlearn mlrepository models nature nemirovsky nesterov neural nevatia nips nonconvex nonlinear notes object obradovic operational optimal optimization order osuna pages pardalos pattern performance pixel point polynomial popescu press preterm probability proceedings processing programming protocol publishers recognition references report repository research robust rooftop rule sage schaible scholkopf science scientific second sensitivity serie series sigma smola society sparse springer springerverlag statistical statistics strengths strohmann studies support symposium systems technical teicher theoretic theory training trans twenty university vandenberghe vapnik vector vectorization verlag vision volume with yang york zeitschrift zhang http://www.jmlr.org/papers/volume7/zanni06a/zanni06a.pdf 157 Parallel Software for Training Large Scale Support Vector Machines on Multiprocessor Systems adam advances alex algorithm algorithms also analysis annual appear application applications artificial assumptions asymptotic athena available barzilai based belmont bengio bernard bernhard bertsekas birgin borwein boser bottou bounds boxconstrained bruce burges cambridge cascade chang chen cheung chih ching chung cjlin class classifier classifiers clint code collobert computation computational computer computing conference constrained convergence convex core cosatto cristianini csie cvpr data david decomposition department dept design detection dimitri donato dong dourdanovic edgar edition editor editors elmer engineering eric ernesto face fast federico fletcher forum freund gaetano generalized germany gilbert girosi gradient graf guide guyon hans haussler hong hsuen http hush ieee igor improved information intelligence interface international introduction isabelle ivor james jian joachims john jonathan jose journal keerthi kernel knoxville kovoor krzyzak kwok large laura lawrence learning lecture leipzig leon library libsvm linear linearly loqo lower luca machine machines making management marco marcos margin mario mart mathematical mathematics mathematik message method methods michael ming minimal mining minos mixture modified murtagh naina national nello networks neural nonlinear nonmonotone notes nova numerical numerische operation optimal optimization optimizer order osuna other pages palagi panos parallel pardalos passing pattern perner peter pittsburgh platt point polynomial practical press prloqo problems proceedings processing programming programs projected projection publisher quadratic raydan recent recognition references regression report research robert roger ronan rong rosenfeld ruggiero samy sathiya saul saunders scale scholkopf sciandrone science scientific scovel second selection sequential serafini sets shawe siam simple singly size smola society software solver sons spectral springer standard stanford statistical step strictly study subject suen supercomputing support svmlight svms svmtorch systems taipei taiwan taylor technical technique techniques tennesse theory thomas throstem time training transaction transactions trends trigiante tsang type university upper user using valeria vapnik variable vector version very vision vladimir volume weiss wiley without working workshop xiong yair york yoshua zanghirati zanni http://www.jmlr.org/papers/volume6/agarwal05a/agarwal05a.pdf 44 Generalization Bounds for the Area Under the ROC Curve academic advances agarwal akademie algorithm algorithms american analysis annals anthony appear area artificial average barbara bartlett based bayesian becker bernhard birnbaum boosting bootstrap bound boundaries bounded bounds bradley buck california cambridge chapman characteristic charles classification classifiers clifford cohen colin combinatorics combining compression conference confidence consistency convergence corinna cormen cortes crammer cruz curve dantzig data dependences detection deviation devroye dietterich differences discrete edition editor editors efficient efron egan empirical erich error estimation exponential formal foundations francisco freund functional gabor generalisation generalization geometry ghahramani graepel gyorfi hall hanley herbrich hill holden hugo inequalities information intelligence international intervals introduction iyer james john journal klaus klose kluwer koby koninklijke large laszlo lawrence learnability learning lectures lehmann leiserson littlestone lugosi machine manfred mann margin martin mathematical matousek mcdiarmid mcneil meaning mehryar method methods minimization model mohri monthly nato nederlandse networks neural nick nonparametric nonparametrics obermayer operating optimization order ordinal pages partition pattern peled peter power pranking precision prediction preferences press probabilistic problem proceedings processing publishers radiology ralf rank ranking ranks rate rayner receiver recognition references regression related relating report research retrieval rivest robert ronald rosset roth roussas saharon sample santa sariel saul schapire scholkopf sebastian second selection series shawe shivani sigir signal simon singer space springer statistic statistical statistics stein surveys systems taylor technical test theoretical theory things thomas thore thrun tibshirani topics under uniform university vapnik variance verlag vladimir volume warmuth weterschappen whitney wilcoxon william with workshop yoav yoram york zaragoza http://www.jmlr.org/papers/volume8/reisert07a/reisert07a.pdf 218 Learning Equivariant Functions with Matrix Valued Kernels algebras amodei analysis analytical applications arbitrary aronszajn averaging ballard banach based berlin boyd burbea burkhardt cambridge canterakis chua cliffs company computation computer conference control cristianini desoer detect digital digitale distinct dodd douglas englewood estimating estimation features filters fitting fogarty foundations from fuer function functions fundamentals gaal generalizing geometric group gruppen haar haasdonk hall harmonic harrison heidelberg hilbert hough ifac image information integral integration intelligent international invariance invariant invariantentheorie ivcnz jain jersey john journal kernel kernels keypoints kimeldorf kirwan kolloquiums learning lecture lenz linear london lowe masani mathematical mathematics maut methods micchelli miller minimale model modules multi multiresolution mumford nachbin neural nips nonlinear nostrand notes over pages pattern paulsen pitman polynominvarianten pontil pose prentice press princenton proceeding proceedings processing rabut radar recognition references reisert represenatation representations reproducing research results scale scandinavian schoelkopf schumaker schur schwerpunkt series shapes shawe siggelkow signal signalverarbeitung similarity smola some sonar sons space spaces spline springer surface systeme systems tagungsband task taylor tchebycheffian theoretical theory topics toronto trans transform ueber university valued vector verlag video vision vollstaendige volterra volumes vorlesungen vossen wahba weighted wiley with york zyklischen http://www.jmlr.org/papers/volume6/elisseeff05a/elisseeff05a.pdf 33 Stability of Randomized Learning Algorithms algorithm algorithmic almost andonova annals applications approximation artificial august bagging bounded bounds bousquet breiman cambridge canada check classifiers combinations combinatorics computation conditions conference cross devroye differences distribution ecai edmonton elisseeff error european everywhere evgeniou free function general generalization girosi gyorfi heuristics ieee information instability intelligence journal kearns kutin learn learning leave leaveone lugosi machine machines mathematics mcdiarmid method model mukherjee nature networks neural niyogi number pattern performance poggio pontil potential predictivity predictors press probabilistic proceedings recognition references research rifkin rules sanity selection simple springer stability stabilization stable statistical statistics survey theory transactions uncertainty university validation vapnik voting wagner wiley york http://www.jmlr.org/papers/volume3/rakotomamonjy03a/rakotomamonjy03a.pdf 282 Kernel Indep endent Comp onent Analysis acoustics adaptive adaptivity additive advances akaho algorithm algorithms alternating amari american analyse analysis annals applications approach approximation architecture architectures archive arias artificial association asymptotic attias bach baker baltimore based becker behavior bell belmont between bickel biometrika bischof blind borga boscolo bound breiman brezis buja cambridge canonical cardoso case certain chapman chebyshev cholesky cichocki clarendon classification classifiers comon component computation computations concept conference constrained constraints continuous contrasts cormen correlation correlations cover cristianini curves data deconvolution density dependent dietterich distribution durrett duxbury edelman editor editors effect efficient efficiently eigenvalue eigenvalues eighteenth elements equations estimating estimation examples extended extraction factor factorizations fast feature fine fixed fonctionel francisco friedman functional fyfe garat gaussian generalized geometry ghahramani girolami girosi golub gradient greedy hall harlow harmeling hasselmo hastie have herault high hopkins hotelling icann icassp icml ieee image imps independent infomax information input integral intel interior international introduction john johns jones jordan journal jutten karhunen kaufmann kawanabe kernel kernels kettenring klaassen knutsson kolmogorov kullback landelius langley larger latent learning least leen leiserson leurgans ligence ligent likelihood linear lkopf loan lodhi london longman machine machines masson mathematical matrix maximization maximum mechanics meeting melzer method methods mixed mixtures models modified morgan motomura moyeed mozer muller multidimensional multiple natural networks neural neuromimetic nonlin nonlinear note numerical nystr optimal optimization order orthogonal orthogonality oxford parametric paris part pham poggio point polynomials press probability problem proceedings processing programming providence psychometric quasi rank rational references regression regularization reiter relations remarks representations reproducing research rinen ritov rivest robust royal roychowdhury saitoh saunders scandinavian scheinberg scia scientific seeger sejnowski semantic semiparametric separation sets seventeenth several shannon shawe siam signal signals silverman smith smola society sons source sources spaces sparse speech speed springer squares statistical statistics string super systems szeg taylor technical tenth text than that theory third thomas through thus tibshirani tokyo touretzky training transactions transformations transmission treatment tree tresp uncertainty unit university unlikely using vapnik variables variance variates verlag vlassis watkins weber welling wellner when widom wiley williams with works workshop wright yang york ziehe http://www.jmlr.org/papers/volume6/khardon05a/khardon05a.pdf 78 Maximum Margin Algorithms with Boolean Kernels adatron advances algorithm algorithmic algorithms analysis annual anthony approximate approximating bartlett becker blum blumer boolean boser bound bshouty cambridge campbell characterizing circuits classes classification classifiers communications complexity computation computational computer computing concepts conference constant convergence cristianini data david decision department dependent depth dietterich disjoint distribution distributions domain editors efficiency efficient ehrenfeucht eiron embeddings equivalence estimating euclidean exact exclusion fast fifth formulae formulas forster fourier fourteenth fourth free friess from functions furst gentile geometry ghahramani guyon half hancock haussler hierarchies holloway ieee inclusion information international introduction jackson journal kearns kernel kernels khardon klivans kowalczyk kucera kushilevitz learn learnability learnable learning letters like limitations linial lnai london machine machines mansour marchetti margin margins maruoka matching maximal maximum membership minimization minsky mistake models monotone neural ninth nisan nlog occam online operations optimal over pages papert perceptron perceptrons polynomial press proc procedure proceedings processing product protassi quasi queries query razor references report research respect risk roth royal rudich sadohara sakai schmitt science sciences separating servedio shawe siam simon simple sixth smola spaccamela spaces spectrum springer statistical structural subexponential support suttorp symposium system systems tamon tarui taylor technical term theory third thirty time training transactions transform trees tsukiji twelfth twentysixth under uniform university using valiant vapnik vazirani vector verbeurgt versus visual warmuth watkins weakly williamson winnow with workshop http://www.jmlr.org/papers/volume7/hamerly06a/hamerly06a.pdf 115 Using Machine Learning to Guide Architecture Simulation aaai accelerate achlioptas acoustics adaptive advances akaike algorithms amir analysis anchors annavaram annnals annual applications architectural architecture architectures artificial august austin automatic automatica automatically basic behavior bellman berkeley best between biesbrouck block bouguet bradley burger calder california carlo center characterizing charney cheng classification clustering clusters code cohn colt comparing compilation compilers computer conf conference control correlation cross dasgupta data database davies december description designs diego dilation dimension dimensional discovery distance distribution dynamic early editors eeckhout efficient efrat elkan embedded encoding estimating estimation experiments explor extending farnstrom fast faster fayyad february fifth find foundations friendly fuzzy generation gersho gonzalez guide hamerly hankins heuristic hierarchical hierarchy high hochbaum icml identification ieee inductive indyk inequality initialization instrumentation intel intelligence intercluster international intervals isbn itanium iterative june kaeli kapoor karunanidhi knowledge kolluri large learning lecam length letters lewis location look machine macqueen madison march markers matching mathematical mathematics matrix maximum mcnames means methods microarchitecture minimize mining model modeling monte moore motivation multinomial multithreading multivariate newsl neyman nips november number observations october operations optimization page pages parallel partial patil pattern pelleg perelman performance periodic phase picking pinpointing pods points polito portions possible predictability prediction press princeton principles probability problem problems proceedings processes processing program programming programs projection projections provost proximity quantization rakvic ramamurthi random references refinement regular reina report representative research revisited rissanen rotated sair samet sampling sampson sanghai scalability scale scaling schoenmackers schwarz science search second selecting september sherwood shmoys shoham shortest sigact sigart sigkdd sigmod signal signatures simplescalar simpoint simulation simultaneous sixteenth smyth software some speech startup statistical statistically statistics strong structure structures support survey survive symposium systems technical techniques theoretical time tool tracking transactions transition triangle twelfth twentieth ucsd uncertainty uniprocessor university using valid validation variable varying vector version volume wisconsin with york http://www.jmlr.org/papers/volume3/kandola03a/kandola03a.pdf 288 ­MDPs: Learning in Varying Environments aamo aamodt abstract adaptive advances agent ahuja algorithms alsp amsterdam annals applications approximate approximation architecture article artifical artificial aspects asynchronous attila august automatic autonomous bari barto basc based behavior bellman between bolyai bounded bourgine boyan cambridge cesses chastic cimmer citeseer cognitive college comp computation computer computing conf conference connell consideration context continuous control controller convergence cowan dayan dean decision decomp delayed dels department dietterich difference discrete domains double doya dular dule dynamic dynamica dynamics dzielinski ector editor editors eecg ehavior eleventh elte endent endulum engineering ensatory eria erience european event examples experimental exploration feedback feudal files first fomin forthcoming framework fransisco from function furuta game games general generalized givan grid gross group gullapalli hagen harmelen hasselmo heger heuristics hierarchical hinton html http hunt hwang indirect information institute integrated intel intelligent international introduction inverse iteration iterative iwashiro jaakkola jersey john jordan journal kaelbling kalkkuhl kalm kaufmann king knowledge leach learning life ligence linear littman lorincz lrincz lyon machime machine maes mahadevan making markov markovian master mataric mateo mathematical mathematics maxq mdps ments metho monro morgan motion move mozer multi multiple national navigation networks neural neuro ninth nipg novemb olicy optimal oral orally oratory organizing osition page pages parameter patha people planning practice precup preliminary press princeton proceedings processes processing programming puterman real references reinforcement representing research resolution results rewards risk robbins robotic robust rozgonyi scales scaling scientific seattle self semi septemb sequential singap singh sons space state static statistics stochastic strategies sugahara survey surveys sutton swing systems szep szita takacs temp tenth tesauro theoretical theory thesis time toronto touretzky toward tsitsiklis twelfth university using utoronto value varela variable volume watkins when wiley with world yamakita york zbikowski zsef http://www.jmlr.org/papers/volume7/decampos06a/decampos06a.pdf 180 A Scoring Function for Learning Bayesian Networks based on Mutual Information and Conditional Independence Tests about abramson academic accuracy acid acyclic adaptive akaike alarm algorithm algorithms amsterdam analysis andersson annals anytime applications approach approximate approximating artificial attributes automatic available averaging bacchus background based bayesian being beinlich belief bell benedict best binder biometrika blake blanco blood bootstrap bouchon bouckaert bound branch brief brown buntine california campos case castellano castelo cattle causal causation centre chang characterization chavez cheng chickering chow classes classification classifiers colony combination communications comparing comparison complexity computation computational computer computing conditional conference connected consortium constructing construction contingence continuous control cooper craig criteria cross cybernetics dags dash data databases denmark dependence dept description different digraphs dimension directed discovery discrete discretization distributed distribution distributions dover driven druzdzel edition efficient eighteenth eleventh elvira entropy environment equivalence equivalent estimating estimation european evans evolutionary experimental expert explanation expressiveness ezawa fayyad feature fern fifteenth fifth first forecasting foulum fourteenth friedman from geiger genetic globally glymour goal goldszmidt good graphical graphs greiner guti hailfinder hall hastings heckerman herskovits hidden hill hogg holland html http huete humans hybrid identification ieee improved incomplete independence independency induction inequalities inference inferred information integral intelligence intelligent international interval introduction inza irani irvine iterated jensen john joint journal kanal kanazawa kaufmann kayaalp kelly knowledge kocka kohavi koller kuijpers kullback kutat larra laskey lauritzen learning lecture lemmer length leung levitt lewitt library limited local long look luna machine madigan magdalena majorization management manley markov marshall mateo mathematical mathware means medicine meek menieur merz method methodology methods metric metrics minimum mining mlearn mlrepository model modeling models modified monitoring morgan multi murga murphy myers ndez neighborhood network networks ninth north norton notes olkin optimization ordering orderings oriented pages papers parameter parameters parentage part partially pattern peacock pearl performance perlman pfleger physica pike plausible poza prediction preliminary prentice press principle principles priors probabilistic probabilities probability proceedings publication puerta rasmussen reasoning recursive references refinement relationships remark repository representation research restarts restricted results right rios risk rissanen rrez russell sahami scheines schwarz science sciences score scoring search searching second selection semantics seventeenth seventh severe shachter singh singly sixteenth sixth sized soft software space spaces sparse spirtes springer squared statistical statistics stochastic structure structures study subset suermondt survey suzuki symposium synthesis system systems tables techniques technologies telecommunications tests that theoretical theory thesis thirteenth tian tools transactions transformational trees twelfth typing uncertainty uniform university using utrecht validation valtorta valued variables verification verlag verma volinsky weather wermuth wiley winkler with wong workshop wrappers yager yang york yurramendi http://www.jmlr.org/papers/volume5/yu04a/yu04a.pdf 18 Efficient Feature Selection via Analysis of Relevance and Redundancy academic adaptive aisen algorithms almuallim analysis application applied approach artificial attributes automatic based bell blum boolean boston broderick brodley cambridge categorization chapman choi class classification clustering cohen comparative comparisions complete computer concepts conference consistency content continuous correlation customer dash data detection dietterich dimensional discovery discrete discretization edition efficient eighteenth elements eleventh elisseeff empirical enabling evolutionary examples extensive fast fayyad feature features filter flannery formalism forman fourteenth frank friedman genomic guyon hall hastie high hill hussain ieee image images implementations induction intelligence intelligent international interval introduction intrusion irani irrelevant isbn java jensen john joint jordan journal karp kaufmann kluwer knowledge kohavi koller kononenko langley learning lung machine many mcgraw menczer metrics microarray miller mining mitchell mitra morgan motoda multi multiple murthy networks neural nineteenth node numeric numerical optimal pages pattern pederson pfleger pracitcal presence press probabilistic problem proceedings programs publishers quinlan recipes redundancy references regression relevance relevant relief relieff research retention retrieval review rivest robnik sahami sampling scheuermann search second selection selective setiono seventeenth sigkdd sikonja similarity sixth solution springer statistical stolfo street study subset swets symposium technique techniques tenth teukolsky text theoretical thirteenth tibshirani tools toward training transactions twentieth university unsupervised using valued variable vetterling vision wang weng with witten wrappers xing yang http://www.jmlr.org/papers/volume8/rifkin07a/rifkin07a.pdf 220 Value Regularization and Fenchel Duality adaptive adrian advances again alessandro alex algorithms altun american analysis andreas andrei annual appearance applied approaches approximation argyriou aronszajn arsenin atomic baker bart bartlett basic basis bazaraa belkin bengio benjamin berlin bernhard between beyond books borwein brabanter cambridge carl cbms charles chen chicago chris christopher clarendon cloud clustering combinations computation computational computations computer conditionally conference constructive continuously convex corinna cortes cristianini darrell data david decomposition delalleau department distance distribution divergence donoho duality edition edward eigenmaps entropy equations equivalence estimation everything evgeniou examples extensions federico framework francois fresh from functions gaussian generalized geometric gert gestel ghaoui girosi grace hanif hastie haussler herbrich historical hybrid ieee industrial inference information institute integral international interpolation interscience introduction isomap jaakkola jean johan john jonathan joos jordan journal kaufmann kernel kernels lanckriet laurent learning least lewis look lwearning machine machines manifold manifolds margin marie massachusetts massimiliano mathematical mathematics matrices matrix matthias maximum mcallester method micchelli michael mikhail minimization miroslav models mokhtar moor morgan mukherjee nachman nello networks neural nicolas niyogi nonlinear norm numerical nystr observational olivier optimization ouimet pages paiement parametrized partha pascal pattern peter poggio point pontil posed positive press probabilistic problems proceedings processes processing programming pursuit rahimi rakhlin ralf rasmusen recht recognition references regional regression regularization report representer reproducing research riemannian rifkin robert rockafellar roger rosset roux ryan saharon sample saunders sayan scattered schapire scholkopf scientific scott seeger semi semidefinite september series shaobing shaw sherali shetty sindhwani smola society solutions sons sparse spectral speed spline splines springer squares stanford statistical statistics structured supervised support suykens systems taylor technical technology theodoros theorem theory thesis tibshirani tikhonov tomaso tommi tony transactions transductive treatment trevor tyrrell uncertainty unifying university unpublished using value vandewalle vapnik variables variational vasilii vecto vector verri video vikas vincent vision vladimir volume wahba washington wets wiley willams williams winston with world yasemin yoshua zhen http://www.jmlr.org/papers/volume3/scheffer02a/scheffer02a.pdf 303 Optimal Structure Identification With Greedy Search academy acyclic algorithm andersson annals artificial automation background ball bayes bayesian belief besnard bouckaert breese buntine carnegie causal causation characterization characterizing chickering choice classes cognitive combination compelled complete computer conference consistent construct cooper cpdag czech darwiche data department determining diagrams digraph digraphs each edge edges editors eighteenth either eleventh else engineering enumerating equivalence every exists explanation exponential extension families family figure finding fisher fourteenth friedman from geiger gillispie glymour goldszmidt goto graph graphical greedy guide hanks haughton heckerman henrion herskovits identification ieee immediate implementation incident inclusion induction inference influence information input institute intel into irrelevance jeffreys journal kanal kaufmann king know knowledge koka koller label labeled laboratory leads learning ledge lemmer lenz ligence ligent literature lowest machine madigan markov mateo meek mellon method microsoft model models moral morgan msrtr network networks optimal order ordered oriented output oxford pages parent partially pastime pearl perlman plausible prediction press probabilistic probability problem proceedings rational reasoning references report republic requisite research reversible scheines science sciences search selecting selection seventeenth shachter simple sixth spirtes springer statistical statistics stratified structure structures studeny such synthesis systems tarsi technical that theory there thesis transactions transformational tutorial ucla uncertainty university unknown using verlag verma which while with york http://www.jmlr.org/papers/volume7/caponnetto06a/caponnetto06a.pdf 194 Stability Properties of Empirical Risk Minimization over Donsker Classes advaced advances agreements algorithmic almost analysis annals annual appear applications approximately approximation artificial asymptotics bartlett bodies bounds bousquet cambridge cambrigde cambrige caponnetto central characterization chervonenkis classes clustering combinatorics complexities computational computer conditions conference consistency constructions convergence convex convexity cube david devroye difficulty distribution donsker dudley editors eiron elisseeff empirical error estimation events everywhere expansion fields francisco free frequences function functions gaussian geer general generalization gine gyorfi haar ieee image importance indexed inequalities information intelligence journal kaufmann koltchinskii komlos kutin learning limit local long loss lugosi machine major mathematics maximizing means mendelson method minimization morgan mukherjee nature necessary neural niyogi number oracle pages partial pattern performance philips poggio pollard potential predictivity press prob probabilistic probabilities probability proceedings process processes processing rademacher rakhlin random ratio real recognition references related relative research results rifkin risk root rudelson rules scandinavian sciences sections shawe singer springer squared stability statistical statistics strong studies sufficient system systems taylor their theorems theoretical theory transactions tusnady uncertainty uniform university vaart vapnik verlag vershynin wagner weak wellner williamson with york zinn http://www.jmlr.org/papers/volume6/chu05a/chu05a.pdf 65 Gaussian Processes for Ordinal Regression adaptive additive advances albert algorithm algorithms american amico analysis application applied approach approaches approximate approximations artificial association back backpropagation barber bartlett based basilico bayesian becker behavior bernhard beyond bhattacharyya biometrics bound boundaries bounds brascamp brunn byrd cambridge cancer carlo cbms cell cess chapman classes classification classifier classifiers clinical cohen collaborative committee compaq computation computing concave concavity conference constrained constraint content correlates crammer cristianini csato curve data december degroeve department design dietterich diffusion discussion domany eachmovie ecml edinburgh edition editors efficient equation error european expectation experts expression extensions fahrmeir family fast febbo filtering fitting fokoue framework frank functional functions fundamenta gaussian gene generalisation generalized genome ghahramani ghaoui golub graepel graphical hagan hall hasselmo hastie hemmen herbrich hofmann http ieee implementation improvements including inequalities inference infinite informaticae information informative institute intelligence international jaakkola jackson january johnson jordan journal july june kantoff keerthi kernel kernels klaus kramer labs ladd lanckriet lander large lawrence learning lecture leen leindler levin lieb likelihood limited linear loda london machine machines mackay manola march margin massachusetts mathematics matrix mccullagh memory methods minka minkowski mixtures modeling modelling models monte mozer muller multiclass multivariate murthy nature neal nelder networks neural nocedal notes november obermayer observational online opper optimal optimization order ordinal pages pattern peled pfahringer platt policy practical pranking pratt prediction prekopa press principle probabilistic proc proceedings process processes processing programming propagation prostate public rank ranking rasmussen references regional regression regularization renshaw report research richie robert ross roth royal sara saul schapire scholkopf schottky schulten science scientific sebastian seeger sellers semidefinite semiparametrically series shashua shevade siam simple singer singh smola snelson social society solla sparse spline springer springerverlag srebro statistical statistics structured support systems tamayo technical technology theorems theory thesis things thrun tibshirani todd toronto touretzky transactions trees tresp tutz twentieth unifying university using vapnik vector verlag volume wahba warped weighted widmer williams winther with workshop yahoo york zimak http://www.jmlr.org/papers/volume7/bach06b/bach06b.pdf 174 Learning Spectral Clustering, With Application To Speech Separation academic achan acoustics advances aided aistats algorithm algorithms american analysis application applications approach arabie arias artificial athena audio auditory bach background bathe belkin belongie berkeley bertsimas bias blatt blind body bottleneck bousquet bregman british brown calculus california cardie chan channel chung circuits classification clustering communication comon comparing comparison component computation computational computations computer conditions conference connections constrained constraints cooke cour cristianini cuts data database demixing department design dhillon differential ding discriminative disjoint distance domany duality eccv ecological econometrics edelman eecs efficient eigenvalues eigenvectors element ellis engineering equivalence estimation evaluating experimental extreme factorization feature finite fourier fowlkes frey from functions geometry gogin gold golub granular graph gray griffin grouping guan hall heckerman hertz hidden higgins higham hillel hopkins hubert human hyvarinen icassp iccv icml ieee ijcai image images independent inference inferring information initialization integrated intelligence international introduction jaakkola jang john johns joint jordan jourjine journal kamvar kandola karhunen kernel kibble kisilev klein knowledge kulis labeled language largest learning likelihood limits linear listeners loan longuet luxburg machine magnet magnus malik mallat manning markov markovian martin mathematical mathematics matrics matrix maximum means measuring meila method methods metric microphone minimizing mining mixtures model models modified moore morgan multiclass multinode multiple multiway music natural negative neudecker neural nonnegative normalized numerical nystr observational optimality optimization organization orthogonal orthogonality other overton partially partitioning partitions pattern pearlmutter penn perception perceptual phaseless pitch prentice press probabilistic problems proceedings processing programming proximity random ratio recognition references relations relaxation relocalisation report representation research rickard rogers roweis russell scene schlag schrodl science scientific scott segmentation segmentations segmented semidefinite separation seung several shawe shental short siam side signal signals simon single singular slonim smith society sons sound source sources sparse spatiotemporal spectral spectrograms speech spline state statistical statistics strathclyde structure sums symmetric systems szummer taylor technical texas theory time tishby tour tracking training transactions transform tsitsiklis typical unified unifying univ university using values vectors view vision wagstaff wahba walks washington wavelet weinshall weiss wiesman wiley wilson with womersley workshop xing yilmaz zeevi zibulevsky zien zomet http://www.jmlr.org/papers/volume7/taskar06a/taskar06a.pdf 163 Structured Prediction, Dual Extragradient and Bregman Projections advances agreement alan alessandro alexander algorithms alignment alignments alistair allerton altun american analysis anders andrew anguelov annual application applications applied approach approaches approximation artificial assignment association associative athena auction baldi bartlett based belmont belongie bengio berlin bertsekas binary bingsheng biological bond bottou boyd burges cambridge canada carlos catholic chandra chatalbashev chekuri cheng christopher classification clique collins combinatorial communication comparison complexity computational computer computing conditional conference connectivity contexts continuous control convex core cost curved cuts daphne data david dependencies dimitri dinkar discrete discriminative disulphide document drago dual durbin eddy editors edmonton efficiency ekonomika empirical energy entire estimation evaluation evgeny exact exercise experiments exponentiated extragradient extrapolation fast fernando fields finding first flow formulation foundations francesca franz functions generalized geneva geremey gradient graeme graph greig guerriero guestrin gupta haffner hastie hebert heitz herman hidden hinrich hofmann human hyper ieee images implementation improvements inequalites inequalities information intelligence interdependent international ioannis ising jaakkola jerrum jianlin jitendra joachims john jordan joseph journal julien kaufmann kernel khanna klein koller kolmogorov korpelevich krogh kumar labeling lacoste lafferty language large lazaros learning lecun leon leonid leslie liao lieven linear linguistics louvain machince machine machines malik manning margin mark markov martial martin matching matematicheskie mateo mathematics matusov maximum mcallester mccallum message method methods metody metric michael mihalcea minimal minimized mitchison model modeling models monotone morgan naor natural nesterov network networks neural nonlinear north object optimality optimization other output pages passing path paths patrick pattern paul pedersen perceptron pereira permanent peter philadelphia pierre platt points polyhedra polymenakos polynomial porteous posteriori prediction press probabilistic problem problems proceedings processing product programming projection puzicha quadratic rada ramin random recognition references regularization regularized related relaxation report representation research reweighted richard robert rosset royal saddle saharon sanjeev sanjiv scale scan schlkopf schrijver schutze science scientific sean segmentation segmenting seheult selection separable sequence sequential serge shape siam simon sinclair smola society solution solving some spaces spatial springer stanford statistical stephen structured support switzerland symmetric symposium systematic systems taskar technical technology test theoretical theory thesis thomas thorsten tibshirani time tommi tracking training transactions translation tree trees trevor tseng tsochantaridis twentieth twenty uncertainty university using valiant vancouver vandenberghe variational various vassil vector vision vladimir vullo wainwright washington what willsky with word xiaojin yasemin york yoshua yurii zabih zens zosin http://www.jmlr.org/papers/volume6/shani05a/shani05a.pdf 73 An MDP-Based Recommender System aaai acapulco accurate agents alberta algorithms american analysis annual antonio anytime application applications architecture arivind artificial association august bacchus balabanovic barto based behaviors bellman bergstorm better bohnenberger bonet book borchers boston boutilier breese brown burke california cambridge canada carolina case categorization chapel chen chickering claypool collaborative combining commerce communications computational computer conditionally conf conference content cooperative cross customer data dearden decision dependency differences digital dimensionality discovery documentation dynamic editor editors edmonton eighteenth eighth electronic eleventh elicitation empirical encyclopedia evidence experiments factored fast faster feedback fetching fifth filtering florida formulation fourth francisco freed full geffner germany goldszmidt good goodman gottlob grouplens grove heckerman herlocker heuristic hill howard hybrid iaai iacovou implicit independent indicators inference information innovative inteligence intelligence intelligent interaction interest interfaces international internet introduction isbn issue item iteration jameson joint joshi journal july kabanza kadie kaelbling karypis kaufmann kent kitts knowledge koller konstan language learning libraries library linguistics littman machine making march marina markov markovian martha massachusetts mdps meek meeting method methods mexico microsoft mining modeling mooney moore more morgan multimedia national netnews networks north open orlando over pages palmer parr personal planning plans policies policy pomdp portalnad posterios prediction predictive preference press princeton probabilities problems proceedings processes programming promotion publishers puterman recommendation recommendations recommender recommending reduction references reinforcement representations research resnick result rewarding rewards riedl riedle rounthwaite saarbrcken santa sarwar schafer score search sell sequences server shani shoham sigkdd significance sixteenth sixth slaney smoothing solution special state states statistical stochastic study suchak supplement supported survey sutton system systems techniques temporal tests texas text than theoretic thiebaux thirteenth thirty tunable uncertainity uncertainty united university user useradapted using varian visualization volume vrieze wased webkdd weights when wiley with work workshop yang york zhang zimdars http://www.jmlr.org/papers/volume8/audibert07a/audibert07a.pdf 235 Combining PAC-Bayesian and Generic Chaining Bounds adaptive advances aggregated aggregation akademie aleatoires algorithms american analysis annals annual applications approach approximations association asymptotic audibert available averaging banach bartlett bayesian berlin better bilites boucheron bounded bounds bousquet cambridge catoni chaining chervonenkis classification classifiers colt combinatorial combining complexities complexity computational concentration conference control convergence course december dembo density deviation devroye dudley ecole econ edinburgh editor empirical error esaim estimation estimators events fields flour frequencies from gaussian geer generalisation generic geometry german gine henri hoeffding http inequalities inequality inst italy journal jussieu koltchinskii laboratoire large learning least lect lecture limit lncs local lower lugosi majorizing massart math mathdoc mathematics mcallester measure measures mendelson methods minimization model modeles models moscow nauka notes optimal oracle pages panchenko paris pattern philips pisier poincare preprints press proba probab probabilistic probabilit probabilites probabilities probability proceedings process processes rademacher random recent recognition references regression related relative report risk russian saint seeger selection series sess sharp shawe some spaces sparse springer square stat statistical statistics stochastic structures sums survey symmetrization talagrand taylor technical techniques textes their theorems theorie theory thesis toulouse translation tscherwonenkis tsybakov uniform universites university upper vaart vapnik varena variables variance verlag volume wapnik weak wellner with without xxxiii york zeichenerkennung zeitouni zinn http://www.jmlr.org/papers/volume3/strens02a/strens02a.pdf 306 Cluster Ensembles ­ A Knowledge Reuse Framework for Combining Multiple Partitions aaai academic accumulation active advances afrl aggarwal agglomerative alamitos alexander algorithm algorithms alogorithm analysis analyzing annual antonio anuka applicability applications applied approach approaches architecture artificial association august austin automation back backpropagation baeza balanced baltimore bangalore barnett barthelemy base based basis baskets bayesian berlin beyond biology blum boley bollacker bradley branko cagliari cambridge canada caruana case categorization chakaravathy chameleon chan chapter cikm classification classifications classifier classifiers cluster clustering clusterings clusters collective colt combiners combining comp comparison completeness computation computational computer computers computing concept conf conference connection consensus constraints construction cover craig cross dasarathy data databases david decemb decision decomp density department design development dharmendra dhillon dietterich dimensional directions discovering discovery discrete dissimilaritybased distributed document domain doug downloadable drastal duin dynamic easier ecml ective edition editor editors edmonton effective effectiveness efficient eled elements elling ensemble ensembles erger ergraph eriments ersp ertext ervised ester estimate estimation etween evidence extended fast fayyad ferligo first fisher force forecasting forecasts frakes framework francisco fred freeman friedman from function fusion future garey generalized generation george ghani ghosh gini granger graphs gross guide hall hanson hasselmo hastings heidelb heidelberg hershb heterogeneous heuristic hierarchical high highdimensional hillol hipc html http huang hypertext icml icpr ieee ijcai impact inductive industrial information informs initial initialization integrating integration intel international intl intractability invited irregular isolation issue issues iterative jain january jarvis jersey johnson joseph journal joydeep july june kahng kannan kargupta karypis kavek kernel kernighan keynote kittler kluwer knoblock know knowledge kohonen kriegel krishnamoorthy krogh kumar kurt laclerc lans large later lattice lavra learn learning lecture ledge leen letters ligence ligent limited linear lippmann lnai lncs local lorien machine machinery mala management many maps market mathematical mathematics means measure measures mehrotra mercer merugu meta methods michael mining minnesota minton mission mitchell mobasher modeling models modha molecular monjardet mooney moore mosenzon mozer multi multiclassifier multilevel multiple multivariate muslea nada natural nearest neighb netlist nets network networks neumann neural next nigam nips noise norton norwell notes novemb oint oints onent optimization oratory order ordered organizing ositions osteriori ottleneck paclik page pages paral parallel park partitioning partitionings partitions patrick pattern pekalska philip phylogenetic pratt prentice press principal probabilities problem problems proc procedure proceedings processing prodromidis prospects publications quality radial rates raymond recent recognition reduction references refinement refining reina relab related relatedness relationship research retrieval reuse rich richard rivest robust roli rule same sampling sander scalable scale scheme science scientific scott search searching sebastian second selective self semi sept sets seventeenth shared sharing sharkey shashi shekhar siam silver similarity simplification slonim society space spangler sparse spatial special springer statistics stemming steve stolfo strehl structure structures sullivan supervision support supra survey symposium systems talk task tasks technical technology tesauro teuvo text than theory thing thomas thrun time tishby touretzky toward training transactions transfer tree trees tumer tutorial twenty uncertainty under university unlab using utexas validation vedelsby verlag view viewp vipin visualization vlsi volume warnow wiley with workshop yates years yooseph zaki http://www.jmlr.org/papers/volume7/castelo06a/castelo06a.pdf 197 A Robust Procedure For Gaussian Graphical Model Search From Microarray Data With p Larger Than n acad academic academy addison algorithm alon american anal analysis application applications applied approach arabidopsis aracne article association associations aveiro basso bayes bayesian berger between bing bioch bioinformatics biology biometrics biometrika blanchette bleuler blocks breast buhlmann building butte califano cancer carter carvalho castelo cellular chain chapman chemotherapeutic chichester chklovskii clinical cnet coefficients coexpression complex computation concentration conditional context continuous correlation correlations covariance cowell data dawid decomposable decomposition definiteness dempster department dependencies design diestel dimensional discovering discovery discussion distribution dobra dressman drton dykstra edition edwards efficient elements empirical establishing estimating estimation experiments expert exploring expression favera fiehn first fitting friedman from fuente functional furholz gaussian gene genetic genetics genome genomic genomics golub graph graphical graphs gruissem hall hans heath heidelberg hennig high hoeschele huang human hyper hypotheses implications independence inference inferring institute integrated intelligent interaction internet interpretation interpreting introduction inverse ishida isoprenoid iterative itzkovitz jones kashtan kaufmann kluwer kohane kohn kurths large laule lauritzen learning lehmann linear london lotem magwene mammalian margalit margolin marks massachusetts mateo math matrix mcclure meaningful mendes metabolomic microarrays milo model modeling modelling models molecular morgan motifs mult multivariate national natl nemenman network networks nevings nevins observing olson order oxford partial pearl perlman physics pinter positive predicting prelic press priors probabilistic proc proceeding proceedings profiles protein publisher publishers reading reasoning reconstruction reference references regulation regulatory relationships relevance report represented richardson rohr rosenberg roverato sample sattath scale scand schafer science sciences selection separators shen shrinkage simple slonim spang sparse spiegelhalter springer statist statistical statistics status steuer stochastic stolovitzky strimmer suppl susceptibility systems tamayo technical testing thaliana theory thiele trans transcription undirected university using verlag vranova washington weckwerth wermuth wesley west whittaker wiggins wiley wille wishart with wong yang yeger york zeros zimmermann zitzler zuzan http://www.jmlr.org/papers/volume7/wright06a/wright06a.pdf 203 On Inferring Application Protocol Behaviors in Encrypted Network Traffic abstract accurate active addison advanced akodkenou alexander algorithm alignment american analysis analytic analyze anders andre andrew anja annals annual anomaly anthony anuual appear application applications approach april architectures area arrivals association asymptotically atkinson attacks august augustin authentication back backbone based baum bayesian behavior behavioral bernaille bernard bhattacharya bioinformatics biology bittorrent blinc boleslaw bounds bram branch breimer brodley broido brown browsing brunskill build building cambridge carla catherine chain chains characteristics charles christine claffy classification classify cleansing clustering codes cohen communication communications computational computer conference connection connections consensus convolutional correlation coull course critique cunningham daniel dark darpa data david dawn december decoding delay denis derived designing detecting detection device discrimination dong doob doors douglas dreger duane durbin dynamic early economics eddy edward emprically encapsulating encrypted encryption eric error esorics etoh european evaluating evaluation evaluations exposition expression extended fabian faloutsos faxon february federal feldmann felix felten finding fingerprinting fips flow flows fried functions gates gene george gerald graeme graf haines hall haussler hidden hiroaki holger identification ieee incentives information intelligent inter international internet intruders intrusion isaac ismael james january joel john joseph joshua journal july june karagiannis kave kendall kent keystrokes kimmen king kohno konstantina kristopher krogh kuai kullback kunikazu laboratory laurent layer learning leibler leonard lili lincoln line lippmann login lorier machine marc march mark markov martingales masson mathematical maximization maximum mcclung mcgregor mchugh measurement measures mian michael michalis mining mitchison modeling models molecular monrose moore motifs multilevel multiple network networking norman november occurring october optimum over packet padmanabhan pages papagiannaki park passive paxson payload peer performed perry perspective petrie physical preparation press privacy probabilistic probability proceedings processes processing profiles profiling protein protocol protocols qixiang reeves references remote renata rescorla research review richard rigsby robert robin robustness rosenberg russell saira salamatian schliep schneider schonhuth scott sean secure security sequence server seth sigcomm sigmetrics simon sjolander solomon sommer song soule soules standard standards statistical statistics steinhoff stephen stepping steve stochastic stones streaming sufficiency supplement supratik survivability symposium system systems szymanski tadayoshi tatu technique techniques technologies teixeira testing theoretic theory third thomas through tian time timing towards tracing traffic transactions university usenix using venkata vern visual visualization viterbi wagner wang weber webster wegman weiss wenke wesley wide wiley wileyinterscience will williams willinger with workshop wright wyschogrod xiang xinyuan xuqing ylonen yoda zhang zissmann zuev http://www.jmlr.org/papers/volume2/cannon02a/cannon02a.pdf 351 Machine Learning with Data Dep endent Hyp othesis Classes abdallah ability academic advances algorithms also ambartzumian analysis annals anthony apparently appeared applications applied approximation ariola artificial automatic bartlett based berkeley boucheron bound bounded bounding buescher calculus california cambridge cannon canonical catch central certain characterizing choice class classification classifiers colt company complexity computational computer computers concerning control convex cortes cover cowen cristianini data densest dependent devinney devroye dietterich digraphs dimension dimensional discussion disjoint distances distribution domination dorato dorit dudley editors electronic error estimates estimation ettinger factorization forthcoming freund generalization geometric geometrical gyorfi hard have hemisphere hierarchies high hochbaum householder hulls hush ieee improved inequalities inequality information intel international introduction john johnson journal kaufmann kernel koltchinskii kumar learning leen letters ligence limit linear lugosi machine machines marchette margin massart mathematics methods minimization model morgan mutual networks neural neurocolt number other over panchenko paper part pattern penalties plane points preparata preprint presented press priebe probabilistic probability problem problems proceedings processing properties psychometrika publishing rademacher random real recognition references replacement report risk rogers sample samples sampling scale scholkopf science scovel selection self serfling series sharp shawe shiryaev shorter smooth soft sons springer statistical statistics structural structures study support symposium systems taylor technical terms that their theorems theoretical theory this transactions tresp trick uncertain uniform university vapnik vector verlag version wiley williamson with without workshop york young http://www.jmlr.org/papers/volume5/vovk04a/vovk04a.pdf 236 Learning with Mixtures of Trees aaai advances aistats algorithm algorithms analysis annals applications approach approximating artificial association attributes autoclass based bashford bayes bayesian bell benjamin biology bishop blake boutilier buntine cached cambridge camp cause cheeseman cheng chickering chow clarendon classfier classification classifiers clustering combination combines comp company complexity computation computational computer computing conditional conference constrained constructing context continous coop cormen cowell cummings data databases datasets dawid dayan dekker delve dempster density department directed discovery discrete discretization discriminant disser distributions doctoral does ecific efficient elief ellis endence endences endency engineering entropy erties erts ervised essen estimating estimators etition evaluation expert face factor fayyad feret field fields finite fitting francisco fredman frey friedman from geiger gene genes getoor getting ghahramani goldszmidt good graph graphical guide hall hasselmo hastie heaps heckerman helsinki herskovits hidden hierarchical hinton hopkins horwood html http icml ieee ijcai improved incomplete incremental indep induction inference information inquiry institute intel international introduction issp jaakkola jacobs jelinek jersey joint jordan journal juan justifies kaufman kaufmann kearns kneser know knowledge koller kontkanen kustra laird language large larsen latent lauritzen learning ledge leimer leiserson ligence ligent likelihood lippmann literature machine machinery maclachlan management manual marcel markov massachusetts mateo mathematics maximum mean meek meila menlo merz method methodology methods michie microsoft mining missing mixture mixtures mlearn mlrep model modeling modelling models molecular monti moody moon moore morgan mozer multinets multiple myllymaki naive neal network networks neural noordewier onacci optimization ository other oxford parametric park pattern pearl pfeffer philips piatesky pittsburgh plausible predoviciu prentice press probabilistic probability proceedings processing produce prop propagation psychometrika publishers publishing puerto rasmussen rauss reasoning recognition recognize references relational repository representation research results revow rico rissanen river rivest rizvi royal rubin saddle saul science scientific sequences shafer shapiro shavlik shenoy similarity sixth sleep smyth society solla sparse speech spiegelhalter springer statistical statistics steitz stochastic structural structure structuring stutz sufficient systems tarjan tation taylor tech technology that thayer their theory thiesson tibshirani tibshrani tirri touretzky towell tractable training transactions tree trees uncertainty university unpublished unsup uses using uthurusamy utoronto variable variants view vision wake watson weiner west with workshop world york zemel http://www.jmlr.org/papers/volume6/almeida05a/almeida05a.pdf 71 Separating a Real-Life Nonlinear Image Mixture aapo acynips advances ahyvarin alberto algorithm algorithms allinson almeida almeidaica amari analyse analysis application applications approach architectures article artificial arxiv ations aussois based batzoglou bayesian bell biology blind brain brauer budillon burel cardoso carlos cichocki codes columbia comon component computation concept cond conference conserving context darmois david deco deconvolution decorrelation denoising docs dynamic editor editors ensemble estimating exisurl factorial faria faster first france fraunhofer generale generalization genome genomebiology geometric girolami granada grassberger haritopoulos harmeli harmeling harri hasselmo helsinki higher homepages hong honkela http hyvarinen iconip ieee image images independent information institut intelligence international islab japan jmlr journal jutten karhunen kawanabed kernel kong kraskov ktdsep lang lappalainen layer lbalmeida learning lecture liaisons life linear loubaton lparra machine maintz mark marques marquesalmeidaica mattera maximization media medical method michael microarrays minimization misep mixture mixtures mlica models mozer muller multi multilayer mutual nara networks neural newton nonlinear nonlineargeo notes number order pages pajunen palmieri papers parra pattern pearlmutter people perceptrons personal physical post predictability preserved press prieto proc proceedings processing public publish puntonet real references regensburg registration repulsion research results review revue riken salk schmidhuber sejnowski sensitive separating separation signal source sources space spain springer state statistical statistique stochastiques stogbauer survey systems taleb tence theis tony touretzky training transactions twan uniqueness unsupervised using valpola valpolanc verlag volume workshop worldscinet yang ziehe http://www.jmlr.org/papers/volume6/fiori05a/fiori05a.pdf 56 Quasi-Geodesic Neural Learning Algorithms Over the Orthogonal Group: A Tutorial adaptive akuzawa algorithm algorithmic algorithms amari american analysis applications applied based batch biometrika blind bodies body calculations california carlin carlo celledoni chain chains chemical cichocki complex component computation computational computing conditions conference constraints coupler deconvolution dept descent diego differential diffusion discussion distributions dynamics edition editorial efficiently entropy equations essay extensions factorization fast fiori fixed flow flows formulation forthcoming from fukumizu gallivan gelman general geodesic geometric geometrical gradient graduate granada grassman grenander group groups hastings hebbian higham hyvarinen ieee image images independent inefficiency inference information integration intelligence international introduction issue jensen john joint journal jump karhunen kass kernel keshava learning lecture letters linear machine machines magazine manifold markov martinez mathematics maximum mcmc method methods metropolis miller minimum modal models monte multi multichannel multivariate mustard mutual natural neal negative networks neural neurocomputing nishimori nonnegative normal notes numerical object olver online optimal optimization orthogonal orthogonality pages park pattern physics planning plumbley point practice proc proceedings processes processing quasi realization recognition reduced references report representation review rigid rosenbluth roundtable sampling second separation siam signal simulation sons spain special spectral springer springerverlag srivastava state statistical statistician statistics stiefel stochastic structures systems technical teller texts their theory training trans university unmixing using valued various vlsi warnes washington weight wiley wilson with works yang zhang http://www.jmlr.org/papers/volume7/MLSEC-intro06a/MLSEC-intro06a.pdf 199 Machine Learning for Computer Security advances analysis application applications barbara based behaviors berlin bratko brodley chan classify communications compression computer conf cormack data detect detection editor editors embedded encrypted executables filipic filtering fumera icdm ieee images inferring information intl into intrusion jajodia journal karp kluwer kolter kruegel kumar learning lippmann lncs lynam machine malicious maliciously maloof masson mining models monrose network newsome notes pages paragraph parthasarathy pillai protocol raid recent references research roli security signature song spam springer statistical text thwarting traffic training using verlag visualization wild workshop wright yurcik zamboni zupan http://www.jmlr.org/papers/volume7/scott06a/scott06a.pdf 127 Learning Minimum Volume Sets acta advanced advances alamos albertos algorithm algorithms amer anal analysis annual anomaly applications applied approach approximate approximation april asso assoc august baillo banach bartlett based belmont berlin blanchard boucheron bounds bousquet breiman brooks cannon casal class classification cluster cohen cole colt comp complexity computation computational concentration concentrations concept conference consistency contour convergence convex cover criteria cscott cuesta cuevas dahmen data daubechies david decision density dependencies detection developments devore devroye diego dimensional dimensions distribution distributions donoho duke durrett dyadic edges editors elements ellipsoid empirical encoding error estimate estimates estimating estimation exact examples excess finding flow fraiman framework friedman functions garcia generalized granulometric grove gyorfi harmonic hartigan heidelberg high howse http hush ieee image inform information intell introduction isopleth iterative john july kelly klemela koltchinskii kutalik laboratory langford lanl learning lectures ledoux letters level levels likelihood lindenbaum lugosi luxburg mach machine machines mass maximum measuring method minimax minimization minimum minmax mode model moguerza muller multimodality multivariate munoz national nearly neighbor network neural neyman nolan nonlinear nonparametric nowak numerica nunez olshen optimal oracle overview pacific pages paper paradigm paris patt pattern pearson penalized penalties platt plug polonik practical prediction prob probabilistic probability proc processes pubs quantile rademacher rates reasoning recent recognition references regions regression regularization related report research rice risk rodriguez rozenholc rtsch sager sawitzki schafer scholkopf scott scovel selection sept sets shawe singer smola smoothing some sons spaces spie springer stat statist statistical statistics steinwart stochastic stone structural submitted support talagrand taylor teacher technical tests their theory thomas trans tree trees trends tsybakov tutorial universit using uwmadison vapnik verlag vert volume wadsworth walther wavelets wedgelets wiley willett williamson with without wolkenhauer york zeger http://www.jmlr.org/papers/volume6/bongard05a/bongard05a.pdf 86 Active Coevolutionary Learning of Deterministic Finite Automata abbadingo active after algorithm algorithms alternatives amsterdam analogical analysis angluin annual applications approach artificial automata automated automating baram barzdin behavior berg bergadano bongard brave cambridge carr cellular change choice cicchello cliffs climbing coevolution cognitive colloquium committee company competition complete complexity computation computational computer conference construction control counterexamples data deterministic dfas driven dupont dynamic editors efficient encoding engineering englewood environmental evidence evolution evolutionary evolvable evolving examples experimentation failure feldman fifth finite first fourth from function garcia gecco generated genetic grammar grammars grammatical gunetti hall hamahashi hardware higuera hill holland honavar icgi identification identify ieee image incremental inducing induction inductive inference inferring information insights intelligence interactive international jonsson journal kauffman kaufmann kitano kremer labelling lang language languages learner learning lecture leucker lexicographically lipson ljung logic london lucas luke machine mahfoud mateo merging methods miclet minimal minimum morgan nasa needed network niching north note notes number oncina online opper ordered pages parekh pattern pearlmutter physical pitt polynomial porat positive prentice press price problem proceedings programming publishing queries query recognition recovery references regular report research results revisited reynolds robot saksena samples science scientific search second sempere sets seung slutzki smart society software solution sompolinsky space sparse springer state supplied survey syntactic synthesis system teacher technical theoretical theory thesis third time tomita trakhtenbrot transactions trials unanticipated universitet update uppsala urbana user using verlag vidal what with workshop world yaniv http://www.jmlr.org/papers/volume3/tsitsiklis02a/tsitsiklis02a.pdf 362 On the Convergence of Optimistic Policy Iteration algorithms approximation asynchronous athena barto belmont bertsekas cambridge computation computational conference convergence dayan delayed dynamic eligibility england eurocolt european fischer fourth from introduction iterative jaakkola jordan learning machine massachusetts neural neuro open pages press proceedings programming questions references reinforcement replacing rewards scientific simon singh springerverlag stochastic sutton theoretical theory thesis traces tsitsiklis umass university watkins with http://www.jmlr.org/papers/volume7/chang06a/chang06a.pdf 179 Adaptive Prototype Learning Algorithms: Theoretical and Experimental Studies adaptive advanced advances alamitos algorithm algorithms amsterdam analysis annals applications approximation architectures associated automation aviation bartlett based berlin between bezdek blake bottou bounds brighton building burges buzo california cambridge case chang chervonenkis classification classifier classifiers cluster clustering comparison computation computational computer computers condensed conf connection consistency consistent control convergence cortes cover cristianini dasarathy data databases denker density department design devi devroye digit discovery discrimination discriminatory distortion dreyfus drucker eckmiller editor editors equations equivalence error estimates estimation evaluation events exponential feature field fifth fogelman frequencies function functions fuzzy generalization girosi gray guyon handwriting hart hettich hodges holland html http hyperrectangle ieee image incremental induced information instance integral intelligence intern introduction irvine jackel john journal kernel klawonn knerr knowledge kohonen kruse krzyak large layer learning least lecun linde lkopf lloyd london lugosi machine machines margin martinez mean means medicine mellish memory mercer merz method methods minimization minimum mining mlearn mlrepository muller multiclass multimedia multivariate murty nature nearest negative neighbor network networks neural neurocomputing newman nonparametric north number objective ordered organization other pages pattern performance personnaz philosophical platt plenum positive ppner press probabilistic probabilities probability proc procedure processing project properties prototype quantization quantizer quantizing randolph recognition reduction references regression relative remote report repository revisited risk royal rule runkler sackinger salzberg sample school science selection self series shaw shawe simard single small smola society sons space sparse springer squares statistical statistics stepwise stone strong study support systems taylor technical technique techniques texas their theory training trans transactions type uniform university usaf vapnik vector verlag view wiley wilson with york zhao http://www.jmlr.org/papers/volume5/bhattacharyya04a/bhattacharyya04a.pdf 25 Second Order Cone Programming Formulations for Feature Selection advances algebra algorithms amaldi annals application applications applied approach approximability atomic basis bertsimas bhattacharyya bissell boyd canberra chebychev chen classification computer computing concentration cone cones connection convex data december decomposition dimensional donoho econ elisseff equations feature fletcher functions ghaoui grate high html http identification inequalities information insead integral interior john jordan journal kann kernel lanckriet learning lebret lecture linear lobo london lugosi machine marshall mathematical mathematics matlab measure mercer methods mian minimax minimizing models molecular molina multivariate negative nemirovskii nesterov neural nonzero norm notes number olkin optimal optimization order over pages philadelphia philosophical point popescu positive practical presented press probability processing profiling programming pursuit radisky references relations relevant report research rizki robust royal saira saunders scholkopf school science scientific second sedumi siam signal simultaneous society socp software sons spaces statistics studies sturm summer symmetric systems technical their theoretical theory tipping toolbox transactions type unsatisfied using vandenberghe variables weston wiley with york zero http://www.jmlr.org/papers/volume4/parra03a/parra03a.pdf 244 Bayes Point Machines ability accepted adaboost adaptive advances aitchison algebraic algorithm algorithms also american analysis annual anthony application applications applied approach approximate approximative artificial aspects automata autralia averaging barber bartlett based bayes bayesian berlin beyond billiards biometrika bishop boser bound brooklyn buntine burges california cambridge campbell candidate carlo categorization cbms chain clarendon classification classifier classifiers comparing compression computation computational computer computing conference connection continuous convergence convolutional cortes cristianini cruz data datadependent decoste density department dietterich direct discrete discussion distribution doktorarbeit download editor editors elimination equations esann estimating estimation european europhysics expectation expectations fast feature features feller field form frequency freund from function functions gaussian generalisation generalization generalize generalized gibbs girosi given glesner graepel graphical guyon haussler herbrich hierarchies hilbert http ieee ijcai inference information input institute integral intelligence international interpolation introduction invariant isbn joachims john joint jordan journal july kernel kernels kinzel kleinz knirsch laboratory labs large learnability learning leen letters linear littlestone lkopf ller location london machine machines mackay madison making many marchand margin margins markov massachusetts mathematical mathematics mcallester mean mechanics media memo mercer method methods micchelli mika minimal minimization minka mitchell mltr model models monte nature nchen neal negative nehl network networks neural novikoff obermayer observational oldenbourg onoda opper optimal optimization osuna over oxford page pages pami parameters pasadena pattern perceptron perceptrons performance philosophical physical physics pitman pittsburgh platt playing point polytechnic population positive practical prediction press probabilistic probabilities probability proceedings processes processing proofs propagation propulsion publication reasonable recognition references regional regions regression relating relevant report representer reproducing research review risk robust rochester royal rule rules rychetsky sampler santa scale schuurmans science search sequential series shawe slicing smola society soft solla sollich some sons space spaces sparsity spline springer statistical statistics structural structures supervised support svms sydney symposia symposium systems taylor technical technische technology test text their theorem theorems theory thesis tolerance toronto training transactions transduction tresp tsch type ucsc universit university using vapnik vector verlag version volume wahba warmuth watkin wiley williams williamson winther wisconsin with work workshop york http://www.jmlr.org/papers/volume8/sutton07a/sutton07a.pdf 228 Dynamic Conditional Random Fields: Factorized Probabilistic Models for Labeling and Segmenting Sequence Data aaai abbeel about abstract accelerated acoustics adam adaptive advances adwait alan algorithms altun american analysis andrew annotated annual antal appear applications approach approaches approximate approximations ariadna artificial association asynchronous audio automatic automation based bayesian beatrice belief bellare berger berkeley bernhard bilmes bosch bottou brill bruce buchholz building califf cambridge carlos carnegie carreira case causation chapter charles chris christopher chunking ciravegna collins comparison composition computation computational computer conditional conference conll constructing context contrastive convergence corpora corpus coupled crammer craven croft crusoe cycle cycles daphne darrell data david dayne dean decision decoding della dependencies dependency descent discriminative discriminatively distance divergence domains dynamic eccv edinburgh edit editors efficient elaine elidan emnlp empirical energy english entropy eric erik erikt estimation european experiments experts extract extraction fabio factorial factorized family fernando field fields fine finite foundations francisco free freeman freitag frietag from function gavin generalisation generalization generalized geoff geoffrey geometric georgios getoor ghahramani grammar graphical graphs guestrin hanna hebert hidden hierarchical hinrich hinton hmms hofmann horn html http human hung icml ieee ijcai image images induction inference informal information informed intelligence international introduction ioannis isbn iterative jaakkola jeff jensen john johnson joint jonathan jordan jorge journal july junhwan kanazawa kaufmann kedar keiji kevin khashayar khudanpur klein koby koller kudo kumar labeling labelling lafferty lance language large lawrence learning leon leonid liang linear linguistics lise loopy luhong machine machines mahadevan malouf manning marcinkiewicz marcus margin marios mark markov martial martin mary match mathematical matsumoto maximum mccallum mcdonald mceliece mcgraw measures meeting mehryar mellon message meta methods michael microsoft miguel minimization minimizing mining minka mitchell model modeling models mohri mooney morgan mukund multiscale murphy naacl naftali narasimhan national natural navigation nefian network networks neural nicol nips nition nocedal north note numerical object observable olkopf online optimization ostendorf pages papers parameter parsers parsing part partially paskin passing pattern paul penn perceptron pereira perpinian persistence peshkin pfeffer pieter pietra pinto policy press probabilistic probability proc proceeding proceedings processes processing products propagation propositional quattoni rabiner ramin ramshaw random ratnaparkhi raymond reasoning recog recognition references relational reparameterization report representation research residual richard riley robert robot robotics rohanimanesh rosenfeld roth rule rules ryan sabine sang sanjiv santorini saul scheduling schemes schmidt scholkopf schraudolph schutze sebastian segmentation segmenting selected semi sequence shai shallow shared shrinkage sigir signal singer single skounakis society soderland some soumya spatial speech springer sridhar srinivas stacked state statistical stephen stochastic string structured study support sutton svetha symposium systems table tagging taku task taskar technical technology text theocharous theory thesis third thomas thrun time tishby tjong tommi trained training transactions transducers transfer transformation tree treebank trevor tsochantaridis tutorial uncertainty unification university ural using variational vector venkatesh verlag very vincent viola vishwanathan vision visual wainwright wallach weighted weiss west william willsky with wolpert workshop wright xiaobo xiaoxiang xing xuming yair yasemin yedidia yoram york yuji zabih zemel zoubin http://www.jmlr.org/papers/volume7/belkin06a/belkin06a.pdf 188 Manifold Regularization: A Geometric Framework for Learning from Labeled and Unlabeled Examples academic academy adams adaptive advances again aistat algorithms alignment american analysis annual applications applied approaches aronszajn artif artificial audibert based basis belkin bengio bennett beyond biological birkhauser blum bousquet brand burges cambridge carmo categorization chapelle chapter chawla chicago chung classification cloud cluster clustering coifman colt combining computation computational conf conference consistency cont continuous corduneanu core cowell cucker cybernetics data definition delalleau demiriz density diffusion diffusions dimensionality discovery discrete documents dokl donoho dumais edinburgh editors efficient eigenmaps eleventh embedding everything evgeniou extensions extracting fields first foundation foundations framework fresh from function functions fung gaussian geometric geometry ghahramani given global graph graphs griffiths grigor grimes harmonic harmonics heat heckerman hein hessian highdimensional historical incorrectly induction inductive industrial inference information input inst institute intel intelligence international interscience isomap jaakkola joachims joint kandola kemp kernel kernels knowledge kondor kwok labeled lafferty lafon langford laplacian laplacians large learning linear local locally look luxburg machine machines maggioni management mangasarian manifold manifolds maps markov mass massachusets master math mathematical mathematics matveeva mccallum measure menlo methods mincuts mining mitchell models nadler national networks neural nigam ninth nips niyogi nonlinear nonparametric observational optimization other pages paiement parametric park partially partitioning philadelphia planck platt poggio point pointwise pontil posed press problems proc proceedings processing random reduction references regularization report representation representations reproducing riemannian rifkin roux roweis sahami sample saul scale scholkopf science sciences second seeger semi semisupervised separation seventh silva sindhwani sixteenth smale smola sobolev society software spaces spectral spline statistical statistics stromsten strong structure supervised support systems szummer task technical techniques technology tenenbaum tenth text theoretical theory thesis thrun tikhonov tool towards training transactions transductive transforms trees tsang twenty uncertainty univ university unlabeled using vapnik vector vert very vincent wahba walks warner weak weighted weston wiley with workshop yale yamanishi york zhou zien zucker http://www.jmlr.org/papers/volume5/hastie04a/hastie04a.pdf 24 The Entire Regularization Path for the Support Vector Machine acta active adaptation adaptive additive advances alessandro alex allgower alpha analysis angelo angle annals appear arrays author available bach bartlett based bernard beyond biostatistics brad cambridge cancer cauwenberghs chang chapman chapter chih christopher chung citeseer cjlin class classification classifiers component computation computational computer conference continuation csie curved data decoste decremental dennis density department diagnosis diehl discovery discussion distribution editors effect efficient efron elements engineering eugene evgeniou expression fine following francis friedman from functions gacv gene generalized georg gerald gert golub guide haifa hall hastie html http iain incas incremental independent inference information input international jerome joachims johnstone joint jordan journal katya kaufmann kernel kernels kiri knowledge kurt labs ladd lander large lasso latulippe learning least libsvm linear loda logistic loss machine machines making margin marron massimiliano mathematics matthias maximizing mesirov method methods michael microarrays mining models morgan mukherjee multi multiclass national nature networks neural nips norm number numerica observational optimization overview pages papers path paths pattern penalized philadelphia piecewise pnas poggio pontil postscript practical prediction press proceedings processing properties publishers quadratic ramaswamy recognition references regression regularization regularized reich report research rifkin ripley robert rosset royal saharon scale scheinberg scholkopf schuurmans science seeding seeger selection series seventeenth shai shrinkage siam sigkdd signature sixth smola society solution special spline springer stanford stat statistical statistics steve support svmlight systems taipei taiwan talk talks tamayo technical theodorus thorsten tibshirani tomaso tracking trevor tumor university using vapnik vector verlag verri vladimir volume wagstaff wahba watkins website weston williams with yeang york zhang http://www.jmlr.org/papers/volume3/dhillon03a/dhillon03a.pdf 278 Shallow Parsing with PoS Taggers and Linguistic Features aaai abney academic acai acquisition advances adwait algorithm algorithms american anno annual antal applied appraisal approach argamon argument arpa artificial assignment association astr automated automatic aycock base based berck bergen bias bies bled bosch brants brill britta buchholz bulgaria canada cardie carnegie cascaded chapter chark chunk chunking chunks church claire cognitive coling comparing compiling computational comunicacion conference conll copenhagen corpora corpus crete cuba cussens daelemans dagan data david denmark dept design disambiguation dordrecht driven dzeroski eacl editor editors eineborg ejerhed emnlp empirical entropy eric erik error esslli euroconference european exceptions extracting fast features ferguson florian forgetting fourth francois free general generator germany gillis grace grammar grammars grammatical greece grosjean gunnel halteren hans harmful http human identification induction inductive information intel internacional international introduction jakub james jciech john joint jorn june kallgren karen katz kaufmann kenneth kluwer krymolowsky lance lane language languages large learning ligence lindberg linguistic linguistics lisbon little logic machine macintyre magnus manuscript march marcinkiewicz marcus mark markov martin mary maximum meeting megyesi mellon memory methods miles mitchell model models montreal morgan morphosyntactic naacl national natural netherlands ngai nikola north norway notes noun osborne overview parser parsing part parts patterns paul pauw penn performance peter philadelphia phrasal phrase phrases pierce pittsburgh portugal predicate principle proceedings processing program programming project pruning psycholinguistic psychology publishers python radu ramshaw ranlp ratnaparkhi recent recognition references relation revised robert role rule saarbrucken sabine sang santiago schasberger seattle second september shallow shared shlomo sigdat simposio skut slovenia social some source speech statistical steven stochastic stockholm structure structures swedish syntactic syntax system tagger taggers tagging task tation technology text texts third thorsten tjong transformation treebank tzigov university unrestricted using veenstra very walter washington wennstedt with wordclass work workshop york yuval zavrel http://www.jmlr.org/papers/volume2/osborne02a/osborne02a.pdf 320 Ultraconservative Online Algorithms for Multiclass Problems abound adaptive adatron additive advances algorithm algorithmic allwein analysis anlauf annual approach approximate artificial attributes bakiri based biehl binary brain breiman brooks burges campbell classification classifiers clss codes computation computational computer conference convergence correcting cortes crammer cristianini design dietterich discriminant duda editors error europhysics exponentiated fast fifteenth foundations freund friedman friess general gentile gradient grove hart helmbold hill implementation information intelligence international irrelevant january jornal journal kaufmann kernel kivinen large learnability learning letters linear littlestone long machine machines margin maximal maximum mcgraw mesterharm methods minimal model morgan multi multiclass networks neural neurocomputing nilsson olshen online optimization organization output pattern perceptron platt prediction press probabilistic problems procedure proceedings processing programs psychological quinlan reducing references regression related relaxed reprinted research results review rosenblatt scene schapire scholkopf schuurmans sciences september sequential simple singer smola solving statistical stone storage support system systems theory thirteenth threshold trainable training trees ultraconservative unifying updates using vapnik vector versus wadsworth warmuth weak when wiley winnow york http://www.jmlr.org/papers/volume2/manevitz01a/manevitz01a.pdf 325 Kernel Methods for Relation Extraction aaai abney abound academic advances algorithm algorithms alignment analysis aone applications applied approach april artificial association attributes august based bednarski berger bernhard berwick bikel bioinformatics biological books brain butterworths cambridge cancer carol case categorization champaign chris chunks classification classifiers collins computational computer conditional conf conference convolution cortes cristianini crystal csli data dean della department description discrete dissimilarity dordrecht duda duffy duin durbin dynamic ecml eddy editor editors edits entropy european event expression extract extraction features fields francisco freitag freund furey generalized goldfarb graepel halverson hampton hart haussler herbrich hidden holloway horn iaai ieee ijcai illinois information innovative institute intelligence international introduction irrelevant jelinek joachims john johnson joint journal july kaufmann kernel kernelbased kernels kluwer krogh kruskal labeling lafferty language large learn learned learning learns linear linguistics littlestone lodhi london mach machine machines macromolecules many margin markov massachusetts matrix maximum mccallum mechanisms menlo message methods microarray miller mitchison models morgan munoz name national natural nebel networks neural neurodynamics nips northholland obermayer optimization other over paclik pages pairwise park parsing pattern pekalska perceptron perceptrons pereira pietra press principlebased principles probabilistic proc proceedings processing progress propositional proximity publishers pulications punyakanok quickly rabiner ramos ramshaw random ranking recognition rees references relation relational relevant report research retrieval richardella rijsbergen robert rosenblatt roth royal samples sankoff santacruz saunders scale scene schapire schummer schwartz science segmentation segmenting selected seqience sequence seventeenth shallow shawe sift spartan speech standars statistical steven stochastic stone string structures study support system systems tagging taylor technical technology tenny text that theory thomas threshold time tissue tutorial undertanding university urbana used using validation vapnik vector voted warps washington watkins weischedel what when wiley with york zelenko zimak http://www.jmlr.org/papers/volume7/langley06a/langley06a.pdf 120 Learning Recursive Control Programs from Problem Solving aaai acquiring acquisition action agent agents aips analogy analytical anatomy approach architecture artificial assignment automated automaten automatic automating autonomous avila bacchus barto based behavior benson berlin bernd biasing billing bonn both cambridge carbonell case chambery choi chunking clocksin cognitive combining competition complexity computational computers concepts concerning conference constrained construction control credit cumulative decomposition derivational detroit development diego discovery doing domain domains effect efficiency eleventh empirical engineering erol estlin evaluation exercises explanation expressivity feigenbaum feldman fern festschrift fifteenth foil fourteenth france francisco from functional general germany givan goal hendler heuristic hierarchical hill human improve incremental induction inductive intelligence international japan joint jones journal kaufman kaufmann khardon kibler klahr knowledge laird langley learning lernende logic lotem machine macro magazine mahr marsella mateo mcgraw mechanism mellish menlo method methods minton mitchell model models mooney morgan multi munich munoz nagoya national neches nejati newell nilsson nonterminal oldenbourg olsson operators ordered park paths persistent planner planning press problem proceedings production program programming programs quality quantitative random reactive reasoning recursive reddy reduction references reinforcement reprinted research results review rogers rosenbloom ruby rule rules sammut scheduling schmid schmidt seattle shapiro shavlik shop simon simple simulates sixteenth skills sleeman soar solution solving specific speed springer steppingstone stockholm storage strategies sutton symbolic system systems tadepalli technik teleoreactive tenth that third thirteenth thought transformation twelfth using utility utilization veloso verlag walks whistler wissenschaft with yoon york zelle http://www.jmlr.org/papers/volume5/wu04a/wu04a.pdf 9 Probability Estimates for Multi-class Classification by Pairwise Coupling advances algorithm algorithms allwein analysis annals annual another applications applied approach architectures artificial available average averaged bartlett based behavior bengio berechnung best binary blake boser bottou bradley breiman brier building california cambridge caption chang chemical citeseer cjlin class classification classifiers cliffs comparison compound computational computer conference consistency control convex cortes coupling cran csie culberson data database databases datasets december decision department division document dotted dreyfus duan each edition editor editors empirical engineering englewood ergebnisse error errors estimates exdb expressed feuston fifth figure five fogelman forecasts forest forests four friedman generalized gradient guyon haffner hall handwritten hastie html http hull hunter ieee information intel international irvine issn john journal keerthi knerr large layer learning lecun leen left letter liacc liaw library libsvm ligence likelihood line machine machines margin mathematische maximumproblem mechanical merz method methods michie minimization mlearn mlrepository mnist modeling models monthly multi multiclass national nerual network networks neural neurocomputing news note november optimal outputs pages pairwise papers pattern personnaz platt plattprob points poly polychotomous prentice presented press price probabilistic probabilities probability procedure proceedings processes processing project qsar random randomforest recognition reducing references refregier regression regularized report reports repository research review revisited right risk rnews ross satimage saul schapire scholkopf schuurmans science second segment sheridan shows singapore singer single small smola software solid sons spiegelhalter springer stanford stat statistical statistics statlog stepwise stochastic study support sventnik svmprob svms systems taiwan taylor technical terry tesauro test testing text theory thrun tibshirani tong tool touretzky training transactions turnier unifying university using usps vallet values vapnik vector verification verlag volume voting wahrscheinlichkeitsrechnung waveform weather weng which wiener wiley with workshop yann zeitschrift zermelo zhang http://www.jmlr.org/papers/volume8/melnik07a/melnik07a.pdf 232 Concave Learners for Rankboost addison advanced aggregation algorithm algorithms analysis annual application applications applied arrow aspects baird based berlin boosting buffalo bunch bunke chapter chellappa choice classifier classifiers combination combining commerce components computational computer computing conf conference confidence decisions department discriminant document dwork editors efficient elastic evaluation experiments face facerecognition fellous feret florida freund from government graph group harman heidelberg http hull ieee image individual information institute intelligence international intl introduction iyer journal kittler krishnaswamy kruger kumar learning lecture lectures leveraging machine malsburg matching mcgill meir melbourne melnik metasearch methodology methods mixed modern moon multiple naor national nist notes number office pages pattern perception performance phillips preference preferences principal printing proc proceedings publ publication rank ranks ratsch rauss recognition references renda research retrieval rizvi roli salton schapire science score singer sivakumar social special springer springerverlag srihari standards state straccia structured swets symposium systems technology tenth text theory thesis thirteenth trans transactions trec university values vardi verlag visual voorhees weng wesley wide wiley wiskott with word world yamamoto york zhang zhao http://www.jmlr.org/papers/volume7/shalev-shwartz06a/shalev-shwartz06a.pdf 161 Efficient Learning of Label Ranking by Soft Projections onto Polyhedra advances agarwal algorithmic algorithms analysis annual applications april artificial based bipartite boosting boundaries boyd burges cambridge category censor class classification classifiers clemenon clickthrough code cohen computation computational conference convex cormen cortes crammer criteria data decomposition dekel development discovery editors eighteenth elisseeff empirical engines erratum european family fast formal fung generalization graepel herbrich hildreth hull ibidem ieee implementation information intelligence interior international introduction joachims jornal journal kernel knowledge krishnapuram label labeled large learning leiserson linear logistics loqo lugosi machine machines mangasarian manning margin marging meet method methods middle minimal minimization mining models mohri mthods multi multiclass musicant naval networks neural niyogi obermayer online optimization optimizing order ordinal overrelaxation oxford pages parallel pattern platt point press procedure proceedings processing programming projecting quadratic quarterly rank ranking rankings recognition references regression research retrieval risk rivest rosales rudin schapire scholkopf schuurmans scoring search separation sept september sequential seventh sigir singer smola software sopopo stability statistical stopping successive support symposium systems theory things training transactions university using vandenberghe vanderbei vapnik vayatis vector watkins weston wiley york zenios http://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf 379 Latent Dirichlet Allocation abramowitz abuse adding advances aitr aking aldous algorithm along alternating american analysis annotated annual appendix applications apply approach appropriate approximate arbitrary arguments arises artificial ascent aspect association assumption attias avery baeza based bayes bayesian begin below berkeley berlin between bibliography blei bound bounding bounds caenorrhabditis cambridge carlin categorical censored center cgcbib chapman chichester class classification clustering cognitive collaborative comparison complexity component computation compute computed computer computing conditional conditionally conference consider considered constrained contain containing content converges coordinate corpus cubic data deerwester defined denote denotes dependence depends derivative derivatives derive deriving describe described diaconis diag diagonal dickey difference digamma dirichlet discussed discussion distribution distributions divergence division document documents dover dropped dumais each earlier easily ecole editor editors eighteenth elegans elements empirical entropy environments equal equation equivalent estimate estimates estimating estimation exactly exchangeability exchangeable expand expands expansions expectation expected experimental exponential exponentiating expression fact factor factorizations family filtering final finally find finding finds finetti first five fixed flour following form found framework from full function functions furnas furthermore gamma gelman general generated generative genetic ghahramani given gives gradient graphical griffiths hall hand handbook harman harshman have heckerman here hessian hierarchical higher hill hofmann http hypergeometric ijcai immediately improving independent indexing individual inequality inference information initialization intelligence international interpretations intractable introduction inversion invoke involves isolate isolating iterating iterative jaakkola jelinek jensen jiang joachims john jordan journal kadane kappen kass kernel labeled lafferty lagrange lagrangian landauer large latent lawrence learning left leisink lemma letting likelihood likelihoods linear lines london lower machine made making marginal mathematical matrix maximal maximization maximize maximizing maximum mccallum mcgill mcgraw meila method methods minimizing minka mitchell model modeling models modern moreover morris multi multinomial multiple multipliers multiplying must naive natural need neto neural newton newtonraphson next nigam normalization notation note notions observe obtain obtaining omitting only optimization order over overall overview oxford pages papadimitriou parameter parameterization parameters parametric particular pennock point popescul posterior practical presented press probabilistic probabilites probability problem procedure procedures proceedings processing progress propagation property provides raghavan raphson recall recent recognition recommendation references related remainder rennie repeatedly report representation reprint requires respect respectively restored retained retrieval ribeiro right ronning rubin saint salton same saul scalar scale scales science second section sections select semantic setting seventeenth several show shows side sigir simplicity simplifies simulation since single size smoothed society solve sons sparse special specific speech springer stationary statistcal statistic statistical statistics steffey stegun step stern steyvers structure subscript such sufficient summing support surrogate swmed systems take taking tamaki technical technique terms text that then theory therefore this those thrun thus time topic topics total translation trec true twenty uncertainty under ungar unified unique univ unlabeled until used useful uses using usually valencia value values variational vector vempala verified viewed where which wiley with word workshop write written xiii yates yields york zero http://www.jmlr.org/papers/volume7/olsson06a/olsson06a.pdf 195 Linear State-Space Models for Blind Source Separation acoustical acoustics adaptive advances aiaa algorithm allen america amplitude anal analysis anemuller annals appear application applications approach approaches approximate approximation array asme audio auto background based basic beamforming bell berkley bermond blind bound brandstein bucy cardoso cassiat channel cicaar comon component computation concept conference context convolutive convolved cosmic current data deconvolution decorrelation delabrouille delayed dempster denmark dynamic dyrholm efficiently engineering estimates eusipco excitation expectationconjugate explicit feder field filtering fmri forecasting from gaussian generalization ghahramani gradient hansen harmonic hyvarinen icassp ieee image imaging incomplete independent information international inverse john journal kalman karhunen kollmeier laird lambert lars learning lehn likelihood linear machine mathematical maximization maximum mcaulay mckeown mean method methods microphone microwave mixture mixtures model modeling models modulation monro more moulines multi multichannel neural neurobiology nielsen noise noisy nonlinear olsson opinion oppenheim optimization overrelaxed page pages parra patanchon pearlmutter petersen prediction proc proceedings processing quateri rauch references regressive rensen report representation results review robbins room roweis royal rubin salakhutdinov schi sejnowski sensitive sensors separation series shumway signal signals simulating sinusoidal small smoothing snoussi society sons souloumiac source sources space speech spence state stationary statistics stochastic stoffer striebel synthesis systems technical than theory time trans transactions tung ucminf unconstrained unifying university using volume weinstein what wiley winther with yellin http://www.jmlr.org/papers/volume8/loosli07a/loosli07a.pdf 214 Comments on the "Core Vector Machines: Fast SVM Training on Very Large Data Sets" advanced alexander algorithm algorithms applied based bernhard burges business chang chen cheung chih chung cjlin classifiers computation conference core csie data editors fast gloosli hsuen html http industry information insa international ivor james john journal kernel kernels kwok large learn learning library libsvm lkopf loosli mach machine machines matlab methods ming minimal models murty narasimha neural optimization order pages platt press proceedings references research rong rouen scholkopf second selection sequential sets simplesvm smola stochastic support theory toolbox training tsang tutorial twentieth using vector very vishwanathan with working http://www.jmlr.org/papers/volume6/kuss05a/kuss05a.pdf 87 Assessing Approximate Inference for Binary Gaussian Process Classification abrahamsen advances algorithms american analysis annealed approximate approximations association baltimore barber bartlett bayes bayesian becker berger bernardo blake bounds bousquet bridge cambridge carlo center chain chang chapelle choosing cjlin classification classifiers comparison computation computations computer computing constants correlation csato csie curve databases dawid department design duane edinburgh edition editors electrical engineering error expectation exponential factors families family fast field fields fitting from functions gaussian gelman generalisation ghahramani gibbs golub hagan handling herbrich hettich hopkins html http hybrid hyperparameters ieee importance inference information informative intelligence john journal kass kennedy kernels kuss large lawrence learning letters library libsvm loan machine machines mackay margin markov matrix mean meng merz methods minka mlearn mlrepository models monte mukherjee multiple neal networks neural newtorks normalizing norwegian note obermayer obtainable online opper optimal ordinal oslo oxford pages parameters path pattern pendleton physics platt prediction press priors probabilistic probabilities process processes processing propagation raftery random rasmussen recognition references regression report repository research review ripley roweth royal sampling scholkopf schuurmans science scientific second seeger series simulating smith smola society sparse springer statistical statistics strategies support systems technical theory thesis thrun toronto transactions tuebingen university using vapnik variational vector williams winther with york http://www.jmlr.org/papers/volume6/gunawardana05a/gunawardana05a.pdf 99 Convergence Theorems for Generalized Alternating Minimization Procedures academic accelerating adaptation algorithm algorithms alternating amari american annals applications approach artificial audio boltzman bound boyles byrne comments communication conference consistency convergence convergent convex cosem csiszar curvature darroch data databases decisions defining dempster digalakis distances distributions divergence editor efficiency efficient efron emission enee estimate estimation estimators european extensions fast fisher foundations from generalized geometry ghahramani gharamani gindi globally graphical gunawardana hall heckerman hidden hinton hopkins hsiao ieee image imaging improvement incomplete incremental information intelligence introduction issue iterative jaakkola johns jordan journal justifies khurd kluwer krishnan laird large learning lecture lehmann liepzig liese likelihood line linear machine markov maryland mathematical maximum mclachlan medical meek meilijson methods michael minimization models neal networks neural nonlinear note notes november number optimization order other pages philosophical prentice press probability problem problems procedures processing programming properties rangarajan ratcliff reconstruction references roweis royal rubin salakhutdinov saul scaling second series society speech spring statistical statistician statistics submitted supplemental technology terms teubner that theoretical theory thesis thiesson tomography training transactions tusnady uncertainty unified university using vajda varational variants verlagsgesellschaft view volume wald wiley with zangwill http://www.jmlr.org/papers/volume7/goldberg06a/goldberg06a.pdf 113 Some Discriminant-Based PAC Algorithms algorithmic allwein analysis angluin annual anthony approach attention based biggs binary bishop blum blumer boolean boosting bound cambridge class classification classifiers codes colt computation computational computer computing concepts conference correcting cristianini cryan dags dasgupta david denis dept dichterman dimension discrete distribution distributions domain duda edition efficient ehrenfeucht equivalence error estimates estimating evolutionary examples focus foundations free freund frieze from gaussians general gilleron goldberg guarantees guruswami hart haussler helmbold ieee information integer international introduction jerrum journal kannan kaufmann kearns label laird large lattices learnability learned learning letouzey letters linear littlestone lnai machine machines mansour margin markov mistake mixture mixtures model models morgan multiclass nature networks neural nips noise noisetolerance noisy occam over oxford pages palmer pattern platt polynomial positive press probabilistic proceedings processing product queries razor recognition reducing references report research restricted rubinfeld sahai scene schapire science sciences second sellie separating shawe siam singer sloan springer state statistical strength support symposium system taylor technical theoretical theory time tolerant tracts transformations trees unifying university unlabeled vapnik vapnikchervonenkis vector verlag volume warmuth warwick weak wiley with workshop york http://www.jmlr.org/papers/volume6/cowell05a/cowell05a.pdf 82 Local Propagation in Conditional Gaussian Bayesian Networks aalborg acyclic acyclicity algebraic algorithm algorithms american andersen annals annual application applications applied area artificial association associations bayesian becker between california cambridge cass chordality city clique close college company complete computation computations computing conditional conference cooper cowell cybernetics dawid decision decomposable decomposing decomposition degenerate denmark department diagrams directed discrete discussion distributions editors electronic elsevier expert fast fill finding formulation francisco gaussian geiger general genetic giving graph graphical graphs hypergraphs identification ieee inference influence information institute intelligence intelligent interaction jensen journal july junction kaufmann kenley kingdom kuijpers larranaga lauritzen lazy leimer linear local london madsen management marked mateo mathematics maximal means methodology methods minimum mixed mixtures models moral morgan mortera multistage murga networks olesen optimal oxford pages part peak pearl poza prime probabilistic probabilities probability problems proceedings propagation publishing qualitative quantitative raphael reduce reduction references report research royal rulff science selectively separation series shachter siam simple sixth small society some space spiegelhalter springer stable state statistical statistics structures subgraph sufficiently systems tarjan technical test their time total transactions trees triangulated triangulation uncertainty united university using variables variances verlag vertices wermuth which with yannakakis york http://www.jmlr.org/papers/volume2/pekalska01a/rev1/pekalska01ar1.pdf 329 Distributional Word Clusters vs. Words for Text Categorization aaai achieved addison advanced advances agglomerative algorithm algorithms allerton allwein american amita analysis annealing annual applied approach apte artificial assistee association athens automated automatic baeza baker bari based basili bekkerman berkeley bethesda beyond bialek binary blum boosting boser bottleneck burges cambridge caropreso categorization celine chapter chechik chemnitz chin cikm claire classification classifiers classify clustering clusters colloquium colt combining communication comparative compression computational computer computing conference confidence control cortes cover craven cristianini croft damerau darmstadt data databases deerwester dempster deterministic development dipasquo distributional document documents dortmund double dumais ecir ecml editor editors efficiently elements english estimating european evaluation examination experiments extract extracting feature features feedback fifth finland fisher foundations france francisco freitag freund friedman from furnas furnkranz generalization germany goetz greece group guyon hall hampp harper harshman hearst heckerman heidelberg hershey hoffman http huji icml idea ieee improved incomplete independent indexing inductive information intelligence intelligent international invited iterative ization jacobs joachims john johnson joining journal kaufmann kernel kernels knowledge koller kraft labeled laird landauer language large latent learing learner learning lecture lewis likelihood linguistics lodhi machine machines madison making management manning many margin massachusetts matwin maximization maximizing maximum mccallum meeting melbourne menlo method methods mining mitchell model modern morgan moschitti mosenzon multiclass multivariate nashville natural nedellec neto networks neural nigam nips notes number oles optimal optimization ordinateur orleans pages paper papers paris park pazienza pedersen pereira performance phrases platt power practical practice predictions prentice presented press probabilistic problems proceedings processing published publishers publishing rated recherche reducing references regression regular related relevance relevant report representations research retrieval riao ribeiro robin rocchio rose round rouveirol royal rubin sahami saunders scale schapire scholkopf schutze science sebastiani selection semantic semi sensitive sequential series shawe side sigir singer slattery slonim smart smola society sons souroujon springer statistical statistics string structures study supervised support surveys symbolic system systems tampere taylor technical techniques text tfidf theory third thomas thrun tishby tong toward training tutorial uncertainty unifying universitat unlabeled unsupervised usefulness using vapnik vector verlag watkins weiss wesley wide wiley winter with word words workshop world yang yaniv yates york zobel http://www.jmlr.org/papers/volume3/strehl02a/strehl02a.pdf 377 Cluster Ensembles ­ A Knowledge Reuse Framework for Combining Multiple Partitions aaai academic accumulation active advances afrl aggarwal agglomerative alamitos alexander algorithm algorithms alogorithm analysis analyzing annual antonio anuka applicability applications applied approach approaches architecture artificial association august austin automation back backpropagation baeza balanced baltimore bangalore barnett barthelemy base based basis baskets bayesian berlin beyond biology blum boley bollacker bradley branko cagliari cambridge canada caruana case categorization chakaravathy chameleon chan chapter cikm classification classifications classifier classifiers cluster clustering clusterings clusters collective colt combiners combining comp comparison completeness computation computational computer computers computing concept conf conference connection consensus constraints construction cover craig cross dasarathy data databases david decemb decision decomp density department design development dharmendra dhillon dietterich dimensional directions discovering discovery discrete dissimilaritybased distributed document domain doug downloadable drastal duin dynamic easier ecml ective edition editor editors edmonton effective effectiveness efficient eled elements elling ensemble ensembles erger ergraph eriments ersp ertext ervised ester estimate estimation etween evidence extended fast fayyad ferligo first fisher force forecasting forecasts frakes framework francisco fred freeman friedman from function fusion future garey generalized generation george ghani ghosh gini granger graphs gross guide hall hanson hasselmo hastings heidelb heidelberg hershb heterogeneous heuristic hierarchical high highdimensional hillol hipc html http huang hypertext icml icpr ieee ijcai impact inductive industrial information informs initial initialization integrating integration intel international intl intractability invited irregular isolation issue issues iterative jain january jarvis jersey johnson joseph journal joydeep july june kahng kannan kargupta karypis kavek kernel kernighan keynote kittler kluwer knoblock know knowledge kohonen kriegel krishnamoorthy krogh kumar kurt laclerc lans large later lattice lavra learn learning lecture ledge leen letters ligence ligent limited linear lippmann lnai lncs local lorien machine machinery mala management many maps market mathematical mathematics means measure measures mehrotra mercer merugu meta methods michael mining minnesota minton mission mitchell mobasher modeling models modha molecular monjardet mooney moore mosenzon mozer multi multiclassifier multilevel multiple multivariate muslea nada natural nearest neighb netlist nets network networks neumann neural next nigam nips noise norton norwell notes novemb oint oints onent optimization oratory order ordered organizing ositions osteriori ottleneck paclik page pages paral parallel park partitioning partitionings partitions patrick pattern pekalska philip phylogenetic pratt prentice press principal probabilities problem problems proc procedure proceedings processing prodromidis prospects publications quality radial rates raymond recent recognition reduction references refinement refining reina relab related relatedness relationship research retrieval reuse rich richard rivest robust roli rule same sampling sander scalable scale scheme science scientific scott search searching sebastian second selective self semi sept sets seventeenth shared sharing sharkey shashi shekhar siam silver similarity simplification slonim society space spangler sparse spatial special springer statistics stemming steve stolfo strehl structure structures sullivan supervision support supra survey symposium systems talk task tasks technical technology tesauro teuvo text than theory thing thomas thrun time tishby touretzky toward training transactions transfer tree trees tumer tutorial twenty uncertainty under university unlab using utexas validation vedelsby verlag view viewp vipin visualization vlsi volume warnow wiley with workshop yates years yooseph zaki http://www.jmlr.org/papers/volume7/mangasarian06a/mangasarian06a.pdf 159 Exact 1-Norm Support Vector Machines via Unconstrained Convex Differentiable Minimization advances algorithm applications applied approximation april arbitrary astronomical athena automated bartlett based becker belmont bernhard bertsekas bradley california cambridge chicago classification classifiers computation computational computer concave conditions conference control cplex dantzig data december department discrimination distribution edition editor editors effect extensions facchinei feature fiacco fifteenth finite first francisco frank functions fung galaxy generalized gradient guide hastie hessian hiriart html http humphreys icml ilog incline information institute international jersey john journal july kaufmann kernel kernels klaus knowledge large lawrence learning least letters linear logistics lucidi machine machines madison mangasarian manual maratos margin math mathematics mathworks matlab matrix mccormick method methods meyer minimization mining mlearn mlrepository morgan murphy natick nature naval networks neural nevada newton nguyen nips nonlinear norm november obermayer october odewahn operations optimality optimization order pages parallel pennington perturbation philadelphia plane press princeton problems proceedings processing products prog program programming programs quadratic quarterly reduced references report reports repository research result rosset rsvm saul scholkopf schuurmans sciences scientific sebastian second selection separating sequential shavlik siam smola software solution sons springer star statistical stockwell strodiot support suzanna systems tech technical techniques techreports theory thrun tibshirani unconstrained university urruty user vapnik vector village wild wiley wisc wisconsin with wolfe york zumach http://www.jmlr.org/papers/volume6/kawanabe05a/kawanabe05a.pdf 46 Estimating Functions for Blind Separation When Sources Have Variance Dependencies abed acad acoustics adaptive advances algorithm algorithms amari analysing analysis application applications approach architecture artefacts artifact artificial asymptotics audio bach background baltimore basawa based beamforming bell belouchrani berlin bernoulli bickel bifurcation biomedical biometrika blind boden brain brazilian cambridge cardoso cell changing chaos chen cichocki circuits clinical code comon complex component components computation concept conditional conf conference convolutive correlations cosmic curio decomposing deconvolution delayed dependencies dependent differential drifting editor editors efficient efficiently eighteenth electroencephalography emergence engineering entropy environments equations estimate estimating estimation eusipco eventrelated evoked extraction fast field finland fixed from functions garrat gaussian geometrical geometry germany ghahremani godambe gradient granada harmeling harva have helsinki herault hierachical hopkins hoyer huang huber hurri hyvarinen icann icassp identifiability identification ieee images independent indeterminacy information injecting inki instantaneous intelligence international into japan john jordan journal jung jutten karhunen kawanabe klaassen kurths learning lecture letters likelihood line mackert magnetoneurography makeig maximization maximum measuring meinecke meraim methods microwave minimum mixture mixtures models monograph moulines muller multidimensional munich murata mutual nara natl natural nature networks neural neuromimetic neurophysiology nguyen niklasson noise nolte nonholonomic notes ocular olshausen optimum order orthogonal oxford pages parra part pham phase physical point presence press prieto principles probability proc proceedings processing puntonet quasimaximum receptive reduction references rensen resampling responses review ritov robust seattle second sejnowski selected semiparametric semiparamtric separation series signal signals simple soon souloumiac source sources spain sparse spatiotemporal speech spence springer stability stationary statistical statistics struc super superimposed supervised switching symposium synchronization systems tdsep techinique techniques that theis through time tong topographic trans transactions tree ture uncertainty unconditional uniqueness univ university unsupervised using valpola variance verlag vigario volume wellner when wiley with works workshop yang york ziehe ziemke http://www.jmlr.org/papers/volume2/tks02a/tks02a.pdf 315 Lyapunov Design for Safe Reinforcement Learning absolute abstract abstraction achieves acrobot action actions adaptive addison advances agent agents algorithms alspector amherst analysis application applications approximate approximation arms artificial asimovian asme athena automatic automatica automation autonomous backgammon barto based basic behavioral belmont bertram bertsekas biological birkhauser blondel boca books boone boston bradford branicky breckenridge brodley call cambridge carnegie chains cohen cohn colorado communications competitive complexity computation computational conference connolly constrained constructive continuous control controlled cowan craig crites cybernetics daniel danyluk decentralized decision decisions dejong delay derivatives design deterministic dietterich difference dimensional discrete discretization dupuis dynamic editor editors efficient eighteenth elevator eleventh elon empirical engineering estimates etzioni exact exploiting feedback financial finite first fleming florida formation francisco frazzoli freeman functions gaits gammon global gordon grantham grossberg group grupen guided gullapalli handbook harmonic hasselmo heuristic hidden hierarchy high hilbert holt hopfield huber ieee infinite information intelligence intelligent international introduction investigation jankovic johansen john joint jordan journal kalman kanellakopoulos kaufmann kearns koditschek kokotovic kretchmar krstic kushner learning level levine limitations line locomotion lyapunov machine magazine markov massachusetts master mathematical mechanics mellon memory method methods meyn mihatsch minimum model moore morgan motion mozer mukhopadhyay multiple munos narendra national navigation nebel network networks neuneier neural neuro nips nonlinear numerical oleng optimal optimization pages parallel pattern performance perkins peterson petsche planning play policies potential press pricing problem problems proceedings processes processing program programming raton references reinforcement research resolution results rimon rinehart risk robot robotic robotics robots robust safe scheduling schneider scientific search second self sensitive sepulchre sets seventeenth seventh shop singh solla solutions soner sons sontag space spaces spong springer springerverlag stability state stochastic stopping storage strengths structure survey sutton swing symposium synthesis system systems tank tasks teaching techniques temporal tesauro theory thesis time touretzky transactions tsitsiklis tweedie twelfth uncertainty university using variable verlag vincent viscosity weld wesley wiley winston with workshop yale york zhang http://www.jmlr.org/papers/volume7/roverato06a/roverato06a.pdf 141 A Graphical Representation of Equivalence Classes of AMP Chain Graphs aalborg acyclic algorithm alternative andersson annals approach approximate artificial association available barndorff based bayesian between block causal chain chapman characterisation characterization characterizing chickering classes complex component components conference cowell dags dawid denmark diagrams digraphs elec electronically equivalence essential expert flags forthcoming francisco from frydenberg fuzziness graph graphical graphs hall horvitz html http inference influence inst intelligence interaction international jensen journal kaufmann kluppelberg knowledge largest lauritzen learning legal machine madigan markov means merging mixed modeling modelling models morgan network networks nielsen operation operations ordering pages perlman probabilistic proceedings properties property qualitative quantitative reasoning recovery references report research review rocca roverato scandinavian some spiegelhalter springer staff statistical statistics stepanova stochastic structure studeny submitted systems twelfth uncertainty unified university utia variables verlag volf wermuth which with york http://www.jmlr.org/papers/volume7/maurer06a/maurer06a.pdf 108 Bounds for Linear Multi-Task Learning academic advances algorithmic algorithms amer analysis ando annals anthony artificial available barry bartlett baxendale baxter berkeley berlin bias bounding bounds bousquet cambridge caruana classifiers colin colt combined complexities concentration conference cristianini data david denoising discovery discrete distributions empirical error estimating evgeniou exploiting feature foundations framework from function functional gaussian generalization gretsi http inductive information intelligence john journal kernel kernels knowledge koltchinskii learn learning lifelong linear local machine machines margin math mathematical mathematics mcdiarmid measures mendelson methods miccheli michael mika mining model models moments muller multi multiple multitask nello network neural online panchenko papers part physics pontil pratt predictive press probabilistic proc proceedings processing rademacher random ratsch reed references regularized relatedness research results risk scholkopf scholz schuller shawe simon smola spaces springer stat statistics structural structures support systems task tasks taylor theoretical thrun university unlabeled vector zhang http://www.jmlr.org/papers/volume6/segal05a/segal05a.pdf 49 Learning Module Networks aaai about acid analysis analyze annual approach array artificial attribute autoclass averaging ball based battle bayesian being binkley biocomputing bioinformatics biology bonferroni bootstrap botstein breiman brooks buntine carlo causation cell cells cellular changes chapman cheeseman cherry chickering classes classification combination comparison computational condition conference cooper data database dean decisions decomposing degroot dimensionality discovering discovery dolinski domain domains dordrecht doucet dwight economic editor efron eighth elidan environmental expression fifth frame freeman freitas friedman from fusion gasch geiger gene genetics genome genomic getoor goldszmidt gordon graphical hall harris heckerman herskovits hidden hill hoeting hope http icml ijcai induction inferring instances intelligence international into introduction jordan journal kanazawa kelly kluwer knowledge koller lander langseth learning linial local london machine madigan matese mcgraw meek method methods model models module modules molecular monte monterey multiple nachman national nature netherlands network networks nielsen nucleic object olshen operations optimal oriented overlapping pacific pages persistence perturbed pfeffer practice probabilistic proc procedures proceedings processes profiles program raftery reasoning recomb references regev regression regulation regulators regulatory relational research response review rich saccharomyces savin scheffe science segal self sequential seventeenth shapira sherlock sixteenth specific springer stanford statistical stone structural structure studies stutz subnetworks suppl symposium system systems taskar taylor their thirteenth tibshirani trees tutorial uncertainty using value variables verlag volinsky wadsworth weng with workshop wyner yeast york http://www.jmlr.org/papers/volume5/schmitt04a/schmitt04a.pdf 6 Some Dichotomy Theorems for Neural Learning Problems abound access acsc activity agnostic akisara algorithm algorithms amaldi amsterdam annual anthony applications approach artificial aspvall association associative attributes australasian australian away bartlett berichte bidirectional binary biological biophysics blum blumer boas boolean bounded bulletin caianiello calculus cambridge certain chapter chervonenkis complete completeness complexity comprehensive computation computational computer computers computing conference conferences consistency constraints cook covering cubes cybernetics dalmau design dichotomy dimension drastal edition editor editors efficient efficiently ehrenfeucht elsevier emde even examples fang fast fifth flow fogelman formulas foundation foundations francisco freeman from function gallinari garey gate golea guide half hall hampson handbook hanson hard haussler haykin hebbian hoffgen horn icann ideas immanent improved improving informatik information international intractability irrelevant itai italian jeavons johnson journal judd kangas kaufmann kearns keep kirousis kohonen kolaitis learn learnability learning leeuwen letters limitations linear littlestone logical maass machine machinery machines marchand mateo mathematical mcculloch memory methods minimal models morgan multicommodity natarajan natural nervous nets network networks neural neuron neurons node oudshoorn pages palm paris patterns perceptron perceptrons perfectly pitt pitts plass practice prentice press problems proceedings processing prospects quantified quickly random reckhow references research retrieval river rivest robust roche saddle satisfiability schaefer schapire schmitt science sciences scientific second sellie servedio shamir siam sigmoidal simon simula simulations singapore single sixth size society sommer souli sparse stored structure stuckey sydney symposium system systems tarjan technology testing theorem theoretical theory threshold time timetable toward trainability training truth turan twenty ulmer universitat university upper using valiant vapnik venkatesh vietri volper volume warmuth weights when winnow wirn with workshop world zheng http://www.jmlr.org/papers/volume2/crammer01a/crammer01a.pdf 332 Benefitting from the Variables that Variable Selection Discards accuracy adaptive advances against algorithms aliferis alspector ambrosino annual applications aronis artificial attention attribute available back backpropagation bakiri ballard baluja based baxter becker benefitting bengio bernard bishop buchanan carnegie caruana case category classification code coherence colorado colt columbia comparing comparison complex complexity computation computational computer conference connectionist cooper cortes cowan dads dana daphne data david dayne dean december denker denver department developing dietterich editor editors electronically eleventh english errors estimation evaluation fawcett feature features fine focus formalism freitag from future genes geoffrey ghosn glymour gordon gray graycode greedy group hanson hanusa hasselmo hidden hild hinton hints holden html http icml ijcai induction information institute intelligence internal international invariances irrelevant janosky jebara john joint jordan joumana journal kaufmann knowledge kohavi koller layer learn learning lecun leen lippmann machine machines many mapping mateo medicine meek mellish mellon methods michael mitchell moody more morgan mortality mostafa mozer multi multimodality multitask national nature network networks neural nips nist noordewier november optimal oxford pages patrice pattern petsche pfleger pneumonia pomerleau predicting prediction present presented press problem proceedings processing prop propagating provost publishers rankprop recognition recognize references related report representation representations research rich richard richardson risk rumelhart sahami same science selected selection sensing sequences shavlik shumeet simard snowbird sort spatial specific specifying speech spirtes standards steve stock studies subset suddarth support suzanna svms symbolic systems tangent task tasks teacher technical technology tesauro text that theory thesis thing thomas through thrun time tony toronto touretzky towards towell training university unlabeled using vapnik vector victorri virginia williams with workshop yann yoshua http://www.jmlr.org/papers/volume3/brodley02a/brodley02a.pdf 294 Minimal Kernel Classifiers active adaptive advances alex algorithms analysis annals applications applied approximation april arbitrary arlington asscociation august austria bartlett bell bernhard blumer boston bottou bounds bradley building burges california cambridge case chapelle checker checkerboard cherkassky chicago citeseer clas classification classifier classifiers clustering company comparison complexity computational computer computing concave concepts conf conference control convex cortes cplex cristianini dantzig data dataset denker department dietterich digit discovery discrimination drucker edition editor editors ehrenfeucht expectation extensions feature ferris festschrift field fifteenth first fischer francisco from fung gaussian generalized greedy grossman guide guyon handwriting haussler heidelberg herbrich hill html http icml ieee incline incremental information informs institute integer intel interior international introduction isbn jaakkola jackel jectable jersey john jordan journal july kaufmann kernel klaus kleinberg know labs lagrangian large learn learning least leave lecun ledge leen letters ligence linear machine machinery machines madison mangasarian mannila margin massive mateo math mathematical mathematics mathworks matlab matrix mcgraw mean methods meyer minimal minimization minimum mining mitchell mixed mlearn mlrepository modeling models morgan motwani mozer mpml mulier muller munson murphy musicant natick nature neural nevada nonlinear norm november occam october operations opper optimization optimizer pages paral parsimonious pattern perturbation petsche philadelphia physica plane point polyhedral press princeton probabilistic proc proceedings processes processing prog programming programs provost proximal pubs ramakrishnan razor recognition reduced references regression relevance report reports repository research riedmueller rissanen ritter robust rockafellar rosen rsvm sackinger scale schaeffler scholkopf schuurmann schuurmans sciences second selection separating shavlik shawe siam sifiers simard smola smooth society software solla solutions sons sparse springer srikant ssvm statistical statistics stochastic stolfo street study support systems taylor tech technical techreports theory thomas tipping todd transactions tresp university user using vapnik vector verlag version vienna village virginia volker volume warmuth weston wiley winther wisc wisconsin york http://www.jmlr.org/papers/volume6/ratsch05a/ratsch05a.pdf 101 Efficient Margin Maximizing with Boosting adaboost adaptive additive advances algorithm algorithms algoritms also analysis annals annu application applications applied approximation arbitrary arcing artifical bach bartlett based behavior bennett berkeley bioinformatics biological boosting bootstrap bounds breiman california center certain chapman class classification classifiers colt column combination combined comput computation computer conf conference confidence conic constructing convergence convex cristianini cyclic data daubechies december decision demiriz department design divergences duality dynamics earlier economic editor effectiveness efficient efron ensembles entropy error experiments explanation fifteenth finite first framework francisco freund from function fusion game games generalization generalized generation genomic germany gesellschaftsspiele greedy grove hall hastie hettich hill hypothesis icml ieee improved inference infinite information intelligence international interpretable introduction isbn issue jordan journal july kaufmann kernel kivinen koltchinskii kortanek lafferty lanckriet langley large learned learning lecture letters limit line linear lnai lncb lncs lozano machine majority mangasarian manuscript march margin margins math maximizing maximum mcgraw methods mika model models monographs morgan muller multiple multiplicative nash national neues neumann neural neurocolt nips noble nonlinear norm notes october onoda operation optimization order pages palais pami panchenko path plane playing potsdam prediction predictions press probability problems proc proceedings processing programming programs projection quinlan rated ratsch recomb references regression regularized relevant report research review robust rosset rudin schafer schapire scholkopf schuurmans science sciences selection semi separating separator september sequence sequential shawe siam singer smooth sofer soft some sonnenburg spaces sparse special springer stanford statistic statistical statistics study svms sydney system systems talk taylor technical techreport theoretic theorie theory thesis three tibshirani twenty university unpublished using verlag version volume voting warmuth watson weak weights with workshop york zhang http://www.jmlr.org/papers/volume3/bartlett02a/bartlett02a.pdf 312 The Set Covering Machine abound aditi alamitos algorithm algorithmic algorithms amsterdam annual anthony applied artificial association attributes bartlett based berlin bias blake bottou california cambridge chervonenkis classes colt combinatorial communications completeness compression computational computer computers computing conference covering cristianini cruz data databases david decision department dependent dhagat dimension discrete eiron embeddings euclidean eurocolt european floyd foundations framework freeman garey greedy guide half hans haussler hellerstein herbrich heuristic hierarchies holloway html http ieee inductive information intel intractability introduction irrelevant irvine john johnson journal july kearns kernel learnability learnable learning lecture ligence limitations linear lisa lists litman littlestone london luckiness machine machinery machines manfred manual marchand massachusetts mathematics merz methods michael minimization mlearn mlrepository nadav nello netherlands notes november operations other over pages press problem proceedings quantifying quickly ralf reference references relating report repository research risk rivest robert ronald royal sally sample santa saunders schoelkopf science shai shawe simon smola society spaces springer statistical stitson structural support symposium taylor technical theory threshold transactions ulrich umesh university valiant vapnik vapnikchervonenkis variability vazirani vector vladimir volume warmuth weston when wiley williamson with york http://www.jmlr.org/papers/volume7/singliar06a/singliar06a.pdf 181 Noisy-OR Component Analysis and its Application to Link Analysis acquisition advances algorithm algorithms allocaurl analysis andrew annals annual arthur article artificial aston attias august authoritative bartholomew base bayesian baysian belief berkeley binary biocomputing bishop blackford blei bounding buntine california cambridge causal cause ceedings cellular chang christopher citeseer classifier cohn college component components computation computing conference connectivity content cooper cooperative cover data david dempster development diagnosis diez dimension dirichlet document documents donald editor editors edward efficient elements ensemble eric estimating european exact extensions factor factorial fast field francisco from gallan generalized ghahramani gideon graphical gregory group hagai harold hasselmo hauskrecht heckerman henrion herskovits hidden hofmann horvitz html http huan hypertext identify incomplete independence independent indexing induction inference information ings intelligence intelligent international internist introduction isbn jaakkola james john jolliffe jordan journal kaufmann kearns kendall knott knowledge laird latent lawrence learning lehmann library likelihood likelihoods linear link lisp lyle machine mackay mansour markov martin maximum mean medicine method methods michael middleton milos mining miskin missing model modeling models morgan mozer multinomial multiple ncrg network networks neural noisy noisyor oxford pacific pages papers part plausible practical predictions press principal probabilistic probabilistically probable proceedings proceedurl processes processing prot publishers qmrdt quantization quantizer references reformulation report research retrieval review richard roger ross royal rubin sample saul schein schwarz selwyn semantic september severino shwe siam sigmoid singliar society sons springer statistical statistics stockholm structure supervised symposium systems technical theory thesis thomas tion tipping tomas tommi touretzky type uncertainty ungar university using variable variational vector volume vomlel wiley with workshop wray wupes xinghua yishay zemel zoubin http://www.jmlr.org/papers/volume1/smola01a/smola01a.pdf 341 Ranking a Random Feature for Variable and Feature Selection academic aknin algorithm algorithms american analyse analysis analytic annual application applications automatique bancaires barnhill based basis bengio bennett billings biology bjoerck bodor boes breneman building cancer chapelle chem chemical chemistry chen classical classification coefficients combinatorial comp computation conception conference context control cowan cramer crawse criterion curie data decision descriptor design destructive detect development distance donnees dreyfus dunod duprat duxbury dynamic dynamical dynamique editor editors effect eighth embrechts engineering estimation etude european eusipco evaluation example experiment experimental exploring feature filtering filters filtres financieres from function fundamentals gene generation goodwin gram grant graybill guyon hansch harget high hill huang huynh hwang identification ieee information informations informationsbehadlung initialization initiation international introduction jing journal kernel lagarde langage learning least leave leontaritis leverages lineaires linear local logp macgraw machine machines marie matching materials mathematics methodes methodology methods model modeles modelisation models modern monari mood morphology mukherjee myers network networks neural neurocomputing neuronaux neurones nonlinear nordisk obsolescence ondelettes orthogonal orthogonalization oukhellou oussar overfitting pages parameterised paris partition payne performance personnaz pierre poggio points pontil pour prediction press principled problems procede proceedings processing processus property pruning publised pursuit qsar quantitative radial ragot ranking realisation reed references regression research reseaux retrieval riao roussel routing rule schmidt science selection sigir signal signatures society soederstroem solving song soudage squares statique statistical statistics statistiques step stoppiglia stricker struct structure studies support survey svms system systems testing tests text their theochem theorique theory thesis throughput tidshrift towards training traitement transactions trec tzoukermann universite urbani using validation vapnik vector vichot vincent vlontzos wavelet weston wiley wilson with withdrawing wolinski http://www.jmlr.org/papers/volume6/devito05a/devito05a.pdf 60 Learning from Examples as an Inverse Problem academic accepted adaptive advanced algorithm algorithms alon amer analyse analysis annals antoch appl applications applied approximation arbib aronszajn arsenin associes authors bartlett berlin bertero best bianchi bias boston bounds bousquet brain bull cambridge canu caponnetto cbcl cbms cesa chen chicago choices classification classifiers communications comp complexities compstat comput computation computational computationals computer computing conference connections consistency control convergence cucker data david deviations devroye different dimensions discrete distribution dordrecht early editor electronic elements elisseeff empirical engl equations error espaces estimating estimation evgeniou fast feedforward fine first formulation foundation foundations fredholm free friedman from function functional gaussian general generalization girosi goncharsky groetsch group gyorfi handbook hanke hastie haussler hilbertiens hoksbergen http ieee industrial inequalities information institute interscience inverse issn january john journal kaipio kecman kernel kernels kind kluwer kohler krzyzak kurkova laboratory lang large learnability learning least leave lecture lectures linear lugosi machine machines margin massachusetts math mathematical mathematics mendelson methodology methods minimization model models msri mukherjee necessary network networks neubauer neural niyogi noisy nonlinear notes noyaux number numerical observational online original other pages paper parameters parametric pattern philadelphia pike pinelis piscataway pitman poggio point pontil posed preprint press princeton probab probabilistic probabilities problem problems proceedings processing program publication published publishers publishing rademacher rates real recognition reconstruction references regional regression regularisation regularization regularized remarks report reproducing reproduisants research revised revision rifkin risk rosasco rsise rudin russian sakhanenko sampling scale scattered scholkopf schwartz sciences scovel selection sensitive series shannon siam signal singular smale smola society soft solution solutions somersalo sons sous sparseness spline springer springerverlag squares stability statistical statistics steinwart stepanov stopping submitted sufficient sums support system systems technical technological technology theory tibshirani tikhonov topologiques toyota trans transaction translated type uniform university values vapnik variance vector vectoriels vectors verlag vito volume wahba walk washington wiley winston with workshop yagola ying york yurinsky zhang zhou http://www.jmlr.org/papers/volume6/rousu05a/rousu05a.pdf 75 Efficient Computation of Gapped Substring Kernels on Large Alphabets advances agarwal algorithms alignment alstrup analysis annual applications approximate arrays automatic baker bartlett based berg berlin biocomputing bounds brodal california cambridge cancedda categorization chazelle classification classifiers colt columbia common communications computational computer computing concave conference contemporary convex convolution cost costs cristianini cruz data department discrete dynamic efficient eppstein erickson eskin european extensions fast features foundations fragments from functions galil gaussier geometric geometry giancarlo goutte grid haussler helsinki indexing inexact information international introduction italiano joachims journal kernel kernels kreveld kuang large learning lecture leslie line linear local lodhi longest lower machine machines makinen many margin matching mathematics methods mismatch model multidimensional nature neural noble notes olap orthogonal other overmars pacific pages parameterized partial pattern point poon press proceedings processing programming protein queries range rauhe references relatives relevant renders report research rosenberg rousu salton santa saunders scholkopf schuurmans schwarzkopf science searching sequence shawe similarity sixth smola space sparse spectrum springer stafford statistical string structures subsequence sums support syllables symposium systems taylor technical tenth text theoretical theory thesis tree tschach university using vapnik vector verlag vishwanathan watkins weston with wong word workshop yang http://www.jmlr.org/papers/volume3/brafman02a/brafman02a.pdf 369 R-max ­ A General Polynomial Time Algorithm for Near-Optimal Reinforcement Learning agent algorithm algorithms annals apllications approximating architectures artificial atkenson aumann average banos barto based bayesian brafman building center certain classes colell conf conference confirming control convera correlated csaba curious data decision dynamic econometrica efficient embedded equilibrium factored framework fudenberg game games gence general generalized hart hebrew hoffman ijcai incomplete information integrated intel international intl introduction jerusalem joint journal kaelbling karp kaufmann kearns koller leading learning less levine ligence littman machine making management markov maschler mathematical mdps megiddo model monderer moore morgan multi near networks neural nonterminating optimal pages planning played players polynomial press prioratized proc procedure programming pseudo rationality reacting real references reinforcement repeated report research reward schmidhuber science self singh statistics stochastic survey sutton sweeping systems szepesv tadepalli technical tennenholtz theoretical theory time university wellman with http://www.jmlr.org/papers/volume3/gers02a/gers02a.pdf 365 Learning Precise Timing with LSTM Recurrent Networks adaptive advances agtm algorithm algorithms applications applied architecture architectures artificial back backprop backpropagation based bengio berlin blues bottou brauer bulsari calculations cambridge carnegie carrasco chapter chauvin complexity computability computation computational computer computing conf context continual continually continuous control cottrell cued cummins department dependencies descent difficult difficulty diploma doya driven during dynamic dynamical dynamischen editors efficient engineering erlbaum error eurospeech existence experiments explicit fachbereich fallside features felix field fixed flow forcada forget frasconi fred free from fully garrison genevieve germany gers gradient guide hillsdale hinton hochreit hochreiter identification ieee ilchmann infeng inference informatik information initial input institut intl jectories jurgen kaiserslautern kaufmann klaus kolen kremer language languages learn learning lecture lecun lehrstuhl letters line long lstm mateo mathematics mathematik mcclements mellon memory mikel modelling morgan muenchen muller munchen nets network networks netzen neural neuroe neuronalen noisy notes nowlan order oscillations oscillator output owens pages pearlmutter peng phase pittsburgh plaut praetzelwolters precise prediction press proc proceedings processing prof propagation prosody rafael recurrent references regular report robert robinson ruiz rumelhart running schmidhuber science second sensitive sequences sheng short siegelmann simard simple size sontag space springer state storage structure survey systems technical technische temporal term their theory thesis time timing townley trade training transactions tricks tsung turing universitaet universitat university untersuchungen using utility verlag volume weiss williams with without works yoshizawa zipser http://www.jmlr.org/papers/volume8/arias07a/arias07a.pdf 223 Learning Horn Expressions with L O G A N - H academic acyclic alberta ales algorithm algorithmic algorithms also amsterdam analysis angluin applications approximations april arias arimura artificial attributes background bain based basile berlin bianchetti biological blockeel blumer botta bottom bounded bratko bryant buntine camacho canada celine center chang chen classic classification clausal clauses closed communications comparing comparison complete computation computational computing concept concepts conference conjunctions consistency constrainedness constraint continuous costa critical data dechter decision deduction definitions dehaspe demoen deraedt determinate discovery discretization domain dougherty down dutch dzeroski edition editor efficiency efficient ehrenfeucht eighth elsevier empirical entailment errata esposito european executing exhaustive experimental experimentation experiments expressions fanizzi fast fayyad features feng ferilli first formalisms forthcoming foundations fourth frazier free from function functional further generalization generation genomic gent giordana haussler hayes hill holland horn human hypotheses hypothesis identification improvement improving induction inductive information integrated intelligence international interpretations interval invention inverse inverting irani issue jacobs jacques janssens jones journal karwath kaufmann kautz kearns keisler kell khardon king knowledge kohavi krishna laer language large learnable learning letters lloyd lnai local logan logic logical london long lookahead machine macintyre mackworth maloberti matching mauro mcgraw methods michele michie mining mitchell model modeling morgan muggleton multi mutagenesis national natural nature networks note november occam oliver order packs pages pearl pereira phase pitt plotkin predicates press problem proc procedure proceeding proceedings processing progol programming programs prolog prosser queries query quinlan raedt ramon range razor reasoning reddy reducing references refinement refining region reiser relational relations relevant research resolution resource restricted robot rouveirol sahami saitta santos satisfaction sattar scaling science scientist search sebag second selected selman sentences september settings shieber sixth solving special springer srinivasan stanford steps sternberg stochastic stolle structure struyf study subsumption supervised suzuki system systems tadepalli techniques tenth theoretical theories theory thesis theta through transformations transitions trees tsang university unsupervised using valued vandecasteele variables verlag walsh warmuth whelam with workshop wrobel http://www.jmlr.org/papers/volume6/chechik05a/chechik05a.pdf 36 Information Bottleneck for Gaussian Variables accuracy adaptive advances alexander algebra allerton analysis antti appl applications applied artificial auxiliary bachrach baltimore based becker belkin berger berkeley between bialek borga bottleneck breese calculus cambridge canada canonical case channel channels chechik chemometric chemometrics chou citeseer classification climate clustering clusters code coding colt communication complexity computation computations computer conditional conference continuous convention correlation correlations cortical cover criterion data decoder decoding demmel development differential dimensionality dimitrov discovers discrete distortion distributions document duality easily econometrics editorial editors educational eigensolution elements engineers equation estimation extension extracting extraction feature fidelity finland first fmri foundations francisco friedman friman from function functional fundamentals gastpar gaussian general genomics gilad globerson golub goyal hebrew hinton holmes hopkins hotelling html http ieee image industrial information ingwersen inst institute intelligence international interpretation invariant inversion isbn issn iteration january jennings jerusalem john johns journal june kaski kaufmann kernel knutsson koller landelius lappeenranta learning leen lemma leong likelihood linear linsker loan local lossy lundberg macine magazine magnus march math mathematics matrices matrix maximization maximize maximum memory metabonomic method mika miller models morgan mosenzon most muller multivariate mutual nature navot network neudecker neural neuroimage nicholson numerical obermayer organization organizing page pages part partial people perceptual pereira philadelphia plenum pradhan predictable prenticehall press problem proc proceedings processing profiling psychology publications quantization radio ramchandran random rate ratsch real record reduction references relevant report research retrieval revisited rimoldi rules sage scandinavian scholkopf scia second self semi seventeenth shannon side sigir signal simultaneous sinkkonen slonim smola society solla solutions sons source sources space spaces standard statistical statistics stereograms stewart storch structures suffice sufficient surfaces synaptic systems technical that theorems theoretic theoretical theory thesis thomas thompson thousands thrun tishby tnkjtg tradeoff transactions transform tutorial uncertainty university uses using vancouver verified version vetterli volume washington weiss weston wiley with word wyner yeung york zamir zwiers http://www.jmlr.org/papers/volume6/keerthi05a/keerthi05a.pdf 42 A Modified Finite Newton Method for Fast Solution of Large Scale Linear SVMs accurate advances algorithm alpha amherst annals applications approaches approximation athena based behavior belmont bertsekas bjorck blake california cambridge carnegie categorization chakrabarti chung classification classifiers clustering computation computing conference consistency convex data databases decomposition decoste dimensional discovery discriminant equations fast finite frommer fung hauptmann high html international irvine joachims journal keerthi kernel knowledge komarek labs language large learning least letters linear logistic lsqr machine machines making mangasarian massachssetts massachussetts mathematical mccallum mellon merz method methods minimal minimization mining mlearn mlrepository modeling modified multiple neural newton nonlinear numerical optimization ordinal pages paige pasadena pennsylvania philadelphia phillips pittsburgh platt practical press problems proceedings processing programming projections proximal references regression regularization report repository research retrieval risk saunders scale scientific seeding sequential seventh siam sigkdd simple software soundalgekar sparse squares statistical statistics support suykens technical text thesis tikhonov toolkit training transactions twentieth university vandewalle vector vldb wagstaff yahoo yang zhang http://www.jmlr.org/papers/volume6/kim05a/kim05a.pdf 32 Dimension Reduction in Text Classification with Support Vector Machines academic advances algebra algorithm algorithms analysis annual appear applications august baltimore based berkeley berlin berry bjorck brien butterworths cambridge categorization centroids chicago classification classifiers clustered computations computer computing conference cristianini criteria data decomposition decompositions deerwester department dimension dimensional diploma discriminant dooren downdating drmac dumais edition efficient elden engineering european examination extraction feature features fierro first fukunaga functions furnas generalized generalizing golub hall harshman heiler hill hopkins howland ieee indexing information intelligence intelligent international introduction jeon jessup joachims john johns journal kernel kernels koutroumbas landauer large latent learning least linear loan lodhi london lower machine machines mannheim many margin mathematics matrices matrix mcgill mcgraw methods mining minnesota modern nature neural nonlinear numerical optimization orthogonal other pages paige parallel park pattern philadelphia prentice preserving press problems proceedings processing rank recognition reduction references relevant report representation retrieval revealing review rijsbergen rosen salton saunders science second semantic shawe siam sigir signal singular smart society sons spaces springer square squares statistical stewart string structure subspace support system systems taylor technical text theodoridis theory thesis third towards tracking transactions university updating using value vapnik vector verlag watkins wiley with workshop yang york http://www.jmlr.org/papers/volume7/castillo06a/castillo06a.pdf 145 A Very Fast Learning Method for Neural Networks Based on Sensitivity Analysis academic accelerating acceleration activation adaptation adaptive advanced aided algorithm algorithms alguacil alkon almeida alonso amaral american analysis antonina application applications applied approach artificial atmospheric back backprop backpropagation barmann based battiti bayesian beale bengio berdinas betanzos between biegler biological black boca boston bottou bowden building buntine cambridge canteli castillo certain chaos chapman chapter characterization chella cherkassky choosing chow circular civil classification cliffs cobo collobert comparisons computation computer computing concepts conejo conference conjugate controlled convergence curvature cybernetics data demuth dennis derivation derivatives descent design determination determining deterministic difference differential digits direct discrete distribution drago editor editors efficient engineering englewood equations erdogmus error errors estimation extended fast feed feedforward fernandez first fletcher flow fontenla forward from function functional garc general generalization gentile global good gradient gradients guijarro gutierrez hadi hagan hall hard hebbian hinton hollander hush ieee improving increased indeces industrial information infrastructure initial initialization instabilities intelligence international iplikci irvine jacobs january john joint jordan journal kanter kaufmann kaynak kluwer konig langlois large layer learning least lecture lecun leen leung levenberg line linear lippmann lncs local london lootsma lorenz mangis marquardt mateo mathematical mathematics matrix menhaj method methodology methods minimax minimization mixtures modeling modelling models moller moody morgan mozer mulier muller multilayered multiple nature network networks neural neurocomputing newton nguez nguyen nonlinear nonparametric nonperiodic notes number numerical optical optimal optimization optimum order ortigosa oscillations other pages parallel parameter parameters park parker pattern peculiar pedregal perturbation pethel petsche plakhov powell prentice press principe problems procedures proceedings processing products programming propagation properties pruneda publishers publishing quaterly rate rates raton recognition reducing reeves references regression representations rescaling restart reuse review ridella rigler romero rovetta rumelhart saad salas scaled scaling scalora scawi schnabel schraudolph science sciences search second self sensitivity society solla solution solving sons sorbello speed sperduti spie springer squares statistical statistically statistician steepest stochastic stock supersab supervised surfaces suykens systems tarantino techniques technometrics theory through time tollenaere touretzky trade training transactions tricks unconstrained university using values vandewalle variables vector verlag vogl volume weigend weight weights weir widrow wilamowski wiley willian with wolfe working york zink zunino http://www.jmlr.org/papers/volume7/munos06a/munos06a.pdf 117 Geometric Variance Reduction in Markov Chains: Application to Value Function and Gradient Estimation actor adaptive advances algorithms analysis andrew annals appl applied approximate approximation approximations artificial atkeson baggerly bartlett based baxter borkar carlo chains chapman comput computation computing conference connectionist control critic decision discrete dynamical editors elements estimates estimation event following friedman function functionals geometric glynn gobet golowich gradient grant greensmith hall halton hammersley handscomb hastie henriksen horizon importance infinite information intelligence iterative john journal kelton kollman konda korobov learning like likelihood linear locally machine maire mansour marbach markov math mcallester method methods monte moore neural numerical optimization overview pages picard policy press probability proceedings processes processing prospective ratio ratios reduction references regression reiman reinforcement research retrospective review reward roberts sampling schaal scientific search sensitivity sequential series siam signal simple simulation singh smola solution sons space spaces springer statistical statistics support survey sutton systems techniques theory thesen tibshirani tsitsiklis type vapnik variance variates vector weighted weiss wiley williams wilson winter with york http://www.jmlr.org/papers/volume6/maurer05a/maurer05a.pdf 63 Algorithmic Stability and Meta-Learning addison algorithmic algorithms almost american annual anthony artificial association bartlett based baxter bayesian berlin bias bounded bousquet cambridge caruana chicago chorus christianini computational computer concentration conference congnition department devroye discrete edelman eleventh elisseeff everywhere explanation foundations generalization gyorfi hoeffding inductive inequalities intelligence journal kluwer kutin learn learning lifelong lugosi machine machines mathematics maurer mcallester mcdiarmid methods minds model models multitask nature network networks neural niyogi pattern performance pratt press probabilistic probability proceedings prototypes random recognition references report representation research robins science shawe similarity some springer stability statistical sums support taylor technical theorems theoretical theory thrun transfer university vapnik variables vector wesley wolpert http://www.jmlr.org/papers/volume2/jonyer01a/jonyer01a.pdf 347 Graph-Based Hierarchical Conceptual Clustering acquisition activity algorithm algorithms allerman analysis analyze appear application approach approaches arlington artificial authoritative autoclass available background ball baritchi based bayesian beckwith benefits boulton bunke carbonell chakrabarti challenge chameleon cheeseman classification cluster clustering company complexity computation computer computing concept concepts conceptual conference conjunctive cook cure data database databases decision department description discovering discovery discrete distributed djoko domain domains drawing dynamic earthquake efficient electronically empirical engineering environment everitt experience expert fellbaum fifth find fisher florida formation framework freeman from galal gene gibson gonzalez graph graphviz gross guha hierarchical holder http hyperlinked ieee incremental inexact inferring information informative inquiry institute intelligence intelligent international into introduction jonyer journal karypis kaufmann kelly kleinberg knowledge koutsofios kumar langley large lattice learning length letters lexical lexicography line link machine maglothin management markov masters matching measure michalski miller minimum mining mitchell modeling morgan ninth north observation parallel partitioning pattern patterns pazzani policy potential predictive principle proceedings project publishers publishing quinlan raghavan rajapolan rastogi recognition references regulatory research rissanen rivest scalable schalkoff science scientific self sequences shim siam sigmod software sons sources stanford statistical stepp stochastic structural structure structured study stutz subdue substructure symposium system systems taylor telecommunications texas theoretical thesis thompson through tioga tompkins tools toxicology transactions trees university unsupervised used using volume wallace wide wiley wordnet workshop world york http://www.jmlr.org/papers/volume1/mahony01a/mahony01a.pdf 345 Prior Knowledge and Preferential Structures in Gradient Descent Learning Algorithms abbot abhandlungen academic academie adam adaptation adaptive addison advances akaike albert algorithm algorithms amari analysis annual applications approach approximate artificial arun barndorff bayes bayesian berlin bernhard blind boca boothby bounds bruno california cambridge carnegie chang chapman choice christian christoffel clarkson classification claudio cliffs colt comprehensive computation computational computer concept conditioning conference continuous convergence courbure crel cruz curvature curved dale data david dawid decision delaware department dependent descent differentiable differential differentialausdrucke dimensions discrete discriminant disintegration distributions douglas dover duda dynamical edition editions editor editors edwin efficiently elwin englewood english euler exploiting exponentiated extended feedforward filtering filters fine flache flatland from fundamentals gauss general gentile geoffrey geometrical geometrie geometry germany gesel gordon gottingen grades gradient graduate grove grunde hall hart hartigan hassoun haykin helmke hermann hill hirotugu hoffmann homogenen hopkins http hypothesen ichi idee ieee improper information international interpretation introduction investigations isspa jagota john johns johnson jordan jorgensen joseph journal jyrki karl kearns kivinen knowledge kong learning leonhard liegen liepzig likelihood limits linear littlestone london loss lschaft machine mahony manfred manifolds many march markov martin massachusetts material mathematics mccleary mellon methodology methods michael models mohamad moires moore multidimensional murray natural neelandica network networks neural nick nielsen niglichen nonlinear norm optimal optimization pages parameter parametric pattern perish peter phil point pollard prediction predictors prentice preprint prequential press prior probability problems proceedings processes processing proper publish published pubn pubs raton raven reading recherches references regression regret relative results revised rgen rice richard riemann riemannian riemannschen robert robustness romance royal santa sara schuurmans science sciences scott separation september series sethares shiryaev shun signal simon society solla solo solutions source spaces sparsity spivak springer statistica statistical statistics stork structured submitted surface surfaces symposium systems tenth terrence teubner texts theory thesis thrift time toward tranlated transactions transfmoration translation twelfth uber ucsc unabridged university unsupervised updates verlag version versus victor view volume volumes warmuth welche wesley weyl wiley william williamson wilmington wissenschaften with works xuan york zweiten http://www.jmlr.org/papers/volume6/sarela05a/sarela05a.pdf 39 Denoising Source Separation abed accurate activity addison advances algebraic algorithm algorithms algos analysis anderson anti applications approach approaches approximate approximations array artefact artificial astola astrophysical asymptopia attias aussois available azancot baltimore based basic bayesian belouchrani bermond biological biomedical blind blocks book bradford brain building cambridge cancellation cardoso case cdma cell chan choudrey cichocki circuits clarendon clinical clustering code coding cognitive communication comparison component components computation computing conference contrasts control conventional correlations cybernetics data david decorrelation delayed dempster denoising diego differential digital distribution djafari donoho douglas dynamical edition editor editors effective eigenvalue electrocardiogram electroencephalography emergence engineering ensemble enst entropy estimate estimates everson experimental extraction factor fast fastica feature fetal field fields fifth filtering finland first fixed flexible foldiak forming framework france from funaro fundamentals gain gazzaniga giannakopoulos granada gunnar hall hamalainen hansen hari hebbian helsinki hierarchical high himberg hoyer http huang human hyvarinen icacentral icann icasso identifiability identification ieee ilmoniemi images incomplete independent indeterminacy inference info information inki instantaneous instrumentation integration international invariance invariances invariant inverse investigating ionescu isspa isssta jade jammer john johnstone journal jousmaki july jutten karhunen kawanabe kerkyacharian kernel knuth knuutila konig kording kovacevic kuosmanen laird lappalainen latent learn learning letters likelihood linear local localization london lopes lounasmaa luenberger maass machine mackay magnetoencephalographic magnetoencephalography marginal markram matlab maximum mean meinecke meraim methodological methods mika minor miskin missing mixture mixtures models modern mohammad molgedey monographs moore moulines muller multidimensional multitaper natschlager natural nature networks neural neurons neuroscience neurosciences niedermeyer nips nnsp noisy noninvasive nonlinear numerical olshausen optimal optimization order overlearning package pages pajunen paraschiv paris pattern percival perturbations pham physical physics picard point practice prague prentice press principal principles problem problems proceedings processing projection projects properties pursuit raiko raju rake rangayyan ratsch real receptive reconstruction recordings references related reliability rensen representations resampling research review reviews rhythmic ristaniemi roberts robust royal rubin sarela scholkopf schuster schwartz second sejnowski sensory separation september sequences series seventh shrinkage sibony signal signals silva simoncelli simple sites skovde slow smola society software solutions sons source sources space spain sparse spectral spectrum spie spread stability stable states stationary statistical statistics structure studies study subband sweden switching symposium synaptic systems tdsep technique techniques theory third time tong topographic toulouse transactions transformation univariate university unsupervised using valpola variable variational vector vetterli vigario vigneron visualization viterbi walden wavelet wavelets wesley wiley wilkins wilkinson williams winther wireless wiskott with without working workshop ziehe http://www.jmlr.org/papers/volume8/chakrabartty07a/chakrabartty07a.pdf 233 Gini Support Vector Machine: Quadratic Entropy Based Robust Multi-Class Probability Regression acoustics advances algorithm allwein alvira analog analysis annals application applications approach architectures artificial assaleh athena auckenthaler audio bakiri baluja based baum bayesian bertsekas beyond bhattacharyya binary boser breiman brooks broun burges cambridge campbell carey categorization cauwenberghs cbcl chakrabartty characterizations circuits class classification classifier classifiers cliffs codes colt communication comparison comparisons computation computational computer conf conference convex correcting cover crammer cross cvpr data decoding decremental department design detection dietterich digital discovery discriminative dissimilarity dortmund elements empirical englewood entropy error estimation eurospeech experts face fast fayyad forward freund friedman functions fundamentals gacv gaussian generative gini girosi gish gopalakrishnan grove growth guyon hastie haussler herbrich hierarchical hilbert holloway huber hybrid icassp icpr identification ieee imitative import improvements incremental independent inequality information informative intelligence internal iscas jaakkola jacobs jaynes jebara joachims john joint jones jordan journal juang kanade kanevsky keerthi kernel kernels knowledge kwok laboratory landmarks large lawrence learnability learning likelihood linear lloyd location logistic london machine machines making manifolds margin math mathematical mathematics measures mechanics media memo methods microwatt minimal mining mixtures modeling models moderating multi multiclass murthy nadas nahamoo nature nayak network networks neural nips normalization object olkopf olshen optimal optimization oren osuna output outputs pacific pages papageorgiou paper parameter pattern pedestrian physics pietra platt poggio polynomial pontil practical prenticehall press princeton probabilistic problems proc process processing programming propagation properties quadratic rabiner randomized rational recognition reducing references regression regularization regularized relevance report reproducing research review rifkin robust rockefeller rowley royal scale schapire schmidt scholkopf science scientific score seeger sell sequence sequential shevade signal singer sinha smola snow solving some sons spaces sparse speaker speech springer statistical statistics support svms symp system systems technical technology templates text theory thesis thomas tipping training trans transactions transformations trees tutorial unifying univ university using vapnik vector verification verlag verri vision vlsi volume wadsworth wahba watkins wavelet weston wiley with workshop york http://www.jmlr.org/papers/volume7/bickel06a/bickel06a.pdf 128 Some Theory for Generalized Boosting Algorithms adaboost adaptivity addison additive advances aggregation algorithm algorithms amer analysis andersen application arcing assoc asymptopia baraud barron bartlett base based baxter bayes behaviour berkeley bickel birge boosting boston bounds breiman buhlmann california chain classes classification classifiers combining comment company computation conference confidence consistency convergence convex convexity counting data department dept design dictionaries discrimination discussion distribution donoho early edts empirical ensembles experiments fields francisco frean free frequency freund friedman func functional functions games gill golden gradient greedy gyorfi hastie ieee improved infinity information international jiang johnstone jordan kauffman kerkyacharian kohler krzyzak large learnability learning linear logistic loss luenberger lugosi machine machines majority mallat mammen mannor margin mason massart matching mcauliffe measures meir methods microarray millar minimax minimization model morgan multicategory neural nonlinear nonparametric northwestern optimal paris part penalization picard potheses prediction predictions predictor press prob probability proc process processes processing programming publishing pursuit radiance random rates reading references regression related report research risk ritov sample satellite schapire scholkopg schurmans selection shrinkage signal singer sinica smola smooth some springer stat statist statistica statistical statistics stopping strength study support tech technical techniques theory tibshirani time tions trans transactions tree trees tsybakov type under uniform university using vaart vayatis vector view wahba walk wavelet weak wellner wesley with yang york zhang zurich http://www.jmlr.org/papers/volume3/forman03a/forman03a.pdf 279 Learning Rules and Their Exceptions abney acker adjective adverb after algorithms allis andr andrei anja annotated antal approach argamon article artificial australian based beijing belz biases birnbaum bosch brill british brkic bruce brunk building buttersworth califf cancedda carb cardie cascades case centre chanod chunk chunking claire claude clausing close cognitive colin coling collins colon combination comma common comparative compilation complementizer computational concept conference conll coordinator corpora corpus currency daelemans dagan dash data decision definitions dejean dennis descriptions determiner discovery disjuncts dollar double driven eacl editors eech effect elaine eleventh empirical endency ending english entropy enumerator ergen eric erik erlative error ersonal esslli exceptions experimental explanation extraction finite first foreign forgetting form france franck free from full good grammars grenoble grover half halteren hammerton hans harmful herv hill holder holte http hybrid identification induction inductive inferred infinitival information input integrating intel interjection international introduction investigation item jakub james jean john joint journal kamal karttunen kaufmann kemp khalil kibler know knowledge koeling kolluri konstantop krymolowski kudoh lance language large lauri learned learning ledge liane ligence linguistics lisb list lists logical machine madrid marcinkiewicz marcus mary mateo matheson matsumoto maximum mcgraw memory methods michael mikheev miles mining miscellaneous mitchell modal modeling models mokhtar montr mooney morgan multi natural nedellec nerb nicola nnps noise noun numb onell onne order oriented orne osition ossessive oulos ound pages paren parser parsing part participle particle past patterns pazzani pierre plural porter portugal predeterminer prep present press prior probabilistic problem proceedings processing pronoun prop provost quinlan quote ramshaw raymond reducing references relational relations research results retrieval rijsb robust role roux rules sang santorini scaling selecting semi seventh shallow shared shlomo sign sima singular small soderland software speedup stasinos state stephen steven stop structured study supp survey symb system systems table tagging tagset taku tamas task technical technologies tense text their theoretical theory there third this thollard tjong tokenisation tolerant tool toulouse tranformation transformation treebank unifying uninflected used using utility vector verb verbs version very volume walter webb with word workshop wpdv xerox yuji yuval zavrel http://www.jmlr.org/papers/volume2/dejean02a/dejean02a.pdf 319 Policy Search using Paired Comparisons accuracy acta adaptive addison advances advantage algorithm algorithms allocation analysis andre annual application approximation artificial australian automatically baird bartlett barto based baxter berkeley biennial boutilier boyan building calculations cambridge carnegie carolina cases choosing college communication complexity computer condition conference continuous control cross data david dean decduction decrease delayed department descent design detection differential direct discretization dorigo dundee dynamics editor editors efficient engineering error evaluation evolution experimenters fitness fogel francisco from function functions garzon general genetic germany global goldszmidt gradient griffiths harada harmon high hitoshi hunter idsia ijcai improve information institute intelligence international introduction invariance john joint jordan journal kalsruhe kalyanmoy kaufmann kelley king koza laboratory lagaris large learning letters likas local logic longman machine mansour marco mathematics mcallester mdps mead mellon method methods minimization minimizing model moore morgan mountain multiplayer munos national nelder neural neurocontrollers north number numerica numerical once optimal optimization over pages paired pegasus personal pittsburgh policy polytope pomdps powell press price problems proceedings processing programming publishers rational references reinforcement remediation report reprinted research residual resolution respectable reward rewards rick riolo russell scheme schmidhuber schoknecht school science sciences scorned search second shaping simple simplex singh solutions spaces stagnation stanford state statistical statistics storn strens sufficient sutton system systems technical teller tests theory thesis thomas through training transformations trials uncertain uncertainty under university using validation variable watkins watson wesley wiley with wright york zhao http://www.jmlr.org/papers/volume7/moser06a/moser06a.pdf 196 On Representing and Generating Kernels by Fuzzy Equivalence Relations aachen about academic aggregation algorithms amsterdam analysis angeles appl applications applied approach approx approximate aspects associative baets based bases bayesian berlin bochner bodenhofer boixader braunschweig california calvo cambridge castineira chapelle classes classical cluster component computation computing congress connectives control covariance cubillo data debrecen demirci determinants dimensional domination dordrecht dover dubois editors eigenvalue eklund equalities equality equivalence european families fields fitzgerald flammarion foundations from functions fuzziness fuzzy gebhardt generating generative generators genton gottwald harmonic heidelberg hohle holland hoppner huang huge hypothese improved indistinguishabilities indistinguishability inequalities inform information inquiry intelligent interdisciplinary internat into jacas johannes john jolliffe journal kecman kepler kernel kernels klawonn klement kluwer knowledge koln kopriva kruse learning linear ling linz logic logics machine many math mathematics matrices mayor menger mesiar metric micchelli mining models moser muir muller neural nips nola nonlinear norms north note october operators orderings ovchinnikov pages paradox paris partitions perspective physica pinkus poincar poincare positive prade preservation preserve press principal probabilistic probability problem proc processes processing publ publishers random reason reasoning references regression reidel related relations representation representing research review revisited rheinland rule saminger scholkopf schweizer science seeger semi semidefinite series sets similarity skala sklar smola soft some sons space spaces springer stationary statistical statistics stochastica structure studies stuttgart supervised suppl surfaces systeme systems technologies termini teubner that theory thesis transitivity treatise trends triangle triangular trillas uncertain universitat university unsupervised vagueness valeur valued valverde ventre verlag versus vieweg volume weston wiley with yaglom york zadeh http://www.jmlr.org/papers/volume5/hoyer04a/hoyer04a.pdf 27 Non-negative Matrix Factorization with Sparseness Constraints academy acoustics adelson advances algorithms also america amsterdam analysis arbib artificial automatic based bell between blind bloch brain brunet buchsbaum cambridge categories cell cells cheng code coding color component components computation computational computer conf content convergent correct cortex cortical cvpr data decomposition density dimensionality discovery distributed does donoho edge editor elsevier emergence environmetrics error estimates event expression extraction factor factorization field fields filters fmri foldiak freeman functions gene genome give globally goal golub gradient handbook hawaii heeger hoyer hyperspectral hyvarinen icassp identification ieee image images independent indexing information intelligence interscience into journal jung karhunen knowledge large learning lecture localized martigny massachusetts matrix mesirov metagenes michael mining model modeling molecular muller multiscale munsell national natural nature negative networks neural neurocomputing neuroimage neuroscience nips notes objects olshausen optical optimal paatero pages parra parts pattern plumbley porrill porter positive press primate probability proc proceedings processing properties published receptive recognition reduction references related relations representations research response revealed sajda scale scenes schutter sciences search sejnowski sensory separation seung shiftable signal simoncelli simple skewed society sources sparse spatially spatiotemporal spectra speech spence statistics stodden stone subsystem switzerland tamayo tapper text theory thorpe through tidor transactions transforms trends unmixing using utilization values versus vision visual what when wiley wilkinson with workshop young zhang zheng ziehe http://www.jmlr.org/papers/volume7/bach06a/bach06a.pdf 166 Considering Cost Asymmetry in Learning Classifiers aaai academic active advances aistats algorithms american analysis annals appear applications artificial association asymmetry asymptotic avis bach bartlett based bayes behavior blake bleistein boyd bremner cambridge characteristic classification classifiers computational computations computing conference considering consistency convergence convex cost cristianini cross curve data databases dover efficient elements entire environments expansions fast fawcett feature flach friedman geometry golub good handelsman hastie heckerman hollander hopkins horvitz hull icml ideal implementation imprecise inference information integrals intelligence international isometrics john johns jordan journal kernel kernels kohavi koller large learning loan loss machine machines margin maros matrix mcauliffe merz method methodology methods metrics minimal minimization multiple neural noise nonparametric operating optimal optimization overfitting path paths pattern pepe platt press preventing processing provost publishers rates receiver references regularization repository research restricted risk robust rosset scheinberg scholkopf seidel selection sequential shawe simplex smola sons space springer statistical statistics subset support systems taylor techniques theory thibaux through tibshirani tong training understanding university using uwer validation vandenberghe vector verlag volume wiley with wolfe workshop wrappers zhang http://www.jmlr.org/papers/volume5/rosset04a/rosset04a.pdf 8 Boosting as a Regularized Path to a Maximum Margin Classifier adaboost additive algorithms angle annals application approximation arbitrary arcing asymptopia bartlett baxter berkeley blake boosting bregman breiman california certain college collins computation computational computer conference convergence convex databases december decision density department dept descent distances donoho early effectiveness efficient efron elements estimation european explanation frean freund friedman function games generalization gradient greedy hastie holloway html http ieee information irvine jmlr johnstone kerkyacharian learing learning least letters leveraging line logistic london machine mangasarian march margin margins mason maximization merz methods mika mlearn mlrepository muller neural neurocolt nips norm onoda operations optimization pages picard plane prediction problems processing ratsch references regression report repository research results rosset royal schapire science segal separating sequential shrinkage singer soft springer statist statistical statistics stopping submitted systems techincal technical theoretic theory tibshirani transactions univ university verlag view volume voting warmuth wavelet with york zhang http://www.jmlr.org/papers/volume6/banerjee05a/banerjee05a.pdf 76 Clustering on the Unit Hypersphere using von Mises-Fisher Distributions aaai abramowitz academic academy adai advances algorithm algorithms amari american analysis annual anti application applications applied approaches approximation april apriori armour artificial assignment association augmented austin automatic auxiliary baeza balanced banerjee bard based battle bennett berkeley bessel bilmes biocomputing biology blum book botstein bounds bradley bregman brown buckley cambridge case cell cellular chakraburtty chapter characteristics circular classification classifiers click cluster clustering clusterings clusters coffey cohn coleman collaborative collections collins colt communications company comparative comparing comparison comparisons compendium competitive computation computational computer computing computional concept conditional conf conference constrained correlated cover criterion dasgupta data databases date death decomposing decompositions demiriz dempster department dhillon diametrical dimensional directional discover discovery discriminative display distribution distributions divergences document documents dong dordrecht dover dowe dubes edition editor editors efficient eisen elements empirical engineering erlbaum estimation evaluation exam exploiting expression extensions external feature filtering finite fisher fitting foundations frakes framework frequency friend from fulfillment functional functions gachotte gaussian gaussians gene generative genes genomewide gentle geometry ghosh gominer good graphical grossman guan hall handbook hard hardin haussler helms hidden high highdimensional hill hinton holland html http hughes hyperspheres icsi identification identifying ieee iida impact incomplete incremental indyk infant inference information intelligence intelligent international interscience into introduction involving ismb issues item iterative jaakkola jain january jersey john joint jolliffe jones jordan journal july june jupp justifies kamath kanal kannan karypis kaski kearns kent kernels kluwer knowledge kogan koller konstan krishnaiah krishnan kumar labeled laird langford large lawrence learning likelihood linear local machine management mansour marcotte mardia markov marton massachussets mathematical maximum mccallum mcgill mcgraw mclachlan means measure measures meila merugu methods metric microsoft mining mises mitchell mixture mixtures model modeling models modern modha molecular mooney mozer multi namburu national natl neal network networks neural newsgroups nigam nong north november objective other pacific page pages parameter patterns pearson peel petsche piater poisson preliminary prentice press probabilistic probability proc proceedings processes processing profiles publ publishers rasmussen recommendation references reidel reidl report requirement research retrieval roberts rocke roshan royal rubin salton sarwar scalable scheme scholkopf science sciences scientific search section segal selected sensitive september sequences series shamir sharan shimizu shoemaker siam similarity simon simulation sinkkonen smola smyth society soft solla some sons space spaces sparse spectral spellman spheres state statistical statistics stegun steinbach stoughton strehl structures study sublinear sudden symposium syndrome systems technical techniques term texas text that theoretic theoretical theory thesis third thomas thrun time transactions treatise tutorial type uncertainty unified university unlabeled using validity variants variation vempala very vetta view visual volume wallace watson weighting whiten wide wiley with wood woodruf work workshop world written yates yeast york zhao zhong http://www.jmlr.org/papers/volume4/lee03a/lee03a.pdf 240 SVMTorch: Support Vector Machines for Large-Scale Regression Problems advances algorithm appear arti available bengio bhattacharyya burges chichester cial cient cjlin classi college collobert computation control conv convergence csie decomposition dept design division drucker engineering external fast flake fletcher freund generalized germond gerstner gilbert giles girosi guppy hasler holloway homepages http icann idiap ieee improved improvements information joachims john jordan kaufman keerthi kernel kohlmorgen large laskov lawrence learning leen london machine machines making mechanical method methods minimal morgan mozer mpessk muller murthy national nature networks neural neurocolt nicoud optimization osuna papers petsche platt practical predicting press principe problems proceedings processing production ratsch references regression reports royal scale scholkopf second sequential series shevade signal singapore smola smorch solla sons springer statistical submitted support svmtorch systems taiwan tech theory time training transaction tutorial university using vapnik vector wiley wilson with workshop york http://www.jmlr.org/papers/volume5/jebara04a/jebara04a.pdf 4 Probability Product Kernels advanced advances algorithm alignment analysis applications applied approach approximation asai based bengio between beyond bhattacharyya bioinformatics biological bishop browsing bull burges calcutta california canu categorisation chapman chapter classification classifiers cluster clustering collections collins composite conference convolution cortes cristianini cruz cutting data database davidson defined density diffusion discrete discrimination discriminative distributions divergence document duffy dynamic econometrics editors elements entropy eskin estimation expected exploiting factorial fast field forecasting frasconi friedman full gather generative ghahramani goldzmidt graphical graphs group haffner hall hastie haussler hidden human hypertext icml ieee inequalities inference information inokuchi input intelligent international introduction jaakkola jebara joachims john jordan karger kashima kaufmann kernel kernels kondor kullback labeled lafferty language large learning lebanon leibler leslie likelihood linear london maccormick machine machines mackinnon marginalized markov matching math mathematics maximum mean measure measures meila methods minimal mismatch models mohri moreno morgan motion multimedia natural networks neural noble optimization output oxford pages pavlovic pearl pederson platt plausible populations positive practice press probabilistic probability proceedings processing progress protein publication pure rational reasoning references regularization rehg related report sahami santa scatter scholkopf september sequence sequences sequential series sets shawe shumway sigir silverman smola smoothing some sons springer stanford statistical statistics stoffer string structures superkernels support switching systems taylor technical tenth text their theory tibshirani time topsoe training transactions tree tsuda tukey tutorial ucsc university using vapnik variational vasconcelos vector vectors verlag vishawanathan watkins weston wiley williamson with xavier http://www.jmlr.org/papers/volume6/aiolli05a/aiolli05a.pdf 58 Multiclass Classification with Multi-Prototype Support Vector Machines advances aiolli algorithm algorithmic algortihm allwein annual approach artificial bakiri based bayesian bhattacharyya binary burges chapter class classification classifier classifiers codes cognition computation computational computer conference control convergence correcting crammer cristianini dags data dept design dietterich discovery discriminant distributed division downs editor editors efficient elisseeff ellis empirical engineering error esann exact explorations extracting facto fast feature first francisco fung gates given glover guermeur handbook helsinki hill hinton horwood html http hynninen ieee ijcai ijcnn implementation improvements improving information input intelligence internal international introduction invariant iterated january joachims joint journal kangas kaufmann kearns keerthi kernel knirsch knowledge kochenberger kohonen laaksonen laboratory large learnability learning leen like local lourenco machine machines making management mangasarian margin margins martin masters mcgraw mechanical metaheuristics methods michie microstructure mika minimal mining mitchell moisy morgan muller multi multiclass murthy national networks neural neurons nnrc nnrcprograms operations optimization output package pages parallel pattern paugam platt practical press problems proceedings processing production program programs propagation prototype quantization quinlan ratsch recognition reducing references relevance report representation research result rumelhart scale schapire schlkopf scholkopf science search sequential series shawe shevade signal simplification singapore singer smola solla solutions solving sona space sparse speigelhalter sperduti starita statistical strategy stutzle support systems task taylor technical technology theory thirteenth tipping torkkola training transactions transformation uniform unifying university using vapnik vazirani vector verleysen versus watkins weighting weston wiley williams workshop york http://www.jmlr.org/papers/volume7/blanchard06a/blanchard06a.pdf 112 In Search of Non-Gaussian Components of a High-Dimensional Distribution algorithm algorithms analysis annals another approach based belkin biomedical bishop blanchard blind bootstrap bounded braun buhmann chapman clustering combinatorics comon component components computation computers concept data densitometry devroye differences dimensional dimensionality dual efron eigenmaps eigenvalue embedding energy engineering estimate exploratory fast fixed flow framework friedman gamma generative geometric global gyorfi hall harmeling huber hyvarinen ieee independent instruments jackknife james jones journal karhunen kawanabe kernel lange langford laplacian lecture linear locally london look lugosi mapping math mcdiarmid meinecke method methods muller multidimensional multiphase networks neural niyogi nonlinear notes nuclear pattern physics point probabilistic problem processing projection pursuit recognition reduction references representation resampling research robust roth roweis royal saul scaling scholkopf science separation series sibson signal silva smola society solutions source springer stability statistical statistics surveys svensen tenenbaum theory topographic transactions tukey using validation what wiley wiliams ziehe http://www.jmlr.org/papers/volume3/bi03a/bi03a.pdf 276 Shallow Parsing using Sp ecialized HMMs abney academic acidca advances adwait aist algorithms allis ambiguity anlp annual antal antonio applied applying approach argamon artificial association atro august basado base based berwick bigramas bigrams boosted boosting bosch brants brill buchholz bulgaria cactus canada carreras cascades case castellano chang chanod chark christer chunk chunker chunking chunks church classification classifier clause clausing coling combination compression computational conference conll connll constantin context contextual corpora czech daelemans dagan damerau david decision detection detector djean dordrecht driven eacl editors emnlp empirical english entropy erik erikt error especializados esslli etiquetador evaluacion evidence exploring ferran finite france fred germany goyal graphs guodong halteren hammerton hans herv hidden hmms http hybrid ichi identification igtree improving incremental information intactico intelligence international introduction ishaan issue james jean jian johansson johnson join joint jorn july june juni kaufman kluwer koeling krymolowski krymolowsky kudo language large lazy learning lenguaje lexical lexicalization lexicalized likelihood linguistics lisbon long machines marcus markov marquez maryland matsumoto maximum means meeting memory merialdo method methods miles model models mokhtar molina monastir montr morfos morgan naacl nara natividad natural noun nptool orasan osborne pages papers para parser parsing part partial parts patrick patterns pennsylvania phrase phrases pittsburgh portugal prague prieto principle probabilistic proce proceedings processing program publications publishers punyakanok ramshaw ranlp ratnaparkhi recent references regularized republic resolution review revista robust roth saarbrucken sabine samiento sang seattle selective sensitive september shallow shared short sigdat single sixth special specialization specialized speech splitting state statistical stochastic structures study support syntactic system tagger tagging taku task tasks tenny term text texts thesis third thorsten tjong tong tongguan toulouse transformation trees tsujii tunisia tzigov university unrestricted upenn using vasin vector veenstra very voutilainen walter washington weijters winnow with workshop wpdv wvlcbook xavier yuji yuval zhang zhou http://www.jmlr.org/papers/volume3/bengio03b/bengio03b.pdf 275 Memory-Based Shallow Parsing aaai abney academic accuracy accurate achieves acquisition advances adwait alexander allis ambiguity anlp anlpnaacl annotated annual antal antonio applied applying approach argamon artificial association attribute austin bankert banko base based beatrice bergen boosted boosting bosch brants brill brunswick buchholz building buttersworth cambridge carreras caruana cascaded case casimir charniak chunk chunker chunking chunks church class classification clause clausing cloud coling collins combination comparison compositional computation computational computer conference conll context corpora corpus daelemans dagan damerau dataoriented david dayne decision derivation detection differences disambiguation discriminative driven durieux dutch eacl ejerhed eleventh emnlp empirical engelson english entropy eric erik eugene evidence exceptions experimental expert feature features ferran finite finland forgetting fourteenth fragments france fred free freitag from george germany gert gillis goyal grammar graphs greedy guide guodong halteren hammerton hans harmful head helsinki herv hong hoste http hybrid hypotheses icml identification improving incorporating induction information inspired institute intel intelligence intensive international introduction irrelevant ishaan issue jakub james jean jetai jian john johnson jorn journal karl kaufman kaufmann kenneth kluwer koeling kohavi kong kryee krymolowski kudoh kulikowski lance language large learn learner learning lecture lexical lexicon ligence linguistics lisbon long machine machines magerman marcia marcinkiewicz marcus markov marquez mary matsumoto maximal maximum meeting memory memorybased methods michael michele miles minimal mitchell model modeling models molina molowski more morgan munoz naacl national natural netherlands nets neural noreen norway notes noun osborne pages papers parse parser parsing part partial parts patrick patterns penn pennsylvania pfleger phrase pittsburgh portugal prediction press principle problem proceedings processing program pronunciation psycholinguistics publishers punyakanok ramshaw ratnaparkhi reasoning recognition reference references regularized rens report representing reranking resolution result retrieval rich richard rijsbergen roth rule saarbruecken sabine sang santorini scaling scandinavian science seattle second selection seventh shallow shared shlomo sholom short shortest significance sloot some sons special speech splitting springer stanford state statistical statistics steven stochastic stress structures subset support syntactic system systems tagging taku task tasks technical term testing tests texas text that theoretical thesis third thorsten through tilburg timbl tjong tong tongguan toulouse transformation transforming translation tree treebank trees twelfth types university unrestricted using utrecht variation vasin vector veenstra veronique version very walter ward washington weiss what wiley winnow with word workshop wpdv wvlc xavier yuji yuval zavrel zhang zhou zimak http://www.jmlr.org/papers/volume6/rakotomamonjy05a/rakotomamonjy05a.pdf 81 Frames, Reproducing Kernels, Regularization and Learning aarhus academic acceleration additive advances agnan akram algorithm amato amer anal analyse angewandte annual antoniadis appear applic applications applied approximation architectures aronszajn arsenin atteia austria based berlinet between birlinghoven boser brezis building burges buseco business cambridge cbms chapman christensen class classifiers colt computation computational computing conference conferences connection creating danemmark data daubechies debnath decomposition dept design discovery duffin econometrics edition editor editors equispaced equivalence estimation evgeniou examples expansions fonctionnelle fourier fractales frame frames function functions gaches generalized girosi golowich gottingen grenoble grochenig groetsch gunn guyon hall harris hastie haussler hilbert hilbertian hilbertienne holland http hyndman ieee incorporating incorrectly index information institut introduction inverse jaakkola jones jordan kearns kernel kernels kimeldorf kluwer knowledge learning lectures library machine machines madison mallat margin masson math mathematical mathematics mathematik method methods mikusinki mining model models monash morosov muller multiscale nature networks neural niyogi nonharmonic north numerische observational ondelettes operators opfer optimal pages pattern penalized pensky pittsburgh poggio pontil posed press presses prior probabilistic probability problems proc proceedings processing processings published publishers recognition references regional regression regularization report reproducing results schaeffer scholkopf sciences series siam signal simard smola sobolev sohn solla solutions solving some spaces sparse spline splines springer statistical statistics support systems tchebycheffian technical theorie theory thesis thomas tibshirani tight tikhonov time tour training trans tsdl tutorial univ universitaires universitt university vapnik vector verlag vienna vieweg virtual volume wahba wavelet wiley winston wisconsin with workshop wwwpersonal http://www.jmlr.org/papers/volume4/mesterharm03a/mesterharm03a.pdf 258 A Generalized Kernel Approach to Dissimilarity-based Classification acad academy advances algebra algorithms almost analysis appear approach arkadev artificial automatic barcelona bartlett based bennett bollmann borg bradley braverman british brussels burges cambridge canada chapman character class classes classification classifiers coefficients cohn columbia compactness complexity component computation computer computers computing conference data database deformable devijver digits discriminant discrimination discussion dissimilarity distance distances dubuisson duda duin editing editors electronic elsevier empty euclidean feature features february friess fukunaga gacv garris gdalyahu geometry goldfarb gower graepel gramming graphs greub groenen hall handprinted handwritten hart hausdorff herbrich hilbert householder ieee image information informs inseparable institute intel international introduction invariance jacobs jain ject john journal kanal kearns kernel kimia klein learning letters ligence linear linearly london machine machines mangasarian matching mathematical memoriam methods metric military modern modified muller multidimensional munich mutual national nature networks neural obermayer optimization pacl page pages pairwise pattern pekalska pierre points port press principal problems processing processings programming progress properties proximity psychological psychometrika publishers randomized recognition references relational report representation representations reproducing retrieval review ridder robust rosenfeld royal scaling scholkopf scia science scientist sdorra sebastian selection semiparametric sets shapes shock similarity smola software solla sons spaces spain springer standards statistical statisticsal stork street support symposium system systems technical technology templates terms their thesis thompson transactions trick tversky unified using vancouver vapnik vector verlag vision volume wahba washington weinshall wiley williamson wilson with york young zongker http://www.jmlr.org/papers/volume8/biehl07a/biehl07a.pdf 216 Dynamics and Generalization Ability of LVQ Algorithms ability adaptation advances algorithm analysing analysis applications approach arbib artificial automated available averaging bachrach barkai barna based becker belgium benchmarking berlin bibliography biehl bioinformatics bloch bode boden bojer bottou brain broeck cagliari cambridge case caticha centre changing chrisley classification classifier classify clustered clustering competitive computation computational computer computing conference convergence cottrell crammer data diego distinction duda dynamical dynamics editions editor editors empirical engel ensembles environments esann estimation european europhysics evere examples feature flotzinger framework freking from fuzzy general generalization generalized ghosh gilad gradient grlvq groningen hammer handbook hart helsinki html http icann ieee image improved information input intelligence international italian italy jain joint kittler koers kohonen kuncheva laws learning lecture leen letters liinwww likelihood line local logic maps marangi margin marinaro mathematics maximum measure mechanics meir merenyi metric minimization modern monitoring multiple navot nearest netherlands nets network networks neural neuro neurocomputing niklasson nimes notes obermayer online organizing otaniemi pages paradigms paris pattern petrosino pfurtscheller physical physics pregenzer press proc processing proteomic prototype quantization recognition reents references relevance remote report research review reviews riegler risk roli rule rules saad sato scaling schleif schweitzer science scientific second seiffert selection self sensing sensitive seung side similarity singapore sites soft solla sompolinksy splice springer statistical stochastic stork strickert studies study supervised symposium systems tagliaferri tasks technical technology tesauro tettamanzi theoretical theory thrun tishby touretzky transactions type univ university urbanczik using vector verleysen versions versus vietri villmann volume watkin wiley windeatt wirn with workshop world wsom yamada york ziemke http://www.jmlr.org/papers/volume4/srinivasan03a/rev1/srinivasan03ar1.pdf 272 Stability and Generalization academic advances algorithmic algorithms alon analysis annals approximate april artificial bagging bartlett based bianchi bonnans bound bounded bounds bousquet breiman building cambridge carnegie cesa check classification combinatorics computation convergence convex cover crossvalidation data david december decision deleted department dependences devroye differences dimensions discrimination distribution efron elements elisseeff empirical entropy equivalent error estimate estimates estimation evgeniou exponential finite framework free function functional generalization girosi gordon guided haussler heuristics hilbert holdout ieee important independence inequalities inequality information inria instability institute intelligence introduction jaakola jebara john journal kearns kernel kluwer laboratory learnability learning leave local look lugosi machine machines madison markov massachusetts maximum mcdiarmid meila mellon memo method model more multilayer network networks neural nonparametric nonsymmetric optimization pages pattern pawlak performance perturbation poggio pontil posterior potential predictors press princeton probabilistic probability problems processes processing publishers recognition references regularization related report reproducing rockafellar rogers rules sample sanity scale science selection sensitive shapiro size solutions spaces springer stability stabilization statistics steele stein support surveys systems talagrand technical technology than that theory thesis thomas topics tour transactions unified uniform university valid vapnik vector verlag wagner wahba weights wiley wisconsin with http://www.jmlr.org/papers/volume6/ando05a/ando05a.pdf 91 A Framework for Learning Predictive Structures from Multiple Tasks and Unlab eled Data algorithms analysis ando annual appear approach artificial banach based baxter belkin berlin bias blum bounded bousquet breiman cambridge cardie caruana character chieu chunking classification classifier clustering colt combination combinatorics combining computational conf conference conll consistency convergence data datasets david descent differences disambiguation discovery discussion documents elements eleventh emnlp empirical entity entropy evegniou exploiting fields florian friedman from functions gaussian ghahramani global gradient harmonic hastie high icml inductive inference information intel international isbn isoperimetry issue ittycheriah jaakkola jing joachims john johnson journal kernels klein know labeled lafferty language large learning ledge ledoux level ligent limitations linear local machine machines manifolds manning markov maximum mccallum mcdiarmid method methods micchelli minimization mining mitchell model models multi multiple multivariate named natural nguyen nigam nips niyogi oles pages partially performance pierce ponti pontil predicting prediction predictive press probability problems proc proceedings processes processing random recogintion recognition references regression regularized relatedness research responses retrieval riemannian risk rivaling robust scale schlkopf schuller semi sense series sixteenth smarr solving sons spaces special springer statist statistical statistics stochastic structures supervised support surveys system szummer talagrand task text theory through thrun tibshirani training transductive university unlabeled unsupervised using vaart value vapnik vector verlag walks weak wellner weston wiley with word yarowsky york zhang zhou http://www.jmlr.org/papers/volume7/wu06a/wu06a.pdf 124 A Direct Metho d for Building Sparse Kernel Learning Algorithms advances algorithms analysis approach bakir bayesian bennett bfgs bhattacharyya bishop bousquet brabanter building burges cambridge chapelle chicago choosing choudhury classification classifiers coefficients combining conference constructing cristianini data decision descriptive differential discriminative dubeau editor editors edward feature features first function gaussian gauvin gestel ghahramani ghaoui greedy hill ieee information inputs intel international jordan journal kaufmann keane kernel kernels lanckriet large learning least ligence limited linear lkopf machine machines mangasarian margin marginal math mathematical mcgraw memory mercer method methods mika minimax mining moor morgan mueller mukherjee multiple nair nature networks neural nocedal nonlinear optimization oxford pages parameters pattern platt press proc proceedings processes processing programming properties pseudo raedt rayleigh recognition reduced references regression relevance research robust rsvm rules saitta scale scholkopf scientific shawe siam simplified singapore smola snelson some spaces sparse sparseness springer squares statistical steinwart study support suykens systems taylor theory tipping transactions tsch university using vandewalle vapnik vector verlag weiss weston with world wrobel york zoubin http://www.jmlr.org/papers/volume4/claveau03a/claveau03a.pdf 266 Machine Learning with Data Dep endent Hyp othesis Classes abdallah ability academic advances algorithms also ambartzumian analysis annals anthony apparently appeared applications applied approximation ariola artificial automatic bartlett based berkeley boucheron bound bounded bounding buescher calculus california cambridge cannon canonical catch central certain characterizing choice class classification classifiers colt company complexity computational computer computers concerning control convex cortes cover cowen cristianini data densest dependent devinney devroye dietterich digraphs dimension dimensional discussion disjoint distances distribution domination dorato dorit dudley editors electronic error estimates estimation ettinger factorization forthcoming freund generalization geometric geometrical gyorfi hard have hemisphere hierarchies high hochbaum householder hulls hush ieee improved inequalities inequality information intel international introduction john johnson journal kaufmann kernel koltchinskii kumar learning leen letters ligence limit linear lugosi machine machines marchette margin massart mathematics methods minimization model morgan mutual networks neural neurocolt number other over panchenko paper part pattern penalties plane points preparata preprint presented press priebe probabilistic probability problem problems proceedings processing properties psychometrika publishing rademacher random real recognition references replacement report risk rogers sample samples sampling scale scholkopf science scovel selection self serfling series sharp shawe shiryaev shorter smooth soft sons springer statistical statistics structural structures study support symposium systems taylor technical terms that their theorems theoretical theory this transactions tresp trick uncertain uniform university vapnik vector verlag version wiley williamson with without workshop york young http://www.jmlr.org/papers/volume3/nevo02a/nevo02a.pdf 371 On Online Learning of Decision Lists abound aditi algorithm algorithms andrzej angluin annual anselm anthony attribute attributes available avrim biggs blum blumer boolean bounds bshouty california cambridge classes closed communications complexity computational computer concept cruz dagstuhl dana david decision dhagat differences efficient ehrenfeucht eiter electronically finitely foundation functions giessen guijarro haussler hellerstein helmbold http ibaraki ieee infintely informatics information institute intersection introduction irrelevant jection journal june kaufmann kazuhisa lavin learnable learning leslie letters line linear lisa lists littlestone littlstone logarithmic machine makino manfred many march martin mexico mistake mistakebound models mona monotone morgan nader nested nicholas norman occam online pages presence press proceedings processing queries query raghavan rank razor references related report research rivest robert rocco ronald sample santa science sciences servadio singh sloan subclass symposium system term terms theoretical theory thesis third thomas threshold toshihide trees university valiant victor vijay warmuth when with workshop http://www.jmlr.org/papers/volume6/tsuda05a/tsuda05a.pdf 64 Matrix Exponentiated Gradient Updates for On-line Learning and Bregman Projection abound action additive advances akaho algorithm algorithms amixture annual application applications artificial asai attributes australia auxiliary barndorff based batch becker bioinformatics biological boosting bounds boyd bregman brodley california cambridge censor chuang clustering cohn common comparison completion computation computational computer conference confidence convex cruz data decision descent distance divergences editor editors efficient entropy estimation exponentiated finding first freund from function functions gaussian generalization generalized germany gill golden gradient helmbold helmholtz huber icann icml improved inference information international interval irrelevant iterative john joint jordan journal july jupp kearns kernel kernels kivinen kwok lafferty learning lent line linear littlestone logarithmic long loss lower machine margin mathematics matrix maximization maximizing method metric metrics mistake mixtures models multidimensional networks neural nielsen nips noble obermayer october online optimization pages parameter phys physics point potsdam predictions predictors press problem problems proceedings processing programming projection pseudo quantum rated ratsch references regression relative relaxation report research robust russell santa schapire scholkopf sciences sets shai shwartz side singer smola solla solution sons springer statist statistical statistics submitted suppl sydney system systems technical theoretic theory thesis threshold thrun tsang tsuda twenty ucsc university using ussr vandenberghe verlag versus warmuth when wiley with xing york http://www.jmlr.org/papers/volume3/crammer03a/crammer03a.pdf 378 Ultraconservative Online Algorithms for Multiclass Problems abound adaptive adatron additive advances algorithm algorithmic allwein analysis anlauf annual approach approximate artificial attributes bakiri based biehl binary brain breiman brooks burges campbell classification classifiers clss codes computation computational computer conference convergence correcting cortes crammer cristianini design dietterich discriminant duda editors error europhysics exponentiated fast fifteenth foundations freund friedman friess general gentile gradient grove hart helmbold hill implementation information intelligence international irrelevant january jornal journal kaufmann kernel kivinen large learnability learning letters linear littlestone long machine machines margin maximal maximum mcgraw mesterharm methods minimal model morgan multi multiclass networks neural neurocomputing nilsson olshen online optimization organization output pattern perceptron platt prediction press probabilistic problems procedure proceedings processing programs psychological quinlan reducing references regression related relaxed reprinted research results review rosenblatt scene schapire scholkopf schuurmans sciences september sequential simple singer smola solving statistical stone storage support system systems theory thirteenth threshold trainable training trees ultraconservative unifying updates using vapnik vector versus wadsworth warmuth weak when wiley winnow york http://www.jmlr.org/papers/volume4/langseth03a/langseth03a.pdf 270 Learning Equivalence Classes of Bayesian-Network Structures aaai able about above absence acyclic acyclicity added adding addition additional additions adjace adjacencies adjacency adjacent admit admits after again akaike algorithm algorithms allen allow allows along already also although always ambrosio among analogous ancestors andersson annals anytime append application applications applied applies apply applying approach approximating arbitrarily argue argument arguments arjan artificial assume assumption automatic averaging away background base based basis bayesian because become been before behavioural being belief belong belongs below bendou besnard between blair bonissone boston both bouckaert break breaking breese broken buntine cannot cardinality carnegie case causal causality causation cercone chain change changed characterization characterizing chickering child chooses choosing chord chordal chordality chosen chow class classes clearly clique closest cognitive combination combining common communications comp compel compelled complements complete completed completely completeness completes complexity component components computations computer computing conclude condition conditions conference configuration conjunction connect connected connecting connects consecutive consequence consequently consider considered considering consist consistent consisting consists constitute constitutes construct constructed constructing construction contain contained containing contains continue contradicting contradiction contradicts control convenient converging convert cooper corollaries corollary correct correctly correspond corresponding could covered create created criterion current cycle cyclic dags dash data databases decision decomposable decreasing deduce defined definition definitions delete deleted deletes deleteu deleting deletion demonstrate demonstrating denote department depend dependence depends derives descendant describe described description desired detailed detecting determine determining developed difference differs difficult digraph digraphs dimension direct directed directing direction directions directly directs disambiguates discovery discrete disjoint distributions does druzdzel each easily easy ecall edge edges edition editor editors efficient either element elements eleventh eliminating elled else emma emphasize empty endix ends engineering enote enough enter entirely enumerating equal equivalence equivalent erator erators essentially estimating every exactly example except exisl exist existed existence exists exit experimental explanation extension extensions extra extract extracting extracts fact favor fayyad fewest fifteenth fifth figure fikes final finally find first fisher flexibility follow followed following follows form forms found foundations four framework frequently from furthermore geiger general generality gets gillispie given glymour goldszmidt graph graphical graphs guarantee guaranteed guarantees guide hanks have heckerman hence henrion heorem herskovits higher hold holds horvitz however hybrid hypergraphs hypothesis identical identification identified identifies identify identifying ieee immediately implies imply important impossible incident include including inclusion increase increases independently individually induction inference inferred information initial initially input insert insertd inserted inserting insertion insertions insertu instead intel intermediate international into introduction invalid inverse involve involves iteratively itself iven jensen jordan journal kanal kaufmann kluwer know knowledge known kocka koller label laboratory largest laskey last later leads learning least lecture ledge lemma lemmas lemmer length lenz less lest ligence ligent linear literature local locally long longer look loss lower machine made madigan main makev mamdani mark markov match mateo matrix maximal maximum mean means meek mellon member members method methods microsoft middle minimum mining model modeling models modification more morgan most msrtr munteanu must nato necessarily necessary need neighbor neighbors neither network networks never next ninth node nodes none notation note notes noting number numbered obtained once only operation operator operators opposite order ordered ordering orderings organized orient orientation orientations oriented orienting orients orollary osition other otherwise output over pages pair paper parent parents part partially participate participated participates participating particular pass passes path paths pdag pdags pearl perform perlman peyton phase place plausible point possibility possible potentially prade precedes precisely prediction preliminary presence presented press previous previously principle probabilistic probability procedure proceed proceedings proceeds process processed processing produced proof proofs prop properties property proposition propositions prove proved proven proves provide proving prow purpose purposes reachable reason reasoning recall reduce refer references refinement regard related relation relies remain remainder remaining remains removed repeated repeatedly replacing report representation requirement requires research respectively restriction result resulting results returns reversal reversed reversible reversing rissanen root rule rules same sandewall satisfying scheines schwarz science sciences score scored scoring search second section sections segment segments select selected selecting selection selectively semi separate sequence series sets seventeenth seventh shachter share short shorter shortest show showing shown siam similar similarly simon simple simultaneously single sition sixth size skeleton smets social some sound sparse specifying spirtes springer staring state statement statistical statistics step still stochastic strict structure structures studeny subgraph subset subtracting such sufficient supp suppose supposition suzuki symmetric synthesis systems table tail takes tarjan tarsi technical terminates test tests than that them themselves then theorem theoretic theory there therefore these thesis they third this those three through throughout thus ties time together topology total transactions transform transformational transforms transition transitivity trees triangle trivial trivially true turn turns tutorial twelfth ucla uivalent unaffected uncertainty understood undirected undirecting unfortunately university unprocessed until used useful uses using uthurusamy valid validity variables various verlag verma vertices violated violating volinsky volume vstructure vstructures well were when whenever where whether which while whose will with within without work workshop would yannakakis yielding yields ylthat york zero http://www.jmlr.org/papers/volume4/stainvas03a/stainvas03a.pdf 254 Supp ort Vector Clustering academic advances alex algorithmic algorithms analysis annals approaches based berkeley bernhard biocomputing biology blake blatt bottleneck burges cambridge chichester classification cliffs cluster clustered clustering clusters computation computational conference cooper current data databases description detection determining diego dietterich dimensional discovering distribution domain domany dubes duda duin editors elisseeff englewood estimating eugenics examination expression eytan fast fisher fletcher fukunaga gene granular guyon hall hart high horn information international interscience introduction irregular jain jiang john kernel klaus learning leen letters lipson machine machines macqueen magnet marcelo markovian mathematical measurments merz method methods milligan minimal model muller multiple multivariate nature networks neural neurons novelty number observations optimization order pacific parametric pattern platt practical prentice press probability problems proc procedures proceedings processing psychometrika recognition references relaxation repository ripley robert roberts sara scholkopf sequential shai shamir shapes sharan shawe siegelmann slonim smith smola solla some sons springer stability statistical statistics stork structure support symposium systems taxonomic taylor theory thomas tishby todd topics training tresp university unsupervised using vapnik vector volker wiley williamson wiseman york zhang http://www.jmlr.org/papers/volume3/sugiyama02a/sugiyama02a.pdf 373 The Subspace Information Criterion for Infinite Dimensional Hyp othesis Spaces academic active adaptation advances akaike albert algebraic algorithm algorithms amari american analysis annals application applications applied approximation aronsza artificial asymptotic atomic ators automatic automatica available bartlett based basis bayes bayesan bayesian berlin bernardo between beverly beyond bias bienenstock biometrika bishop bootstrap bounds bunke burges cambridge camp center central chapelle chen cherkassky choosing clarendon classification clustering cognitive cohn combination comments communications completely complexity component computation computing conference connection continuous control correct corrections correspondence craven cristianini criteria criterion cross curves data daubechies decision decomposition decompositions degree degroot delve dependencies description design determining devroye dilemma dimension dimensional discovery distribution donoho doursat droge dual dubes edinburgh editor editors efficient electronically empirical equivalence error estimates estimating estimation estimators evaluation expectation experimental experiments extension fedorov filters finite fisher four from fujita fukumizu function functions fundamentals further gammerman gaussian geman generalization generalized ghahramani gibbs girosi gpros graphical greedy hall hasselmo henkel heskes hidden hills hinton http huber hunter hypothesis icml ideal ideals identifiable identification ieee ieice image imai imaizumi implementation incremental industrial inference infinite information institute intel international interpolation intro introduction inverses issue jain japanese jective john johnstone jolliffe jordan journal kernel kernels kimeldorf kitagawa know kohonen konishi kustra laplace lasso learning lectures ledge letters ligence likelihood lindley linear london look lugosi machine machines mackay mallows manual maps mathematical mathematics mathematik matrix meeting method methods mika mining miyakoshi model modeling models modified moore mozer mulier muller multilayer multiple murata nature neal network networks neural nishii noising noisy norm number numerische numo observational ogawa oper operators optimal optimizing organizing other oxford pages papers parameter parameters pattern pennsylvania penrose perceptrons philadelphia prediction prentice presented press principal prior probabilistic procedure proceedings processes processing properties pruning pseudoinverse publication pursuit quadratic radial rasmussen ratsch recognition references regional regression regressors regularization regularizers report reproducing research restoration revow ridge rissanen robust royal sage saunders schatten scholkopf schwarz science sciences scientific selection self series shao shawe shibata shinomoto shirinkage shortest shrinkage siam signal significance smith smola smoothing society soft some sons spaces sparse spatial special spline splines springer statistical statistics stochastic subset subspace sugiura sugiyama support survey systems takeuchi tanaka taylor technical technometrics term tests their theoretic theoretical theory thesis thresholding tibshirani toronto touretzky transactions tsuda tutorial types unified units university using valencia validation validity vapnik variables variance vector verlag volume vovk wahba watanabe wavelet wavelets wiley williams williamson with yamanishi york yoshizawa http://www.jmlr.org/papers/volume7/gardner06a/gardner06a.pdf 140 One-Class Novelty Detection for Seizure Analysis from Intracranial EEG abrupt acoustics adaptive advances ajmone alessandro algorithm algorithms analysis annual anuzis apparatus application applied approach artificial atlanta automated automatic back based behavior bennett biology biomedical boser brain burges campbell changes channel class classification classifiers clinical comparison computational computer conference counteracting data davy dept description desobry detecting detection device dietterich dimensional discovery dissertation distribution document echauz ecog editors efficient electrical electroencephalography energy engine engineering epilepsia epilepsy epileptic esteller estimating evaluating evaluation excess explorations feature fifth frei from gabor gardner georgia gibb giftakis gotman guyon hall hallelujah hayton herzog high hype icassp ieee ieeg implantable improvements information ingram institute international intracranial introduction jersey jessel journal kandel kernel khorasani kishawi knowledge krieger lancet laskov learning leen length lesser like line liss litt lkopf ller long longterm machine machines manevitz margin marsan medicine methodology microsoft mining monitoring nature network networks neural neurology neurophysiology novelty ongoing online onset optimal ordon organizing osorio other part patent patient patients pattern performance peters pittsburgh platt pless possible precursors prediction prentice press prevent principles prior proceedings processing quantitative ratsch real reassessment recognition redmond references research richardson risch rise risinger sackellares schaffner schwartz science seizure seizures self series shawe short sigkdd signal signals smola society spatial specific spectra speech springer statistical strategies structure suitable supplement support svms synergy system systems tarrasenko taylor tcheng technology temporal term theory time training transactions tresp tsuda turnbull tutorial using utility vachtsevanos validation vapnik vector verlag vibration viglione walczak warning webber weng wennberg western wilkinson williams williamson wilson with workshop york yousef zaveri http://www.jmlr.org/papers/volume4/almeida03a/almeida03a.pdf 246 Prior Knowledge and Preferential Structures in Gradient Descent Learning Algorithms abbot abhandlungen academic academie adam adaptation adaptive addison advances akaike albert algorithm algorithms amari analysis annual applications approach approximate artificial arun barndorff bayes bayesian berlin bernhard blind boca boothby bounds bruno california cambridge carnegie chang chapman choice christian christoffel clarkson classification claudio cliffs colt comprehensive computation computational computer concept conditioning conference continuous convergence courbure crel cruz curvature curved dale data david dawid decision delaware department dependent descent differentiable differential differentialausdrucke dimensions discrete discriminant disintegration distributions douglas dover duda dynamical edition editions editor editors edwin efficiently elwin englewood english euler exploiting exponentiated extended feedforward filtering filters fine flache flatland from fundamentals gauss general gentile geoffrey geometrical geometrie geometry germany gesel gordon gottingen grades gradient graduate grove grunde hall hart hartigan hassoun haykin helmke hermann hill hirotugu hoffmann homogenen hopkins http hypothesen ichi idee ieee improper information international interpretation introduction investigations isspa jagota john johns johnson jordan jorgensen joseph journal jyrki karl kearns kivinen knowledge kong learning leonhard liegen liepzig likelihood limits linear littlestone london loss lschaft machine mahony manfred manifolds many march markov martin massachusetts material mathematics mccleary mellon methodology methods michael models mohamad moires moore multidimensional murray natural neelandica network networks neural nick nielsen niglichen nonlinear norm optimal optimization pages parameter parametric pattern perish peter phil point pollard prediction predictors prentice preprint prequential press prior probability problems proceedings processes processing proper publish published pubn pubs raton raven reading recherches references regression regret relative results revised rgen rice richard riemann riemannian riemannschen robert robustness romance royal santa sara schuurmans science sciences scott separation september series sethares shiryaev shun signal simon society solla solo solutions source spaces sparsity spivak springer statistica statistical statistics stork structured submitted surface surfaces symposium systems tenth terrence teubner texts theory thesis thrift time toward tranlated transactions transfmoration translation twelfth uber ucsc unabridged university unsupervised updates verlag version versus victor view volume volumes warmuth welche wesley weyl wiley william williamson wilmington wissenschaften with works xuan york zweiten http://www.jmlr.org/papers/volume3/fung02a/fung02a.pdf 372 Minimal Kernel Classifiers active adaptive advances alex algorithms analysis annals applications applied approximation april arbitrary arlington asscociation august austria bartlett bell bernhard blumer boston bottou bounds bradley building burges california cambridge case chapelle checker checkerboard cherkassky chicago citeseer clas classification classifier classifiers clustering company comparison complexity computational computer computing concave concepts conf conference control convex cortes cplex cristianini dantzig data dataset denker department dietterich digit discovery discrimination drucker edition editor editors ehrenfeucht expectation extensions feature ferris festschrift field fifteenth first fischer francisco from fung gaussian generalized greedy grossman guide guyon handwriting haussler heidelberg herbrich hill html http icml ieee incline incremental information informs institute integer intel interior international introduction isbn jaakkola jackel jectable jersey john jordan journal july kaufmann kernel klaus kleinberg know labs lagrangian large learn learning least leave lecun ledge leen letters ligence linear machine machinery machines madison mangasarian mannila margin massive mateo math mathematical mathematics mathworks matlab matrix mcgraw mean methods meyer minimal minimization minimum mining mitchell mixed mlearn mlrepository modeling models morgan motwani mozer mpml mulier muller munson murphy musicant natick nature neural nevada nonlinear norm november occam october operations opper optimization optimizer pages paral parsimonious pattern perturbation petsche philadelphia physica plane point polyhedral press princeton probabilistic proc proceedings processes processing prog programming programs provost proximal pubs ramakrishnan razor recognition reduced references regression relevance report reports repository research riedmueller rissanen ritter robust rockafellar rosen rsvm sackinger scale schaeffler scholkopf schuurmann schuurmans sciences second selection separating shavlik shawe siam sifiers simard smola smooth society software solla solutions sons sparse springer srikant ssvm statistical statistics stochastic stolfo street study support systems taylor tech technical techreports theory thomas tipping todd transactions tresp university user using vapnik vector verlag version vienna village virginia volker volume warmuth weston wiley winther wisc wisconsin york http://www.jmlr.org/papers/volume7/scheinberg06a/scheinberg06a.pdf 183 An Efficient Implementation of an Active Set Method for SVMs academic active advances algorithm algorithms annual applications arising balcazar baltimore bixby blake boser burges cambridge case cauwenberghs certain chapter classification classifiers combining computational computations computer computers computing conference convex cristianini data databases decremental department dietterich dual edition editor editors efficient entire extension extensions fast ferris fine fletcher forrest foundations frangioni freund general girosi goldfarb golub gregory guyon hastie haussler hopkins html http idnani ieee improved incremental information institute interior international introductin joachims johns joint journal kaufman kernel kernelbased kiwiel large learning leen library linear loan london lootsma lustig machine machines macines madison making margin marsten massive mathematical mathematics matrix meeting merz method methods mininal mining mlearn mlrepository munson national networks neural newton nlearning nocedal nonlinear nonsmooth numerical numerically olkopf operations optimal optimization orsa osuna other pages path platt poggio point positive practicle presented press problem problems proc proceedings processing programming programms provably quadratic rank references regularization report repository representations research rosset scale scheinberg scholkopf sciences scientific semidefinite sequential shanno shawe siam signal simplex smola solving springer sringer stable statistical strictly study subroutines support systems taylor technical theory thomas tibshirani tims todd training trainning tresp trining university using vanderbei vapnik vector verlag very volker watanabe wisconsin with within workshop wright york http://www.jmlr.org/papers/volume5/mangasarian04a/mangasarian04a.pdf 15 Knowledge-Based Kernel Approximation aaai academic adaptive advances advice algorithmica anouar applications approximation approximations archives aspirates august bartlett barto based baudat becker berlin best breast burges cambridge cancer cherkassky classification classifiers colt computational computer computerized computing concepts conference control convex creating data department deutsch diagnosis discrete drucker edited edition editor editors eighteenth emphasis estimation evgeniou fine from function fung generalized golowich guiding heisey hilbert html http icml ifip information initial inner institute international interpolation introduction inversion john joint jordan jose journal july kaufman kearns kernel kernels klaus knowledge kuhlmann language large learner learners learning linear machine machines maclin madison mangasarian march margin math mathematical mead method methods micchelli minimization mining mlearn mlrepository mooney mozer mulier murphy musicant natural nature needle nelder networks neural nonlinear obermayer october operations operator optimal optimization pages patients pattern petsche philadelphia poggio pontil press prior proceedings processing product prog prognosis programming recognition references regression regularization reinforcement report reports repository research results robocup scale scaling schoenberg scholkopf schumaker schuurmans sciences sebastian second shavlik siam signal simard simplex smola smoothing soccer solla sons space spaces special splines springer statistical stone street subset supervisory support surgery survival sutton suzanna systems taking talks tech technical theory thrun time toward university utreras vapnik vector verlag volume warmuth washington wiley williams wisc wisconsin with wolberg workshop york http://www.jmlr.org/papers/volume6/steinwart05a/steinwart05a.pdf 38 A Classification Framework for Anomaly Detection abnormal advances aggregation aided algorithm algorithms amer anal analysis annual anomalies anomaly anuzis applications applied approach approaches aproach artificial aspects assoc astley based bennett brady cambridge campbell carr cerneaz chan chang chow classification classifiers cluster clustering clusters communications comput computat computation computer concentrations conditions conference consistency contour control convex cooper cowley cuevas dagupta data david dempster density desforges detect detection detectors dietterich dimensions discrimination distribution distributions duda dudley editors engine engineering engineers estimates estimating estimation evolvable examples excess fast fault febrero fraiman from further gaul gaussian genetic gfkl gonzalez hart hartigan hayton high highdimensional http icdm icpr identification ieee images incomplete influence inform information ingo institution integrated integrity international intrusion intrusions jacob kernel kernels king knowledge known laird lanl learn learning leen level levels library libsvm likelihood lindenbaum linear mach machines magazine mammen mammograms manikopoulos markou mass masses maximum measuring mechanical methodology miller mining modality monitoring muller multi multimodality multispectral nairac negative network networks neural nonparametric novelty optimal organization other page pages papavassiliou paradigm part parzen pattern pebl pfeifer plant platt polonik practical press probability proc proceedings processing programming publications rates real recognition references regularized resampling review ripley royal rubin sawitzki scholkopf science scovel selection sets shawe signal singh smola smooth society spectra spie springer statist statistical steinwart stolfo stork submitted support system systems tarassenko taylor teacher techniques tests theiler their theoretical theory townsend trans tresp tsybakov university unknown using utete valued vector vibration wiley window with without yeung york http://www.jmlr.org/papers/volume6/murray05a/murray05a.pdf 57 Machine Learning Methods for Predicting Failures in Hard Drives: A Multiple-Instance Application aaai abrera advanced advances against allan alon always american analysis andrews annals application applications applied approach approaches approximate april artificial association asymptotically attributes august autoclass awareness axis based bayes bayesian behrens bendat bengio berlin better bickel biometrics biometrika bled bratko bridge brunner bulletin burges business cambridge catlett changing chapter charles cheeseman cherkassky christopher classification classifier classifiers clinical comparative comparison comparisons computation computer concepts conference constructing continuous current daniel data december delgado department dietterich dietz discovery discrete discretization discriminative disk diskdrive distributions doksum domingos dortmund dougherty drive drives dzeroski econometrica edgar edition editor editors effective eibe eighteenth elkan emerson epidemiology estimating estimation european event events exact failure failures faming features filip fisher fourth francisco frank from generative global good gordon greenland greg hadley hafner hall hamerly hamilton hard henry hettmansperger hirsh hofmann holden hughes icann ieee illinois impact importance improved increasing individual inference information instance institute intelligence international into ioannis istanbul jacquelin jean jordan joseph journal julius junan june kaufmann kendall kenneth kernal killeen klotz knowledge kodratoff kohavi kreutz langley larger lathrop lazy learning lehman lehmann liang lippencott logistic loss lozano machine machines madan maintenance making mann manual march master matching mathematical maurice mean mehta menlo methods michael mining modern monotone morgan moses motor mulier multiple multivariate munzel murray mysvm naive narayana nature network networks neural nips nonparametric nonparametrics note obtaining october optimal optimality ordered orlitsky other outcomes pagano pages parallel parametric park pascal patel pattern pazzani pedro perez permutation permutational pharmaceutical philadelphia physicians piersol planning polynomial power predict predicting prediction predictive prentice press preusser probabilities probability problem problems proceedings process processing publishers puri pursuit random randomization rank ranking ranks rare raven recognition rectangles references regression relevance reliability report research restricted richard river rothman rules ruping saddle sahami samples sampling santhanam sawilowsky science senchaudhuri september sequences series session seventeenth shifts signature significance simple sixteenth slovenia small solving sparse springer stanford statistic statistical statistics stefan stochastically stuart study stutz supervised support systems tables technical technology test tests than theodossiou theory thesis thomas ties time timothy tipping tomas tool transactions trend tritchler tsochantaridis turing turkey tutorial ullrich under unit university unsupervised upper using vapnik variables vector verlag vincent vladimir volume waikato wang warnings weiss whether whitney wilcoxon wiley with witten working york yoshua zealand zero zhang zucker http://www.jmlr.org/papers/volume1/allwein00a/allwein00a.pdf 338 Variable Selection Using SVM-based Criteria adaboost advances algorithm analysis annual appear applications arakotom artificial backward barnhill based behaviormetrika bengio biowulf boser bounds bousquet bresler cambridge cancer chapelle choosing classification classifiers colt computation couvreur cristianini criteria duan editor elisseeff error evaluation expectation feature gallinari gene gradient greedy guyon haussler http hyperparameters information informations insa intelligence introduction john journal keerthi kernel kohavi learning leray linear machine machines margin margins matrix measures methods models mukerjhee mukherjee muller multiple networks neural neurocomputing norm onoda optimal optimality optimization pages parameters perception performance pittsburgh poggio pontil press problem processing rakotomamonjy ratsch references report rouen scholkopf selection shawe siam simple soft statistical subset support svms systeme systems taylor technical theory training tuning univeristy using vapnik variable vector weston wiley with workshop wrappers http://www.jmlr.org/papers/volume5/devito04a/devito04a.pdf 23 Some Properties of Regularized Kernel Methods academic advances algorithms american analyse analysis applied approximation arsenin artificial associes asymptotic bartlett bayesian berkeley berlin bertero best between bias bounds bousquet burges cambridge chan chicago choices classification comp complexities complexity computation computational computationals computations computer concentration conference consistency convergence convexe convexity correspondence creation crisp cristianini cucker data dealing december department dept dietterich dimensional early editor editors ekeland empirical equivalence espaces estimation estimators evgeniou examples foundation foundations from function functions gacv gauthier generalization generalized geometric germany girosi helmbold herbrich hilbert hilbertiens http identification ieee ifac infinite information institute intelligence introduction jersey jordan journal kernel kernels kimeldorf kluwer knowledge laboratory large learning lectures likelihood linear local machine machines margin massachusetts math mathematical mathematics mathematique mcauliffe mendelson methods micchelli michael minimization models mukherjee necessary networks neural neurocomputing niranjan niyogi note notices noyaux number observational olivier optimization other pages paper parameters paris penalized peter piscataway poggio pontil posed prediction press prior problem problemes problems proceedings processes processing publishers rademacher rakhlin randomized references regularization regularized related report representer reproducing research rifkin risk scattered scholkopf schwartz science separable series shahar shawe siam signal smale smola smoothing society solution solutions some sous spaces sparse sparseness splines springer stability stat statistical statistics steinwart stochastic submitted sufficient sullivan support symposium system systems taylor technical technology teman theorem theorems theory tikhonov topologiques torre transactions tresp turnbull twelfth uncertainty uniqueness university valued vapnik variance variationnels vector vectorials verri villards virtual vision volume wahba washington wiley williamson winkler winston wisconsin with workshop york zhang http://www.jmlr.org/papers/volume7/peer06a/peer06a.pdf 110 MinReg: A Scalable Algorithm for Learning Parsimonious Regulatory Networks in Yeast and Mammals active alon altschuler analysis analyze annual antonio april architecture armour artificial auctions bacteria ball barash bard barkai bayesian bennett bergmann between binkley biocomputing bioinformatics biology biotechnology botstein bourne brown bussemaker campbell carmel causal cell cells cerevisiae chakraburtty chandy changes cherry chickering choi church classification clusters code coffey coli collado combination combinatorial combining commerce compendium computational condition conference consortium cooper correlation current danford dasgupta data database davierwala decamp decreasing detection determination discovery dolinski dordrecht duda dwight editor eisen electronic element elidan environmental escherichia estimation eukaryotic experimental expression fifteenth fraenkel framekwork friedlander friedman friend from function functional gachotte gasch geiger gene genes genet genetic genetics genome genomewww genomic gerber gifford glasner global gordon goto graphical hannett harbison harel harris hart hartemink hartwell heckerman herskovits hopfield hsueh http hughes identifying ihmels imoto induction inferring intelligence jaakkola jennings john jones jordan joseph journal kellis kidd king kluwer knowledge koller lander large learning lehmann leibler linial location lymphocytes machine macisaac mangan marginal martinez marton matese method meyer microarray microbioly milo minreg mixture miyano mock modelling models modular module modules molecular motifs murray nachman nature netherlands network networks nisan nonparametric observational odom ontology opinion organization overlapping pacific page pages pathway pathways pattern perturbed pokholok polytrees prediction principled probabilistic proceedings profiles program promoter recomb references regev regression regulation regulator regulators regulatory relationships research response revealing reynolds rinaldi roach robert roberts robinson rolfe rourke saccharomyces sambrano sarig scale scene science segal sequence series shapira shen sherlock shoemaker siggia signal simon sinkovits slade sons specific spellman stanford statistical stepaniants storz stoughton structures subnetworks suppl symposium systematic tagne takusagawa tanay tavazoie tenth their thompson thorsson time tool transcriptional transduction tutorial uncertainty unification unravelling using utilities verghese vides volkert weng wiley with wyrick yeast york young zeitlinger http://www.jmlr.org/papers/volume7/keerthi06a/keerthi06a.pdf 158 Building Support Vector Machines with Reduced Classifier Complexity accuracy adler advances algorithm annals approach approximation arenas asilomar assp asymptotically bach bakir bartlett based basis bayesian benchmark bengio bennet bennett berkeley boosting bounds building burges california cambridge chapelle cholesky classification classifiers coefficient column comparison complexity computation computer computers conference controlled convex data decomposition decoste delgado department dictionaries downs embrechts exact fast feature figueiras finite first forward fraunhofer frequency friedman function garc gates gaussian generation germany girosi gradient greedy growing herbrich hernandez heterogeneous http icml ieee improving informative ings input international invariant iterative jimenez joachims jordan journal keerthi kernel kernels knirsch kreutz large lawrence learning linear machine machines making mallat mangasarian margin mark massachussetts masters matching mathematics matrix meijerink method methods microsoft mika minimal mining mixture models modified momma mora muller naviai neural newton nips optimal optimization osuna pages parrado pattern philadelphia platt potsdam practical predictive press primal proceedings proceedt process pursuit pursuits quadratic raetsch rank ratsch recognition redmond reduced reducing references regression relevance report repository research robust rsvm scale scholkopf science second seeger selection sequential sharp siam sigkdd signals simplification smola software solution solutions some space sparse sparseness speed statistics steinwart study submitted support svms symmetric systems technical thesis thies time tipping training transactions twenty university updates vazquez vector vectors vidal vincent vorst weber which williams with workshop zhang http://www.jmlr.org/papers/volume7/demsar06a/demsar06a.pdf 104 Statistical Comparisons of Classifiers over Multiple Data Sets aaai accuracy advances against algorithm algorithms alpaydin alternative american among analysis annals annual application applied approach approximate approximations arch artificial asia associates association assumption attributes august australia avoid bagging balanced based bayes beck bellazzi belmont bengio between biometrics biometrika biostatistical blake bled bonferroni boosting bouckaert bratko brazdil brighton brodley calibrated california case chambery chapman choosing class classification classifier classifiers cleveland clifs cohen combined combining communications comparing comparison comparisons computation computational computer conf conference continuous control course crafting critical cross cumulative curves data databases davenport demsar designs development dietterich discovery discretization distribution document dunn dunnett dzeroski earth ecai edition editor editors englewood erlbaum error escolano estimating estimation estimator european evaluating evaluation experimental experiments faculty fawcett fayyad feature fidel fifteenth first fisher fold france francisco frank free friedman from galindo generalization grandvalet grobelnik guerrero hafner hall hamilton handbook harlow hochberg holland holm hommel honghua html http hull hypothesis icml idamap iman implicit individual induction inference information ingwersen intelligence intelligent interactive international interval intl irani iwann jersey joint journal july june junquera kaufmann knowledge kohavi langley lawrence learning ljubljana locally machine mathematical means measures medicine merz methods mining mishra mladenic mlearn mlrepository model modern modified morgan mulaik multi multiboosting multiple nadeau naive national natural nemenyi networks neural neurocomputing nonparametric normality notes november orange pacific pages pairwise pakdd paper papers parametric pathol pedersen performance pharmacology pitfalls pizarro portland position prentice press princeton problem proc procedure procedures proceedings processing provost psychological psychologist psychology publishers publishing quarterly quinlan ranking rankings ranks recommended references region regression rejective repeated replicability repository representations research retrieval review riano robust round routing salzberg scandinavian scatterplots schmidt schultz schutze science scientific selection sequentially seventeenth several shaffer sharper shavlik sheskin sigir significance simple sixteenth slovenia smoothing soares springer srikant stagewise stanford statement statistic statistical statistics supervised sydney systems technique test testing tests there thesis thirteenth three treatments tukey twentieth twenty unbalanced unbiased university using validation valued variance vazquez verlag wadsworth wagging washington webb weighted were what white wilcoxon with workshop york zhang zupan http://www.jmlr.org/papers/volume7/sugiyama06a/sugiyama06a.pdf 109 Active Learning in Approximately Linear Regression Based on Conditional Expectation of Generalization Error academic active addisonwesley akaike algorithm algorithms annals applications artificial automatic automatica based beverly biased camp cohn communications computation computer control covariate data delve description design designs dimension electronics estimating estimation estimator experimental experiments fedorov fukumizu function functions fundamentals generalization generalized ghahramani henkel hills hinton http identification ieee ieice improving incremental inference information intelligence john jordan journal kanamori kiefer knuth kustra learning least likelihood linear loglikelihood look mackay manual massachusetts maximum model modeling models multilayer neal networks neural objective ogawa optimal optimization optimum perceptrons planning points polynomial predictive press programming publication pukelsheim rasmussen references regression research revow rissanen robust royal sage sample schwarz sciences selection seminumerical series shift shimodaira shortest significance simultaneous society sons squares statistical statistics sugiyama systems tests theory tibshirani toronto transactions trigonometric under using vapnik volume weighted weighting weights wiens wiley with york http://www.jmlr.org/papers/volume8/raiko07a/raiko07a.pdf 209 Building Blocks for Variational Bayesian Learning of Latent Variable Models accelerating accurate acoustics adaptive advances algorithm algorithms alhoniemi american analysis anderson annals annealed annual application applications approach approaches approximate approximating approximation approximations april arising artificial association astronomical attias australia automatic available back barber based bayes bayesian beal belgium belief berlin best beyond bishop bits blind blocks boca bottou breaking bruges budapest bugs building buntine calculating california cambridge camp canada cardoso carlin carlo carolina cause cell cells chan change chapman charleston choudrey clarendon classification cliffs clusters coding cohn college communications competition complex component components comprehensive computation computational computer computing conf conference continuous control cortex cruz cyclic data dayan december decomposition dempster densities dependencies dept derivation description detection deterministic developments diego dietterich dimension dipl discrete distributions doucet driven dynamic dynamical early edinburgh edition editor editors effect electronics emergence empirical energy englewood ensemble esann espoo estimating estimation european evans everson evidence expectation experts extended extensions factor family feature features february fiala field fifth filtering filters finding finland fischer florida form formation fort foundation framework francisco free freitas frey from fundamentals galaxies game gaussian gelfand gelman geman ghahramani gibbs gilks girolami gordon granada graphical gray hall hansen harmeling harva hasselmo hateren hawaii haykin helsinki hidden hierarchical hierarchy higher hiit himberg hinton honkela honolulu hoyer http hungary hyvarinen icann icassp icci iconip ieee ieice ijcnn ilin image images implementation importance incomplete incremental independent industrial inference information initialisation instantaneous integrals intelligence intelligent international into introduction invariant istanbul jaakkola january japan joint jordan journal july justifies kaban kadane kalman karhunen kaski kaufmann kearns keeping kernel koczkodaj kohonen laird lappalainen latent lauderdale learning lecture leen length letters library likelihood linear local logistic london lundqvist machine mackay magnetoencephalograms maps march marginal markov master matlab maximum mean medical message method methodological methods minima minimization minimizing minimum minka miskin missing mixtures model modelling models moments monte monthly montreal moore morgan mozer multidimensional multilayer multiple multivariate murphy nara natural nature neal networks neural nips nolan nonlinear nonstationarity nonstationary nonsymmetric notes notices november observed optical optimal order organized organizing ostman other oxford pages parameter parhankangas park parra partially passing pattern pearl penny perceptrons pham phase plausible poland populations posterior practice prentice press prieto primary principles priors probabilistic proc proceedings processing projects propagation properties pruning publishers puntonet raiko raton raychaudhury reasoning recognition rectified references regression relational relaxation rensen research restoration roberts robinson roussos rowe roweis royal rubin ruderman sajda sampling santa saul schemes schumann schwarz science sciences scoring scotland searches seattle sejnowski self separation september sequences sequential series shift signal signals similar simple singapore sixth smith society software solla somtoolbox source sources south space spain sparse spatio spectra speech speeding spence spiegelhalter springer state statistical statistics stellar stern stochastic stockholm structure structures subspace subspaces swartz sweden sydney symmetry symp symposium systems technology temporal that their theoretic theory thesis thomas tierney toolbox topographic topography tornio touretzky trans transactions tresp turkey type uller uncertainty undercomplete univ university unmixing unsupervised update using valpola values variable variables variance variants variational various verlag version vesanto view vigario visual volume wallace warsaw washington waterhouse wavelets weights weiss wiley winn winther wisconsin with workshop yields york young zemel http://www.jmlr.org/papers/volume6/micchelli05a/micchelli05a.pdf 68 Learning the Kernel Function via Regularization academic additive advances alberta algorithm algorithmic alignment amer anal analysis annals annual anova appear appl applications applied approximation argyriou aronszajn aubin augmentation australia bach banach banff bartlett based basic bayesian becker behavior bennett bertinoro between borwein boston bounds bousquet bull cambridge canada canadian caponnetto chapelle choosing city classes classics classification classifiers college colt combinations company completely completion complexity component computation computational computer computing conf conference conic consistency continuously convex cosso covering cristianini cucker data decomposition definite department dept description diego dietterich discrimination domain dual duality duin economic edition elements elisseeff empirical entire entropy equivalence error estimating evaluating evgeniou examples fast flexible formulation foundations friedman function functions game gaussian gaussians generalization ghahramani ghaoui girosi graepel hasselmo hastie herbster herrmann hoffmann holland hong hyperkernels icann icml inequalities inference information institute interpolation italy iterates jaakkola january jebara jersey jordan july june kandola kaufmann kernel kernels kimeldorf kong kwok lanckriet learnability learning letters lewis linear london loss machine machines macmillan majorization mangasarian marshall math mathematical mathematics matrix maximum meila methods metric micchelli mimeo minimization mining models monotone morgann mozer mukherjee multi multicategory multiple ncsu necessary networks neural niyogi nonlinear north note number obermayer observational october ohio olkin online operators optimization pages parameterized parameters path paths pattern philadelphia piana poggio polynomial pontil prediction predictions preprint press princeton proc processes processing programming properties publishing pure rasmussen rates real recognition references regression regularization regularized relative report representation reproducing research results rifkin risk risks rockafellar rosasco rosset royden sammut saturation schoenberg scholkopf science scovel selection semi series sharpley shawe siam smale smola smoothing society some spaces sparse spline splines springer stability stanford state statis statistical statistics steinwart structured studies sufficient support sydney systems tanner target taylor tchebycheffian technical theory thesis thibaux thrun tibshirani time touretzky trans transductive university using valued vapnik variance variances vector verlag verri vito wahba wiley williams williamson with wong yeung ying york zhang zhou http://www.jmlr.org/papers/volume8/ying07a/ying07a.pdf 212 Learnability of Gaussians with Flexible Variances adapting advanced advances aggregation algorithm alon american analysis annals annual anthony appear applications applied approximation aronszajn artificial association banach barron bartlett based basis behavior between bianchi bounding bounds bousquet bulletin cambridge campbell cantelli capacity caponnetto central cesa chapelle chen choosing classes classification classifiers cohn combined complexities complexity computation computational conference connections consistency convergence convex convexity cortes covering cristianini cucker data david density dependent devroye differentiability dimensions discovery distributions dudley dynamically edition empirical error esaim estimating estimation evgeniou fast forthcoming foundations from function functional functions gaussian gaussians generalization generalized ghaoui gine girosi glivenko gyorfi harmonic haussler herbrich hierarchies http hypothesis ieee influence information integrals intelligence isoperimetry jersey jetter john jordan journal kearns kernel kernels knowledge koltchinskii lanckriet learnability learning least lecture lectures ledoux limit lugosi machine machines manifolds margin mathematical mathematics matrix mcauliffe mendelson methods minimization mining mixture model models mukherjee muller multi multiple networks neural niyogi notes number observational optimal over panchenko parameters pattern penalties poggio point polynomial pontil press princeton probabilistic probability proceedings processes processing programming properties pugachev rademacher radial rakhlin rates reading recognition reconstruction references regularization regularized relationships representer reproducing research results riemannian risk rosasco sample sampling scale scholkopf scientific scovel selection semidefinite sensitive shannon shawe siam sigkdd singapore singular sinitsyn smale smola society soft solla sons spaces spline springer springerverlag squares staff statistical statistics stein steinwart structural studies support systems talagrand task taylor theorem theorems theoretical theory todd transactions tsybakov uniform universal university vaart values vapnik vector verlag viewpoint vito wahba weak wellner wiley williamson with world ying york yosida zhang zhou zinn http://www.jmlr.org/papers/volume7/fumera06a/fumera06a.pdf 201 Spam Filtering Based On The Analysis Of Text Information Embedded Into Images aaai advances analysis androutsopoulos anti approach asian athens attack automated available baird based bayes bayesian building burges captcha categorization cbandrinos chellapilla cheng classification classify cloete communications comparison computer computing conference cormack corpus curb cybernetics czerwinski data defense developments document drucker dumais editors eighth enron european evaluation event experimental features filter filtering forthcoming fourth fumera geer graham greece gvcormac hard heckerman help hips holmes horvitz html http human humanfriendly ieee image information intelligence interactive international joachims junk kernel keyword klimt koutsias language large larson learning legible line lynam machine machines madison mail making mccallum messages methods models moll naive networks neural nigam noisy olkopf pages pattern paulgraham personal plan practical press proceedings processing proofs reading recognition references report research resistant retrieval riopka sahami scale scattertype sebastiani second segment segmentation simard smola spam spamcormack spie spyropoulos standards statistical supervised support survey surveys systems technical techniques text transaction transactions using uwaterloo vapnik vector vinciarelli visual wang wars weinstein will wisconsin with workshop yang zhang http://www.jmlr.org/papers/volume3/guyon03a/guyon03a.pdf 273 Covering Numb er Bounds of Certain Regularized Linear Function Classes abound additive advances agnostic akad alexander algorithm algorithmic alon amer american analysis annals anthony application applications approximation association asymptotic attributes average averaging banach barron bartlett based bayesian berlin bernhard bianchi boosting bounded bounds burges cambridge cantelli capacity carnegie cesa chapter characteristics chervonenkis christopher city class classes classification classifiers colt combinatorial completely complexity comput computation computer convergence course covering cristianini data david decision density dependent devroye dimension dimensionbased dimensions discriminant discrimination dokl dual dudley editors effectiveness efficient empirical entropy events explanation exponentiated families formulation foundations frequencies freund from function functional functionals general generalization generalizing gentile geometric glivenko gradient greedy grove gurvits gyorfi haussler hierarchies hilbert hinge hoeffding ieee important inequalities information interior introduction iowa irrelevant isbn isoperimetry issn jaakkola jebara jection john jones journal kernel kernels kivinen kolmogorov langford learnability learning lecture ledoux leen lemma line linear littlestone loss lugosi machine machines march margin marina martin math mathematical mathematics maximum mcallester meila mellon mendelson methods metric michael michel minimization model more muller nauk nello network networks neural nips note notes number numbers operators other over pages pattern performance peter point pollard pontriagin potential practice prediction press probabilistic probabilities probability proc proceedings processes processing programming propri pursuit quickly random randomized rates recognition reduction references regression regularization regularized relative report results risk robert sample sauer scale schapire schnirelmann scholkopf schuurmans science seeger sensitive series sets shahar shawe sigmoidal simple size smola solla some sons space spaces springer sssr statist statistical stochastic structural sums superpositions support symposium syst systems talagrand taylor technical than their theoretic theoretical theory threshold tihomirov todd tommi tong tony training trans transactions transl trique uniform universal university updates vapnik variables vector verlag versus volume voting warmuth weights when wiley williamson with ying yoav york zhang http://www.jmlr.org/papers/volume7/meinshausen06a/meinshausen06a.pdf 138 Quantile Regression Forests absolute adaptive additive american amit andy annals anomaly applied artificial association asymptotics austin austria barnett bartlett belmont berkeley bernoul biometrika bivariate boosting breiman california cambridge carlo chaudhuri classification computability computation computing conditional conjectures consistency core correlation data department detection development effectiveness environment error estimates estimating estimation explanation forests foundation framework freund friedman gaussian geman hare hastie hodge http huber hush hyndman intel isbn jeon john journal koenker language laplacian learning lewis liaw ligence linear logistic machine margin markou matthew methodologies methods model monte multiple nearest neighbors neural news nicta nonparametric novelty olshen optimal outlier outliers packages portnoy press processing project quantile quantiles quantization random randomforest randomized recognition references regression report research review robust royal sample schapire science scovel sears shape signal simple singh smola smoothing society sons splines squared statistical statistician statistics steinwart stone survey team technical tibshirani tortoise transformations trees university using versus vienna view voting wadsworth weisberg wiener wiley wisconsin with http://www.jmlr.org/papers/volume8/nilsson07a/nilsson07a.pdf 224 Consistent Feature Selection for Pattern Recognition in Polynomial Time aaai abundance acyclic alexander algorithm algorithms aliferis almuallim analysis andre anil annual application approaches approximate arthur artificial assumption avril barnhill based bayesian before bell berger bernhard biological bjorkegren blake bloomfield blum bound boundaries bousquet buhlmann california caligiuri campenhout cancer casella cellular characterization chemistry chickering christopher class classifiation classification coller comes communications comparing computation conditional conference consistency consistent constantin corinna correct cortes cover cybernetics daphne darnell data databases david devroye dietterich dimensional directed discovery donna douglas downing duxbury edition efficient eighth elisseeff eric estimating european examples expression factors faithfulness feature features filters finding formalism francisco fransisco friedman from gaasenbeek gabor gaussian gene genetics george golub graphical graphs gretton guyon gyorfi hard hardin heckerman herbrich hettich high holland html http huan huard hussein hyperparameters identifying ieee independence inference inferring influence ingo intelligence intelligent international introduction ioannis irrelevant isabelle iterative jain james jason jesper johan john jose journal judea kalisch karl kauffman kaufmann keerthi kenji kernel kira kohavi koller lander langley large larry lazlo learnable learning leslie lugosi machine machines magnitude many margin markov markus maxwell measurement measuring meek mehran merz mesirov methods michael milan mlearn mlrepository models molecular monitoring morgan multivariate national nature networks neural newman nilsson ninth nodes number olivier olkopf optimal orderings orders over pages pathways pattern patterns pearl pena peter pfleger possible prediction press principled probabilistic problem proceedings publishers quantitative radius ralf reasoning recognition redundancy references relevance relevancy relevant rendell report repository research reviews roger roland sahami sample sathiya scalable science selection seminar slonim smola springer statistical statistics statistik steinwart stephen structure structures studeny subset supervised supplement support switzerland symbolic systems tamayo targets technical tegner tests theoretical theory therapy thomas todd towards traditional transactions transcript transcription tsamardinos tuning uncertainty under using valiant vapnik variable varies vector verlag vladimir volume waller wang weston william with workshop wrappers yeast york zurich http://www.jmlr.org/papers/volume8/fukumizu07a/fukumizu07a.pdf 217 Statistical Consistency of Kernel Canonical Correlation Analysis academic advances aihara aistats akaho akihiro alamos alexander alfred algorithmic alternating american analysis anderson andreas annals applications approach aronszajn arthur artificial association asymptotic augath bach baker banach barry based belitski bernard bernhard bioinformatics bischof blanchard bousquet breiman buja cambridge canonical chaotic charles chobanyan clint clusters company component computation conference constrained convergence correlated correlation correspondence covariance cristianini cross curves data david dependence description detecting dimensionality discriminant distributions douglas dunford dynamical edition editors eigenspaces eigenvalue equations estimating explicit extraction feature first fisher fmri francis fredholm friedman friman from fukumizu functional gaussian gene general generalized genomic georg gilles greenacre gretton groetsch gunnar guozhong hans hardoon herbrich hidden hilbert hiromichi holland horst horth hush icann ieee image imps independent information ingo integrable international interscience introduction jacob jain jane jason jean jerome john joint jordan journal kanehisa kazuyuki kcca kenji kernel kernels kind klaus laboratory larsen laurent learning least leurgans linear ling logothetis london machine machines mathematical measurement measures measuring medical meeting melzer method methods michael micheal mika minoru moyeed muller multiple multivariate nachman nakaya national nello nelson networks neural nikolai nonlinear norms olivier olkopf operators optimal other pages part pauls peter philippe physics pitman platt ponent press principal probability problem proceedings procersses processing psychometric publishing rana ratsch rayama reduction reed references regression regularization reidel reiter remarks renyi report reproducing research robert royal schmidt scholkopf schwartz scovel sebastian sergei series shawe shotaro signal silverman simon smola society sons spaces springer square squares statistical statistics steinwart stochastic suetani supervised support synchronization systems tarieladze taylor technical theodore theory third thomas tikhonov tomita transactions transformations understanding university using vaart vakhania variates vazha vector vert volume wang weiss weston when wiley wilson with yamanishi yoshihiro yukito zwald http://www.jmlr.org/papers/volume1/heckerman00a/heckerman00a.pdf 336 Sufficient Dimensionality Reduction academic accuracy advances agglomerative algorithms allerton amari american analysis annals annual application approach artificial arxiv association asymmetry becker berger bialek bottleneck cambridge categories categorization channel chapman classification classifications classified clogg clustering clusters communication comparison computation computational computing conf conference contingency control coordination correlation cover cross csiszar darroch data deerwester della devroye dimensionality discriminant distributions divergence document dumais efficient elements embedding entries entropy estimation eugenics european examination factorization features fields filter filtering fisher framework friedman from functions furnas geiger generalized geometric geometry george global globerson goodman gyorfi haberman hall harshman having hinton hofmann http icml ieee ijcai indexing inducing information informative intelligence intrinsic iterative jaynes joachims journal kearns lafferty landauer lang langford language lanl lapidoth latent learning linear linguistics local locally lugosi machine machines maloney malouf many math matrix maximally maximum mccallum mechanics method methods minimax minimization missing modeling models mosenzon multidimensional multivariate mumford nagaoka narayan natural nature negative neural news newsweeder nigam noise nonlinear number objects ordered oxford pages pami parameter parts pattern pereira phil physical physics pietra pitman precision press principle probabilistic probability problems proc proceedings processing queries qulitative random ratcliff recognition reduction references relevant reliable review roweis rudra saul scaling science semantic sets seung seventeenth sigir silva simple sixth slonim society springer statist statistical statistics sufficient support symposium systems tables tenenbaum text texture theory thoery thomas tishby tolerant transactions uncertainty under university unordered unsupervised using vector volume wiley with without wolf word workshop yang http://www.jmlr.org/papers/volume3/bshouty02a/bshouty02a.pdf 361 Learning Monotone DNF from a Teacher that Almost Do es Not Answer Membership Queries algorithm algorithmic algorithms almost amsterdam analogical angluin annals annu annual answer april artificial auer berlin bianchi boolean bounds bshouty california castle cesa chen closure colt comput computational concept conference cruz dana david does donna eiron environments equivalence eurocolt european fallible formulae formulas from germany goldman incomplete inductive inference inform intel international issue jority july kearns learnability learning lecture leonard leslie ligence line linear littlestone logarithmic machine malicious manfred mathematics mathias membership michael ming mistake monotone nadav nader netherlands nick nicol noise noisy note notes october oracle owshanko pages peter pitt press proc proceedings queries randomly rectangles references regular reinhardsbrunn report sally santa sets sixth slonim special springer sympos teacher teachers technical term that theory thesis threshold ucsc univ using valiant verlag volume warmuth weighted with workshop york zhixiang http://www.jmlr.org/papers/volume6/banerjee05b/banerjee05b.pdf 88 Clustering with Bregman Divergences academic acoustics acta akhizer algorithm algorithmic algorithms allerton alternating amari american analysis annals annual answers application applications approach arbitrary arimoto associated automatic auxiliary axiomatic azoury banerjee barndorff based basis bayesian berg berger berkeley bernoulli between bialek bilmes biology blahut bottleneck bounds bregman buzo california capacity carnegie censor channel channels christensen classical classification cluster clustering coding coefficients collins colt combining common communication communications company complexity component compression computation computational computer computing conditional conference continuous control convex covariances cover csiszar dasgupta data dawid decision decisions definite densities density department dept design devinatz dhillon discrepancy discrete dissimilarity distance distances distortion distributions divergences diversity divisive documents duality dubes duda duke dynamical elements ensembles entropy equality estimation expectation expected exponential exponentially extensions families family feature finding forster framework from functions game gaussian generalization generalized gentle genton geometry gersho ghosh gibson gneiting graphical gray grunwald hafner hall harmonic hart hidden hungarica icml icsi ieee inference information instantaneous integrals international interpretation interscience inverse isit issue jain jensen jersey john jordan journal july kazakos kluwer knowledge kolmogorov krishnan kumar labeled lafferty landmarks language laplace latent learing learning least letters likelihood linde line linear logic loss lossy machine macqueen mallela mapping markel markov mathematica mathematical mathematics maximum mccallum mclachlan means measures mellon memoryless merugu method methods minimization minimum mitchell mixture models modha moment multivariate nagaoka negative networks neuhoff neural nielsen nigam nips observations optimal optimality optimization oxford pages palus parallel parameter partitionings pattern pennsylvania pereira physics pietra points population positive prediction predictor prentice press princeton principal probability problem problems proc procedures processes processing programming projections publishers publishing quantization quantizer questions rate rates redner references related relative relaxation report representation research ressel reuse review robust rockafellar rose schapire school schuurmans science semigroups sets siam signal slonim society solution some sons source spangler spectral speech springer squares stationary statistical statistics stieltjes stork strehl supplement symposium systems technical terms text theoretic theoretical theory thomas thrun tishby transactions tusnady tutorial unified university unlabeled upon using ussr variable variational vector verlag vitanyi wainwright walker wang warmuth weighting weiss wiley with zenios http://www.jmlr.org/papers/volume1/herbrich01a/herbrich01a.pdf 343 Use of the Zero-Norm with Linear Models and Kernel Methods aizerman alamos algorithm alizadeh alon amaldi analysis annual appear application approximability arrays artes artificial assessing automation barkai barnhill based bayesian bengio bennett biology blum boser bradley brain braverman broad brussels cambridge cancer cell chapelle citeseer class classification classifiers clustering colon colt computational computer computing concave conference control cortes courant cruz data delve department dependences detection determination diffues discrimination distinct editor elisseeff empirical english error esann estimation examples expression face facto feature features fifth figueiras foundations francisco franke freund function fung gene generalization girosi gish guez guyon hilbert html http icml identified ieee improved inference information informs inseparable institute intelligence interscience john journal kann kernel kernels langley large learning levine linear linearly logistics lymphoma machine machines mack madison mangasarian margin massive mathematical method methods minimal minimiza minimization minimizing mining model models moscow mukherjee multi multiplicative nadeau nature nauka naval navia neal networks neural nips nonlinear norm normal notterman oligonucleotide operations optimal optimization organization osuna pages parallel pattern patterns people perceptron perez perspectives physics pittsburgh poggio pontil potential press probabilistic probed proc proceedings processing profiling programming psychological publication publishers quadratic quarterly recognition references relations relevance relevant remote report research revealed review risk robust rodr rosenblatt rozonoer russian schapire scholkopf science sciences selection separation series sets signal smola society software sons sparse springer statistical storage street submitted support svms systems takimoto technical theoretical theory tion tipping tissues training transaction transactions translation tuebingen tumor types university unsatisfied updates using vapnik variables vazquez vector verlag verleysen vidal vision volume warmuth watkins weston wiley wisconsin with wolfe workshop ybarra york zero http://www.jmlr.org/papers/volume3/kalai02a/kalai02a.pdf 310 Learning Probabilistic Models of Link Structure abbeel abduction academic acsc algorithms american annual applications approaches approximate artificial association authoritative automating based bayes bayesian belief british california cambridge canada categorization chakrabarti classification clustering cohn columbia combining comparison computer concurrency conference connectivity construction content craven data david dipasquo discovering discrete discriminative document domains dzeroski editor editors edmonton eighteenth ellis empirical enhanced environment event extract fifteenth frame francisco freitag friedman from getoor ghani graphical haddawy heckerman heidelberg hofmann horn horwood hyperlink hyperlinked hyperlinks hypertext inductive indyk inference information intelligence intelligent international internet joint jordan journal kaufmann kleinberg kluwer knowledge koller lavrac lawrence learning lecture link logic logistic loopy machine madison management massachusetts mccallum meta methods mining missing mitchell model models morgan muggleton multi murphy naive networks neural nigam ninth notes page pages patterns pearl pennock pfeffer poole popescul portals press probabilistic proc proceedings processing programming propagation reasoning references regression regularities relational rennie retrieval science seattle segal seventeenth seymore siam sigmod sixteenth slattery sources springer stanford statistical stockholm structural study sweden symbolic symposium systems taskar techniques test text thesis towards tutorial uncertainty ungar university using vancouver verlag volume washington weiss wide williamstown wisconsin with workshop world yang york http://www.jmlr.org/papers/volume6/jaeger05a/jaeger05a.pdf 41 Generalization Bounds and Complexities Based on Sparsity and Clustering for Convex Combinations of Functions from Random Classes ability advances algorithms analysis andonova annals annual anthony applications applied arcing artificial balancing bartlett based berkeley birkhauser boosting borowkow boston bound bounding bounds breiman burges california cambridge cannon central certain cervonenkis class classes classification classifiers clustering combinations combined communications complexities compression computation computational computer concerning conference confidence convergence convex cruz data department dependencies dependent detection dimensionality discrete distributions dokl dudley effectiveness electronic empirical error estimation ettinger event experimental explanation extensions fast feature feller frequencies freund games generalization heisele hierarchy hush hypothesis ieee images improved inequality intelligence introduction john journal kernel kohonen koltchinskii learnability learning lecture limit littlestone lozano machine machines margin margins math mathematics methods moscow mukherjee nauka neural notes object organizing other panchenko pattern performance poggio prediction predictions press probabilities probability proceedings rated recognition reduction references relating relative report research result russian santa schapire schoelkopf scholkopf scovel self serre shawe singer smola society some sons soviet sparseness sparsity springer springerverlag statistical statistics steinwart submitted support taylor technical their theorems theoretical theorie theory thesis uniform university using vapnik vector verlag video vision volume voting wahrscheinlichkeits warmuth wiley with york http://www.jmlr.org/papers/volume7/cesa-bianchi06b/cesa-bianchi06b.pdf 147 Worst-Case Analysis of Selective Sampling for Linear Classification abound about active adaptive advances algorithm algorithms american analysis angluin annals annual apple applications approximate artificial association atlas attributes auer automata average azoury based best bianchi block bordes bottou bounded bounds brain budget california cambridge campbell case cesa classification classifiers cohn committee computation computational computer computing concept conconi conference confident connectionist continuous control convergence corpus cristianini cruz dasgupta dekel density descent discrete discriminant distributions duda dynamic edition efficient electronic ertekin estimates estimation exponential familiy fast fixed forgetron forster freund functioning fundamentals game games general generalized gentile gradient grove hart helmbold herbster hinge hoeffding identification inequalities information intelligence international interscience introduction irrelevant jagota journal kalai kaufman kaufmann kernel kernels kivinen koller label ladner large learning least line linear littlestone lnai lncs logarithmic long loss lugosi machine machines margin mathematical mathematics maximal maximum mistake model models modern monteleoni morgan multidimensional networks neural nonlinear norm novikov online order organization pages panizza pattern perceptron perceptrons physics potential prediction predictor press probabilistic probability problems proceedings processing proofs psychological queries query quickly random references regression relative relaxed research researchandstandards reserarch results reuters review reviews robustness rosenblatt sampling santa schapire scholkopf schuurmans sciences second selective self seung shalev shamir shawe shwartz siam singer smola some springer squares statistical statistics stochastic storage stork sums support symposium system systems tasting taylor text theory thesis threshold time tishby tong tracking training university updates using vapnik variables vector warmuth weston when wiley with worst http://www.jmlr.org/papers/volume3/herbrich02a/herbrich02a.pdf 367 Algorithmic Luckiness advances algorithm algorithmic algorithms alignment alon american analysis annual anthony application applications applied approximants approximate approximation artificial association bartlett based bayes bayesian becker berlin blumer boser boucheron bound bounded bounds bousquet brain brussels california cambridge campbell chervonenkis class classifier classifiers clustering code combinatorial compression computational computer computing concentration conditions conference consistency convergence cortes cristianini cruz data delsarte density dependences dependent devroye dietterich dimension discrete discriminant douglas editor editors ehrenfeucht elisseeff empirical erdos error esann estimation events facto families feder feller fisher floyd foundations frequencies functions generalisation generalization generalized ghahramani graepel greene guyon haussler herbrich hierarchies hoeffding ieee image inequalities inequality information introduction isbn john journal july kahane kandola kearns kernel kernels large larsen learnability learning leen linear littlestone lkopf luckiness lugosi machine machines macwilliams makovoz marchand margin massachusetts massart mathematics mceliece method methods mika minimization model more motion muller multi nature necessary networks neural optimal organization over pages pattern peled perceptron performance pittsburgh point press probabilistic probabilities probability proceedings processing psychological random rate ratsch recognition references relating relative relevance report representer research result review risk rodemich rosenblatt rumsey sample santa sauer scholkopf schuurmans science series sets sharp shawe signal smola some sons sparse spencer springer stability statistical storage structural structures sufficient sums support svms symposium systems target taylor technical than their theorem theory tipping training transactions tresp uniform university upper vapnik vapnikchervonenkis variables vazirani vector verleysen volume warmuth watkins welch weston wiley williamson wilson with work york http://www.jmlr.org/papers/volume5/fleuret04a/fleuret04a.pdf 29 Fast Binary Feature Selection with Conditional Mutual Information aaai adaboost additive advances algorithm algorithms amit analysis annal annu annual applied artificial automata bagging bartlett based battiti baxter bayesian bednarski berkeley binary bioinformatics bonnlander boosting boser brain breiman california cambridge cancer challenge christiani classification classifier classifiers coarse collaborative comput computational computer conditional conf conference convergence correlation coupling cover cristianini data decision density department descent detection dimensional dror duda duffy editors elements elisseeff ensembles estimation experiments expression extraction face fast feature features fifth filter filtering filters fine fleuret forests frean freund friedman furey game geman gradient gunn gunopulos guyon hart hastie haussler high hybrid icpr ieee induction infinity information informative input inria intelligence international interscience introduction isann john joint jordan journal kaufmann kearns kernelbased kohavi koller langley learning line linear logistic machine machines margin mason mathematical maximization methods michael microarray miyahara model morgan muller mutual naive naquet nature networks neural nips nonparametric november novikoff object october onoda optimal organization other pacific pages pairwise parametric pattern pazzani perceptron perceptrons pose precise prediction predictors press probabilistic proc proceedings processing proofs psychological random ratanamahatana ratsch recognition references regression regularizing report research result review rosenblatt sahami samples sara scene schapire schummer selecting selection shape shawe simple solla solution some sons springer statistical statistics storage subset supervised support symposium systems taylor technical theory thomas thompson tibshirani tissue torkkola toward training transactions tree trees ullman university using validation vapnik variable variables vector verlag vidal view vision visual volume weigend wilder wiley with workshop wrappers york http://www.jmlr.org/papers/volume7/malioutov06a/malioutov06a.pdf 176 Walk-Sums and Belief Propagation in Gaussian Graphical Models advances algebra algorithm algorithms analysis anisotropic annals applications arbitrary artificial available based belief biometrika boman bounds brydges cambridge chandrasekaran chen classical clifford communications computation conditions conf consensus convergence correctness correlation covariance cowell critical cycles dawid december decoding decomposition densities distributions edition effects eigenvalues embedded errors estimation expert exploring extended factor fields finite fisher framework freeman frohlich gauss gaussian general generalizations gibbs godement graph graphical graphs growing hammersley hill horn http ieee ihler inequalities inference information intelligence intelligent interpretation ising iterative jaakkola jasonj johnson jones jordan journal kappen kaufmann kiiveri kirkland kumar laplacian lattices lauritzen learning linear loopy machine malioutov manuscript march markov math mathematical matrices matrix mcdonald mcgraw measures message millennium moallemi models mooij morgan multilinear networks neural november optimization over oxford parekh passing patterns pearl people physical physics plarre positive preparation press principles probabilistic proc processes processing product propagation quadratic random references related reparameterization report representation require research review rudin rusmevichientong science series sign signal sokal speed spiegelhalter spin springer statistical statistics strang subgraph sudderth sufficient summable symmetric systems tatikonda technical tees temperatures that theory toledo topics topology trans tree trees tsatsomeros turbo uncertainty understanding undirected university unpublished upper varga verlag wainwright walk weiss west width willsky with yedidia http://www.jmlr.org/papers/volume7/sahbi06a/sahbi06a.pdf 178 A Hierarchy of Support Vector Machines for Pattern Detection abdel accuracy accurate across active advances ahlers ahuja aktuell algorithm amit analysis annals annual antifaces appearance application applied apprentissage approach approximation architecture asian asymmetric automatic background baker bakir baluja based battiti bellingham bernier blake blanchard boosted borgefors boser boujemaa brown burges burl carnegie cars cascade cccd center chamfer changements chen class classification classifier classifiers clustering cluttered coarse collobert color comparing complex computation computational computationally computer computing conference convolutional cootes dagm decision deficient delakis department design designs detection detector detectors devinney distance ecole edge editors efficient eighteenth elad electrical engineering etection eveland evgeniou example expansions face faces fast feature features feraud fifth finder finding fine fleuret floatboost flow focalisation france franz freund gangaputra garcia geman gesture girosi gotsman graph gravity guyon hanson hausdorff haynor heisele hierarchical hierarchique hierarchy hopkins human huttenlocher ieee image images imaging improving informatik information institute intelligence international interpretation interscience jain johns jones jordan journal jung jussieu kanade keren kernel keshet kienzle knirsch koch kyoto labelled learning lecture letters leung levi linear loew machine machines manbeck marchette margin massachusetts matching maximal mcclure medical mellon method methodology miao michael model models mottaleb mozer mukherjee mullin multi multiangle multimedia multiple multiscale mustererkennung nayar network networks neuheisel neural notes novel object objets optical optimal osadchy osuna pages papageorgiou parametric paris pattern perona petsche picture pittas poggio polytechnique pontil press priebe principle priors problems proceedings processing publication random rank rapid recognition recogntion reconnaissance reduction references rehg rejection report representations resolution robust romdhani roth rowley rucklidge rules sahbi scale scales scenes schanz schneiderman scholkopf science search segmentation selection sequential serre shape shen sicily simple simplified skewed skin smola snow sobottka socolinsky spaces specific speed spie statistical statistics statistique strategy sung support symposium system systems taromina taylor technical technique technology template testing theory thesis thomas torr trace tracking training transactions university using vapnik vector versailles viallet video view viola visages vision visual volume walker wang wiley with workshop yang zhang http://www.jmlr.org/papers/volume8/cawley07a/cawley07a.pdf 234 Preventing Over-Fitting during Model Selection via Bayesian Regularisation of the Hyper-Parameters academic accuracy adaboost adaptive advances alamdari alberta algorithm algorithms allen anal analysis anderson annual applic applied approaches approximate approximation april arsenin artificial assessment august automatic azar back backprop backpropagation baltimore banff barber based bayesian between bias bienenstock biology bischof bishop blackford bootstrap bousquet brabanter brailovsky buhmann buntine cambridge canada cawley challenge chapelle characters choice cholesky choosing classification classifiers classify comparisons complex computation computational computations computer conditionally conference conjugate connection constructive cortes cristianini criteria cross croz csrp curvature data december decomposition definite demmel demsar determination dilema discriminant distance dongarra double doursat driven dror drucker dual duan econometrica edition editor editors efficient entropy equations error estimation evidence expectation extraction fast feature february feedforward fifth fisher fourteenth framework function functions gaussian geman generalization gestel ghahramani golub gradients gradshteyn greenbaum guide guyon hammarling heidbreder hopkins hyperparameters icpr ieee ijcai ijcnn improving inference information integral integrals integrate intelligence interface international interpolation invariant inversion jiao joachims john johns joint journal july june kauffmann kaufmann keerthi kernel kernels kibernetica kimeldorf kingdom kluwer kohavi lachenbruch lapack laplace learning least leave leaveone leen letters loan logistic london lssvms luntz machine machines mackay margins marquardt mateo math matrices matrix maximum mckenney mead mercer method methods micchelli mickey mika minimization minka model models moor morgan mukherjee muller multiple negative nelder networks neural november observational obtained onoda optimally optimise other over overfitting oxford pages parameters pattern performance philadelphia philosophical picard pittsburgh posed positive practical prediction predictions predictive press prior problems procedure proceedings processes processing products propagation pruning publishers rasmussen rates ratsch recognition references regression regularised regularization relationship relevance remidies report research results royal russian ryzhic saadi saffari sarle scaling scattered scholkopf science scientific seaks selection september series sets shawe shevade siam signal simplex size smola smoothing society soft solla solutions some sorenson spaces sparse spline squares statistical statistics step stone stopped study submitted sundararajan support sussex suykens syminv symposium systems table talbot taylor tchebycheffian techicheskaya technical technometrics text their theory third tikhonov tipping training transactions tsuda type uller uncertainty united university users using validation validatory vancouver vandewalle vapnik variable variance vector vert volume wahba wang weigend weighted weston wiley williams with world york http://www.jmlr.org/papers/volume6/bordes05a/bordes05a.pdf 84 Fast Kernel Classifiers with Online and Active Learning academic active adaboost adatron advances aistat aizerman algorihms algorithm algorithms amari american analysis annual appear applications applied apprentissage approximate aronszajn artificial asymptotically atlas automata automatic automation bakir barbados based bastian becker bengio berlin bernhard bianchi bordes boser bottou bounds brain braverman breaking brooklyn budget burges business california cambridge campbell canu capacity cascade case cauwenberghs cesa chang chattopadhyay chervonenkis cheung chosen cjlin classification classifier classifiers classifying cohn collobert colt comparison complexity compression computation computational computer conf conference confrence connectionist control convergence core cortes cosatto cowan cowell crammer cristianini cross cruz csie data datasets decomposition decremental denker dependences dependency design dietterich digit dimension directions domingo dourdanovic drucker dynamics ecml editor editors efficient eisenberg electrical elsevier empirical engineering estimation european even examples experiments fast feasible fedorov fifteenth foundations francisco freund frie fulk function functions generalized gentile ghahramani gilbert giles glaskova graf grenoble guyon handwritten hanson hill hotel http huller iapr icml ieee image incremental industry information institute integer intelligence international introduction isbn isee jackel jerusalem joachims john journal june kandola kaufmann keerthi kernel kernelbased kernels koller koscheev kwok ladner langley large lawrence learnability learning lecture lecun leon lerner less library libsvm line linear linearthreshold liquiere littlestone long loosli machine machines mackay madaboost making marc margin mateo mathematical maximal maximum mcgraw method methods michel mikhailski minimal mixture model models modification more morgan muller multiclass murata murty narasimha national nauka networks neural nilsson nishi notes november novikoff objective october offline online optimal optimization organization other outils pages parallel pattern perceptron perceptrons platt poggio polytechnic portrait potential pour practical presentation press presses probabilistic problems procedure proceedings processing programming programs proofs psychological publishers queries query random rapide recognition references regression relating relaxed remote report reproducing research review rivest robert rosenblatt rozonoer russian sackinger sample sampling santa saul savannah scale schapire schohn scholkopf schrijver science sebastian sebban section selection selective sequential seventeenth sharp shavlik shawe signal simard simple simplesvm singer smola society some sons sparseness springer statistical statistics steinwart stochastic storage study support svms svmtorch symposium systems taiwan takahashi taylor technical tenth termination text theoretical theory third thrun tighter tong touretzky trainable training transactions tsang tuning ultraconservative universitaires university updated using vapnik vector verlag very vishwanathan vision volume warmuth watanabe weiss weston wiley with workshop worst york zaniboni zoubin zoutendijk http://www.jmlr.org/papers/volume6/quinonero-candela05a/quinonero-candela05a.pdf 95 A Unifying View of Sparse Approximate Gaussian Process Regression advances alexander anton approach approximate artificial aspects assessing augmentation barbara bartlett bayesian becker bernhard bhaskar biasvariance binary bishop brendan cambridge candela carl christopher classification cohn committee computation computational conference corinna cortes csato curve david denmark dietterich discussion dong edinburgh editors edward fangyu fast fitting forward frey gacv gatsby gaussiam gaussian ghahramani grace greedy hasselmo healing inductive inference information inputs intelligence international jason joaquin journal keerthi kernel kerns klaus klein kuss lawrence learning leen lehel lyngby machine machines malte manfred massachussetts matching mathias matthias method methods michael mozer nature neil network neural neuroscience ninth nystrom obermayer online opper pages palmer parametric perspectives peter platt prediction press process processes processing pseudo pursuit quinonero randomized rank rasmussen reduced references regression relevance report research ronald sara sathiya saul scholkopf schwaighofer scotland sebas sebastian seeger selection servations silverman smola smoothing snelson society solla some sparse speed spline springer stat statistical statistics support suzanna systems technical theory thesis thomas thrun tian tipping todd touretzky tradeoff transductive tresp uncertainty unit university using vapnik vector verlag vladimir volker wahba weiss williams wipf with workshop xiang xiwu zoubin http://www.jmlr.org/papers/volume6/hutter05a/hutter05a.pdf 52 Adaptive Online Prediction by Following the Perturbed Leader active adaptive adversarial adversary advice against aggregating alamitos algorithm algorithmic algorithms alphabet annual application approximation armed artificial arxiv auer bandit based bayes bayesian bayespea belgium benelearn berlin better bianchi blum boosting bounded bounds carolina casino cesa colt combinatorics computation computational computer conf conference confident contributions convergence decision decisions differences dresher editors efficient enschede expert experts focs following foundations francisco freund gambling game games general generalization gentile geometric hannan haussler helmbold http hutter idsia ieee increasing inductive information intelligence international journal july kalai kaufmann leader learning lecture line littlestone lnai lncs london loss ltbip machine majority marcus master mathematical mcdiarmid mcmahan method morgan multi netherlands norm north notes online optimality optimization padova pages park pascal perturbed plays poland prediction presented press princeton principles probability problem problems proc references repeated report research rigged risk robustness rochester schapire science sciences seiden self sequence sequential series setting society springer strategies surveys symposium system technical theoretic theory third trans triangle tucker uaibook universal university values vempala versus volume vovk warmuth weighted weights with wolfe workshop yaniv yaroshinsky york http://www.jmlr.org/papers/volume3/chan02a/chan02a.pdf 364 Variational Learning of Clusters of Undercomplete Nonsymmetric Independent Components aapo accepted adali advances algorithm amari analysers analysis anthony appearance approaches artificial attias aussois automatic bases bayes bayesian beal bell biomedical bishop blind boston brain cambridge cardoso cell chan choudrey christopher class classification classifier clusters code comparison component components computation conf conference constrained context correlation deconvolution department diagnosis diego disc distributions douglas dynamics editor editors edwin elisabeth emergence engineering ensemble eric eriksson erkki estimation examination factor field findings frans gassiat gaussian germany ghahramani glaucoma goldbaum gradient graphical guan hagai hansen hitchings hyvarinen icann icassp ieee images imaging independent inference inferring information intelligence international introduction jaakkola james jaynes jean johnson jordan juha jung karhunen karvanen kluwer koivunen kwokleung lappalainen lars larsen latent lawrence learning leen lewicki likelihood machine makeig markus martin matthew maximum mckeown mean measurements methods michael miskin mixture mixtures model models moulines muller munich natural nature neal networks neural ninth noisy nonsymmetric ocular olshausen ophthalmol optic over oxford pages paliwa pamela papers parameters parg pattern pedro perimetry philadelphia physics ping piscataway press probability proc proceedings processing properties radford receptive recognition references rensen report rizwan robert roberts sample saul saunders science scott sejnowski separation signal signals simple society solla source sources spaeth sparse springer statistical statistics stephen structure switching systems technical terrence testing thesis todd tommi traditional transactions tzyy uncertainty undercomplete university unsupervised using variable variational verlag visual volume weber weinreb welling wide widrow wiley wilson winther with workshop york zadnik zoubin http://www.jmlr.org/papers/volume4/costa03a/costa03a.pdf 265 Recommender Systems Using Linear Classifiers aaai advances algorithm algorithms analysis andrew applications approach approaches apte architecture artificial asia august automating baltimore based basic bayes bergstrom biased blake bottom breese california cambridge carl categorization chickering christopher chute classification classifiers clustering collaborative combination commerce compaq comparison compression computations computer conald conference context continuous cortes cscw damerau data databases david decision dependency dept discovery dumais eachmovie ecml edition editors emperical empirical estimation european evaluation event example fast features fifteenth fifth filtering formal foster fourteenth frank generalization glassman goetz golub grouplens hampp heckerman hoerl hofmann hopkins html http iacovou icml ieee inductive inference inform information intel international irvine iyengar joachims john johns johnson jordan journal july kadie kamal karypis kaufmann kearns kennard keogh know konstan learing learning ledge ligence ligent linear loan machine machines maes management mansour many mapping markov matrix maximizing maxwell mccallum meek merz method methods mining mlearn mlrespository model models morgan mouth mozer naive national near netnews networks neural nigam nonlinear nonorthogonal oles open optimal pacific pages pattern performance petsche platt predictive press problems proceedings processing programs properties pruning publishers quinlan recognition recommendation recommender references regression regularized relevant repository representations research resnick retrieval ridge riedl ripley robert rounthwaite sahami sarwar science shardanand shtarkov social sons statistical study suchak support systems technometrics text their theory third tjalkens tong trans transactions tree tresp uncertainty ungar university using vapnik variables vector visualization weighting weiss wiley willems with word workshop yang yiming york zhang http://www.jmlr.org/papers/volume6/hyvarinen05a/hyvarinen05a.pdf 54 Estimation of Non-Normalized Statistical Mo dels by Score Matching advances aistats algorithm algorithms amari analysis approach artificial barbados based bases basis besag biometrika blind bouman cambridge carreira cell cichocki coding complex component computation computer contrastive dept divergence edge edition efficiency employed energy estimating estimation experts field fields from garrat gaussian generalized hinton hoyer hyvarinen ieee image images imaging independent inference information ings inki intel inter interaction journal karhunen lattice layer learning learns ligence likelihood machine mackay markov matching mathematical maximum minimizing mixture model modeling models natural neural olshausen osindero overcomplete pages perpinan pham preparation preserving press proceed processing products pseudolikelihood quasi random receptive references report representations research royal sauer schervish science score separation series signal simple society source sources sparse spatial springer statistical statistics strategy systems technical theory through topography toronto training transactions university vision welling wiley windows with workshop yang http://www.jmlr.org/papers/volume2/steinwart01a/steinwart01a.pdf 349 On the Influence of the Kernel on the Consistency of Supp ort Vector Machines ability advances analysis annual anthony applications approximation areas artificial associative available bartlett based berlin biological bottou bound burges cambridge carl central chapman classifiers comp compactness compint complexity computational computer conference consistency convergence crisp cristianini cross cybernetics data department dependent deutsch devroye dudley editors egham electronically empirical entropy fakultat frankfurt friedrich generalization generalized germany gradstein gyorfi hall harri herbrich hierarchies holloway http ieee influence informatik information integrals integraltafeln intel introduction jena jordan kernel learn learning leave lecture leen ligence limit lkopf london lugosi machine machines main manual math mathematik measures methods minet minimization muller neural nonlinear notes operators optimal orem other over pages pattern pedersen performance poggio port press probab probabilistic probability proceedings processes processing products produkt real recall recognition reference references report reports representer research rhbnc risk royal ryshik saunders schiller scholkopf science series shawe smola solution sources springer statistical steinwart stephani stitson structural submitted summen supo support tables taylor technical theo theorems theory tong trans uniqueness universit university vaart validation vapnik vector verlag weak wellner weston wiley williamson with york zhang http://www.jmlr.org/papers/volume7/MLOPT-intro06a/MLOPT-intro06a.pdf 149 The Interplay of Optimization and Machine Learning Research academic active advances algorithm algorithms analysis application applications approach approaches athena bach backpropagation based bayesian bazaraa belief benchmarking bennett bergkvist bertsekas bhattacharyya bilinear bishop boyd bradley bregman building burer burges california cambridge catholic chapelle chatalbashev chen classification classifier clustering combinatorial complexity computational concave cone conference conic constrained constraints convergence convex core cortes cristianini cross damaschke data decoste defense definite dempster differentiable distributed dolan dual duality editor editors efficient empirical ensemble error exact extragradient extrapolation factorization fast feature fifteenth first francisco from fung gain gehl generalized glasmachers goberna golub graph graphical guestrin hall handling heiler hettich hierarchical hinton hypotheses icml ieee igel implementation incomplete incremental independent inequalities inference infinite information integer internal international jersey joachims john jordan journal julien kaufmann keerthi kernel klautau kluwer koller kortanek kruger label lacoste laird lanckriet large laskov learning likelihood linear lopez louven luthi machine machines making mangasarian margin markov mathematical matrix matt maximum mcclelland meltzer method methods minimal minimization missing mitchell mittelmann models more morgan mozer muller multilabel multiple multiprocessor musicant negative nemhauser nesterov network networks neural newton niculescu nocedal nonlinear nonmonotone norm numerical onto operations optimization order other overrelaxation oxford pages parallel parameter pattern performance perturbed petsche platt polyhedra practical prediction prentice press probabilistic problems proceedings processing profiles programming programs projections propagation pruning radin ranking ratsch recognition reduced reemtsen references related relaxations representations research review rifkin rousu royal rubin ruckmann rummelhart saunders scale schafer scheinberg schlkopf schnorr scholkopf scientific second selection semi separation sequential serafini serial series sets shalev shavlik shawe sherali shetty shivaswamy shwartz siam singer smola society socp soft software solodov solvers solving sonnenburg space sparse springer statistical statistics street structured study success support svms systems szedmak taskar taylor theory training transaction transduction twenty uncertain unconstrained university using validation vandenberghe vapnik variational vector weiss wiley williams with wolsey working wright yanover york zanghirati zanni zhang http://www.jmlr.org/papers/volume4/sarela03a/sarela03a.pdf 252 Kernel Partial Least Squares Regression in Repro ducing Kernel Hilb ert Space academic additive advances algorithm algorithms alternative american analysis annimab application applications applied approach approximation architectures arnold aronsza assessment association available avenue bartlett based basis between beyond biological birkhauser birl boston brain brooks burges calibration cambridge chemdata chemical chemolab chemometric chemometrics choice cichocki classification coherent college collinearity communications compact component components comprehensive computation computations computer computers computing conference connection constructed continuum control cristianini cross curse data department detection dimensionality discussion displaymonitoring dual dunn dynamics edition editor eigenvalue electronically elements embracing engineering entropy equations equivalence erps estimation evanston event extraction factor fast feature feed fewer formance forward foundation frank friedman from functions gacv gammerman gani garthwaite gaussian geladi generalization generalized girolami girosi golub graphical hall haykin helland hilbert holloway honour hopkins hoskuldsson human illinois indices inference information integral intel intensive international interpretation introduction inverse iterative jects john jolliffe jones jong jordan journal kernel kernels kimeldorf kluwer konig koska kramer krishnaiah laboratory language latent learning least lewi ligent lindgren linear loan london machine machines madison malthouse manne many maple martens massarat mathematical mathematics matrix memo mercer methods model modeling models monitoring montgomery muller multivariate naes nature negative network networks neural neurocolt nnppss noising nonlinear nonlinearized northwestern note number numbers observational operators ordinary pages paisley papers part partial parts pattern peck performance perspectives philadelphia philosophical poggio point positive potentials prediction predictions prentice press principal priors probability problem proceedings processes processing psychology quarterly radial randomized rannar readable recognition references reflection regression regulariza regularization related report repository reproducing results ridge rkhs rosipal royal ruhe saunders scholkopf scientific scotland sequentially series sets shawe shensa siam signal simpls simulation sirovich smola smoothers society soft some sons spaces sparse spline splines springer squares stabilizers statistical statistics stone structure structures support sweden systems tamhane taylor tchebycheffian technical technometrics tegorg tensor their theory thesis tion tools transactions trejo turbulence type ungar university upenn using validated validation validatory vapnik variables vector verlag view volume vovk wahba wavelets wide wiley williams williamson wisconsin with wold york http://www.jmlr.org/papers/volume3/torkkola03a/torkkola03a.pdf 286 Variational Learning of Clusters of Undercomplete Nonsymmetric Independent Components aapo accepted adali advances algorithm amari analysers analysis anthony appearance approaches artificial attias aussois automatic bases bayes bayesian beal bell biomedical bishop blind boston brain cambridge cardoso cell chan choudrey christopher class classification classifier clusters code comparison component components computation conf conference constrained context correlation deconvolution department diagnosis diego disc distributions douglas dynamics editor editors edwin elisabeth emergence engineering ensemble eric eriksson erkki estimation examination factor field findings frans gassiat gaussian germany ghahramani glaucoma goldbaum gradient graphical guan hagai hansen hitchings hyvarinen icann icassp ieee images imaging independent inference inferring information intelligence international introduction jaakkola james jaynes jean johnson jordan juha jung karhunen karvanen kluwer koivunen kwokleung lappalainen lars larsen latent lawrence learning leen lewicki likelihood machine makeig markus martin matthew maximum mckeown mean measurements methods michael miskin mixture mixtures model models moulines muller munich natural nature neal networks neural ninth noisy nonsymmetric ocular olshausen ophthalmol optic over oxford pages paliwa pamela papers parameters parg pattern pedro perimetry philadelphia physics ping piscataway press probability proc proceedings processing properties radford receptive recognition references rensen report rizwan robert roberts sample saul saunders science scott sejnowski separation signal signals simple society solla source sources spaeth sparse springer statistical statistics stephen structure switching systems technical terrence testing thesis todd tommi traditional transactions tzyy uncertainty undercomplete university unsupervised using variable variational verlag visual volume weber weinreb welling wide widrow wiley wilson winther with workshop york zadnik zoubin http://www.jmlr.org/papers/volume3/nair02a/nair02a.pdf 301 Rademacher and Gaussian Complexities: Risk Bounds and Structural Results andrew annals annual anthony application applications applied averages banach bartlett baxter bias bodies boost boosting boucheron bounded bounding bounds cambridge cantelli chervonenkis classes classification classifiers codes coding combinatorics combined comparison complexities complexity computational conference convergence convex correcting corrects cristianini dana data decision department dependent devroye dietterich differences dimensional distance distributions does editors effectiveness empirical error estimation events experimental explanation explicit finite foundations fourteenth frequencies freund function gabor gaussian generalization geometry gine glivenko golea gyorfi hierarchies high ideals ieee important improved information international introduction isoperimetry jaegermann john jonathan july kaufmann kearns koltchinskii kong learning ledoux llew lugosi machine mansour margin margins martin mason mathematics matter mazur mcdiarmid mendelson method methods mexico michael minimisation minimization model modelling monographs more morgan mostefa multiclass nello network networks neural nips norm number october operator optimization output over pages panchenko pattern penalties peter phane phase pisier pitman positivity press probabilistic probabilities probability problems proc proceedings processes pure rademacher recogniao references relative report results risk robert sample schapire selection shahar shawe size space spaces springer statistics stochastic structural support surveys talagrand taylor technical than their theoretical theory through tion tomczak transactions transitions trees uniform university using vapnik variance vector vladimir volume voting weights wellner williamson with yishay yoav http://www.jmlr.org/papers/volume6/wingate05a/wingate05a.pdf 59 Prioritization Methods for Accelerating MDP Solvers aaai abhandlungen acquisition adaptive advances albuquerque algorithm algorithms alpert andre andrew angeles annual approximate artificial asynchronous athena atkeson automatic autonomous aztec baird barrett barto based behavior belmont berry bertsekas birmingham blocks boston bounds bradtke building cache california cambridge carlos carnegie chan charles christopher cliffs coarse coding combinatorial complexity computation computations computer computing conference connectionist continuous control convergence cued daphne data david dean decision demmel differences dimitri discounted discrete discretization distributed donald donato dongarra dupuis dyna dynamic edition efficiency efficient eijkhout eleventh eligibility englewood european examples exploration factored fifteenth first fixed fourteenth framework friedman functions game gavin generalization generalized geoffrey george gesammelte givan gordon graph graphbase graphs greedy guestrin guide gullapalli hall harold henk hermann heroux hutchinson hypergraph ieee imperfect incremental indirect infeng information intelligence intelligent international irregular iteration iterative iterator jack james jing john journal june kaelbling karypis kearns kevin kingdom knuth koller kumar kushner laboratory lambda learning leemon leslie less levinson line linear littman machine mahesan management markov martin mathematics mathematische mdps mellon method methods metrics michael mike miller minimization model modified moon moore multi multidimensional multilevel munos national near neural neuro nevin niranjan northeastern numerical observable official optimal pages parallel parr parti partially partitioned partitioning paul peng performance philadelphia philippe pittsburgh planning platform plenum points policies policy polynomial pozo predict prentice press preux prioritization prioritized priority problems proceedings processes processing programming propagation publishing puterman raymond real references reinforcement remi replacing report research resolution reynolds richard robert robot roldan romine ronald rummery saad sandia satinder scheme schwarz science scientific second seppi seventh shadid shin shobha siam simulation singh solution solutions solvers solving sons space sparse speeding springer stanford state step stephen stochastic stuart successful sutton sweeping systems tabular technical templates temporal thatcher thesis thirteenth thomas tight time tony traces transactions tsitsiklis tuminaro twenty uncertainty united university user using value values variable venkataraman verlag version victor vijaykumar vipin volume vorst weihong weiyu wiley williams wingate with within workshop york yousef zhang http://www.jmlr.org/papers/volume6/luo05a/luo05a.pdf 50 Active Learning to Recognize Multiple Types of Plankton accepted active advances algorithm algorithms analysis annual application applications association august baram based binary bootstrapping brinker cambridge campbell catch cauwenberghs chan chang chemical choice cjlin class classiers classification classifiers clustering cohn collected comparison computational computer concurrently conference convergence counter coupling cristianini csie cybernetics dags data decremental deep defense development discovery diversity document drug efficient empirical engineering error estimates estimation evaluation extraction feature feedback first florida forthcoming from gale goldgof gulf hall hastie high hopkins http ieee image images imaging incorporating incremental information intelligence international introduction japanese joint journal kearns kernel kernels klautau koller kramer langebrake large learning lemmen less letters lewis liao library libsvm likelihood linguistics machine machines margin mathieson mccallum meeting methods mexico mitra more multi multiclass multiple multispectral murata murthy nature networks neural nguyen noise northeast ocean oceanic oceanographic online onoda optical optimal orthogonal other outputs pages pairwise papers park part particle patten pattern pillar plankton platt poggio press probabilistic probability proceedings process processing profiling queries query ratsch recognition recognize recognizing recorder reduction references regularized relevance remote remsen remson research resolution retrieval rifkin sampling samson sassano scaling schohn scholkopf sciences segmentation sensing sequential shadow shadowed shankar shawe sigir smeulders smola society south springer statistical study support sutton system systems tang taylor text theory thesis through tibshirani tolerant tong toward training transactions twentieth twenty types university unlabelled using vapnik vector vectors version vision volume wang warmuth weng what with word yamada yaniv zhang zooplankton http://www.jmlr.org/papers/volume5/valentini04a/valentini04a.pdf 1 Bias-Variance Analysis of Support Vector Machines for the Development of SVM-Based Ensemble Methods aaai accurate adaptive advances algorithm algorithms allwein analysis annals applications approach approaches approximate arcing artificial austin automatic bacterial bagged bagging bailey bakiri bang bari bartlett based basu bauer baxter becker berkeley berlin bezdek bias bienenstock binary biostatistics boosting bounds bousquet breiman breva brief buciu burges cagliari california cambridge canada canu carson chan chapelle choosing classification classifier classifiers codes coding cohen combination combining comparing comparison complexity computation computer conference constructing correct correcting curse cybernetics data databases dean decision decomposition decompostion department detection development dietterich dilemma dimensionality discovery distributed diversity domingos doursat duin editor editors edwards effectiveness elisseeff empirical engineering enns ensemble ensembles error estimators evgeniou experimental experiments experts explanation explicit expression face fawcett feature features first forests foundation francisco freund friedman function functions fusion fuzzy geman gene general generalization generated grandvalet handwritten hatef heskes html huang hybrid icip icml ieee ijcnn improve improved information inns inspired intelligence international intrator introduction investigation issues italy james joachims joint journal kargupta kauffman kaufmann keller kernel kernels kittler kleinberg knowledge kohavi kolmogorov kong kotropoulos kuncheva langley large learning lecture library likelihood logic loss machine machines making marcialis margin margins mason masulli matas mathematically mcadoo measures medicine merz meta method methods mining mishra mlearn mlrepository model morgan mukherjee multiclass multiple murphy muselli national network networks neural neurobjects neurocomputing nips notes numerals obermayer object optimization oriented output pages pang parameters pattern perceptron perez performance pitas plus poggio pontil portland practical prediction predictors press preventive problems proc proceedings processing prodromidis radial random randomization recognition reducing references report repository research rigorous roli ruffino rules scale scaling schapire scholkopf science seattle second selection seventeenth shipp singer smola solving springer springerverlag squared stability stanford statistical statistics stolfo subsets subspace suen supervised support svms systems technical templates test theory thirteenth thomas three through thrun tibshirani toronto trans transactions trees twentieth unconstrained unified unifying university using valentini vapnik variance variants vector verlag vitanyi volume voting wang washington whitaker wiley with wolpert workshop york zero http://www.jmlr.org/papers/volume3/ling02a/ling02a.pdf 298 Using Confidence Bounds for Exploitation-Exploration Trade-offs accessible adaptive advances adversarial advertisements agrawal alamitos algorithm allocation alon american annual applied armed association associative asymptotically auer azuma bandit banner barto based berry bianchi bounds cambridge casino certain cesa chapman computation computer concepts conf confidence design disjunction eared ects efficient encer endent eriments exploitation exploration foundations francisco freund fristedt functions gambling generate hall herbster hoeffding holloway http ieee index inequalities information international internet introduction jority journal kaelbling kaufmann learning letin linear littlestone london long machine math mathematical mathematics mean method morgan multi nakamura neurocolt offs olicies online optimally ounded ounds pages preliminary press probabilistic probability problem problems proc proceedings random references regret reinforcement rigged robbins royal rules sample schapire schedule science sequential society some statistical sums sutton symposium technical test tohoku tracking trade university using variables version warmuth weighted wiley with york http://www.jmlr.org/papers/volume5/laub04a/laub04a.pdf 3 Feature Discovery in Non-Metric Pairwise Data acids adaptive advances agents altschul analysis annealing annual approach april archive arnold autonomous banerjee based becker biologically biomedical blast bollen braun buhmann cambridge chapman chicago class classification clickstream cluster clustering cognitive common component components compstat computation computer computing conference connectionism contextual contrast dagan data database denoising design designing designs deterministic dimensional dimensionality dimensions discussion dissimilarity distance distances distinctive distributed duda duin edition editors eigenvalue embedding engineering estimate everitt experimental extimation feature features framework from function gapped gati gdalyahu generalized generation gentner geometric ghosh global going goldstone goodness gutenberg hall hart headlines hesketh hofmann householder http human hypothesis ieee image immune independent information intelligence jacobs john journal judgments kawanabe kernel knowledge kruskal lange langford language laub learning linear lipman lkopf ller locally london longest machine madden management marcus mareschal markovitch medin meinecke methods metric miller mining model modified motivated multidimensional mutual nasa navarro neural news nonindependence nonlinear nonmetric notions nucleic obermayer optimal optimizing other pacl pages pairwise pattern pekalska perception performance points preserving press principles problem proceedings processing programs project promo protein proximities proximity psychological psychology psychometrika puzicha qualitative quantitative rabe recommendation reduction references relational representation resampling research retrieval rocha roth roweis saul scaling schaffer science search second segmentation shepard siam silva similarity smola society soft sons sparse speech stability statistics stimulus stork structure submitted subsequences system systems talkmine tenenbaum terms texture their theory thomas thrun torgerson transaction transactions trends tversky unknown unsupervised using validation version volume weighted weinshall wiley with word workshop york young zhang ziehe http://www.jmlr.org/papers/volume7/shivaswamy06a/shivaswamy06a.pdf 150 Second Order Cone Programming Approaches for Handling Missing and Uncertain Data academic adapting advances algebra algorithm algorithms anal analysis annals appl application applications applied approach approximate approximation australia available based bennett bernhard beyond bhattacharrya bhattacharyya biology bishop blake bottou bounded boyd brescia calafiore cambridge campi chandrasekaran chebyshev classification classifiers classifying climate codes collins computation computational cone cones conference connection constraint control convex cortes covariance data databases dempster design discriminative dynamic editor editors embeddings empirical engineering equations equivalent estimation farias formulation from function functions fung gaussian ghaoui golowich golub graepel graphical grate guestrin herbrich hidden html http huber hyperplane imputation incomplete inequalities inference information input integral interior invariant john jordan journal june kernels kluwer knowledge koller laird lanckriet language lawrence learning least lebret lectures leon likelihood linear lobo london machine machines mangasarian margin markov marshall math mathematical mathematics matlab matrices matrix maximum mean mercer merz method methods mian mika minimax missing mlearn mlrepository models modern molecular mozer multicategory multivariate national natural negative nemirovski nemirovskii nesterov networks neural nicta nips noise nonparametric number olkin oper operations optimization order over pages pannagadatta parameter pattern petsche philadelphia philosophical point polychotomies positive prediction presence press problems proceedings processes processing profiling programming pucci quantile quoc ratsch recognition references regression regularization report repository research roac robust royal rubin saira sampling saul sayed scenario schneider scholkopf sears sebastian second sedumi semidefinite separation shavlik shivaswamy siam signal smola society socp software solutions sparse squares statistical statistics studies sturm submitted support symmetric systems taskar technical their theory thrun tikhonov toolbox training transactions type uncertain uncertainties uncertainty universita university using values vandenberghe vapnik vector volume weiss wiley williams with yair zhang http://www.jmlr.org/papers/volume2/meek02a/meek02a.pdf 353 The Learning-Curve Sampling Metho d Applied to Mo del-Based Clustering aaai advances algorithm algorithms among analysis application applications applied august autoclass basford bayesian benefit cambridge cheeseman chickering clarendon classification clustering computationally computer conference cooper cost curves data decision dekker dempster department discovery discussion distributions domingos dynamic editors efficient eighteenth fayyad federation fifth finite fourth from general graphical heckerman herskovits hertz howard hulten illinois incomplete induction inference international jensen john journal kadie kaufmann know laird langley learning ledge likelihood machine macmillan makov marcel mateo maximum mclachlan meek meeting melese menlo method methods mining mixture mixtures models morgan networks oates operational oxford pages park pearce piatesky press probabilistic proceedings progressive provost references regression research results royal rubin sampling scaling science second seer selecting shapiro sixth smith smyth societies society sons speed static statistical statistics stutz theory thesis thiesson titterington university urbana uthurusamy valencia versus wiley with york http://www.jmlr.org/papers/volume6/zoeter05a/zoeter05a.pdf 97 Change Point Problems in Linear Dynamical Systems about accepted advanced advances aistats algorithm algorithms american annual applied approximate approximation approximations artech artificial association auai audio barber bayesian becker belief bernhard bishop cambridge cemgil change changepoint cluster conditionally conference constructing corvallis data dempster dept deterministic dietterich during duysens dynamic econometrics edition editors efficient elsevier energy estimation exact expectation factor fearnhead field forecasting francisco free freeman frey from gaussian generalized generative ghahramani graphical graphs handbook harrison heskes house human hybrid ieee incomplete induced inference information intelligence intelligent international intro iteration jaakkola john journal july kappen kaufmann krishnaiah kschischang laird lancaster lauritzen lawrence likelihood linear loeliger markov math maximum mean means mechanically method methods miao minka mixed model models morgan morphing mulder multiple multivariate music networks neural neuroscience ninth novel onno oregon pages pearl points practice press principles probabilistic probabilities problems proceedings processing product propagation publishers reasoning references regions report review royal rubin saul schemes schillings scholkopf sebastian shalom shumway society software sons space speech springer stat state statistical statistics steffen stevens stoffer structured stumbling submitted switching systems technical techniques theory thrun tracking transactions transcription treadmill tree tutorial uncertainty university variances variation variational volume walking weiss welling west wezel whittaker wiegerinck wiley with workshop yedidia zoeter http://www.jmlr.org/papers/volume7/wainwright06a/wainwright06a.pdf 169 Estimating the "Wrong" Graphical Model: Benefits in the Computation-Limited Setting adaptive advanced albers algorithm algorithms analysis annealing approximate approximation approximations april arbitrary artificial arxiv athena august baraniuk based bayesian belief belmont benveniste berkeley bertsekas besag bethe beyond binary biometrika boltzmann bounds brown cambridge capacity carlo carmichael casella check class codes computation computer computing conditions conference constrained constructing convergence convex correctness crouse cuts data decoding della densities density department determinant deza directions discrete distributions editors effects efficiency embeddings energy errors estimating estimation experiment exponential families family features february field fields finite fisher fitting framework free freeman from function fundamentals gaussian generalized geometry gibbs gibbsian graphical graphs hayward henri heskes hidden higher idiosyncratic ieee ihler improving inducing inference info information inst institute intelligence intl iterative jaakkola january jordan journal journey july june kaebling kappen lafferty lattice laurent lauritzen learning leisink level linear local loops loopy machine machines makov markov matching mathematical mccallum mean measures mechanics message messagepassing methods metivier metric minka mixture models monte mooij motion national networks neural nijmegen nips nonlinear nowak object opper optimality optimization order oxford pages parity partition pasztor pattern piecewise pietra poincare practice press princeton principle priouret probability proc procedure processing product programming propagation properties proportional pseudolikelihood pseudomoment random references related relaxation reparameterization report research response reweighted richardson ripley robert rockafellar ross rusmevichientong saad scientific segmentation september serfling series signal simple smith solutions spatial springer static statistical statistician statistics stochastic strategies submitted sufficient sutton tatikonda technical theorems theory thesis titterington training trans tree turbo uncertainty under undirected university upper urbanke using variational verlag vision volume wainwright wavelet weiss welling wiegerinck wiley willsky with workshop yedidia york younes http://www.jmlr.org/papers/volume6/brown05a/brown05a.pdf 85 Managing Diversity in Regression Ensembles aarhus abbass academy active activity adaptive ambiguity analyzing annals anticorrelation application approaches artificial austin australia australian averaging best between bias bienenstock birmingham brain brown california canberra categorisation center chemical classifier classifiers college combiners combining computation computer conference connection convex correlated correlation creation cross datalogisk decompositions decorrelated defence dilemma diversity doursat ensemble ensembles error estimation estimators evolutionary examining experimental expert extensions extremes finance fischer flow force foundations friedman fumera function fusion geman general generalization ghosh guildford hansen harris higuchi ieee improving information institut institute intelligent international issue joint journal krogh kuncheva learning linear lncs losel machine markowitz mckay measure measures meta methods monterey multiple multivariate nabney nakano negative negatively network networks neural nips objective optimization order otago overfitting pages pattern perrone portfolio prediction predictors press proc proceedings processing produce quantitative reduction references regression relationships report research results roli rosen school science sciences selection sollich some south special splines springer statistics structure surrey survey systems technical telecommunications texas their theoretical thesis tino traffic transactions tumer ueda universitet university useful using validation variance vedelsby vision wales washington whitaker williams with workshop wyatt zealand http://www.jmlr.org/papers/volume3/bach02a/bach02a.pdf 360 Kernel Indep endent Comp onent Analysis acoustics adaptive adaptivity additive advances akaho algorithm algorithms alternating amari american analyse analysis annals applications approach approximation architecture architectures archive arias artificial association asymptotic attias bach baker baltimore based becker behavior bell belmont between bickel biometrika bischof blind borga boscolo bound breiman brezis buja cambridge canonical cardoso case certain chapman chebyshev cholesky cichocki clarendon classification classifiers comon component computation computations concept conference constrained constraints continuous contrasts cormen correlation correlations cover cristianini curves data deconvolution density dependent dietterich distribution durrett duxbury edelman editor editors effect efficient efficiently eigenvalue eigenvalues eighteenth elements equations estimating estimation examples extended extraction factor factorizations fast feature fine fixed fonctionel francisco friedman functional fyfe garat gaussian generalized geometry ghahramani girolami girosi golub gradient greedy hall harlow harmeling hasselmo hastie have herault high hopkins hotelling icann icassp icml ieee image imps independent infomax information input integral intel interior international introduction john johns jones jordan journal jutten karhunen kaufmann kawanabe kernel kernels kettenring klaassen knutsson kolmogorov kullback landelius langley larger latent learning least leen leiserson leurgans ligence ligent likelihood linear lkopf loan lodhi london longman machine machines masson mathematical matrix maximization maximum mechanics meeting melzer method methods mixed mixtures models modified morgan motomura moyeed mozer muller multidimensional multiple natural networks neural neuromimetic nonlin nonlinear note numerical nystr optimal optimization order orthogonal orthogonality oxford parametric paris part pham poggio point polynomials press probability problem proceedings processing programming providence psychometric quasi rank rational references regression regularization reiter relations remarks representations reproducing research rinen ritov rivest robust royal roychowdhury saitoh saunders scandinavian scheinberg scia scientific seeger sejnowski semantic semiparametric separation sets seventeenth several shannon shawe siam signal signals silverman smith smola society sons source sources spaces sparse speech speed springer squares statistical statistics string super systems szeg taylor technical tenth text than that theory third thomas through thus tibshirani tokyo touretzky training transactions transformations transmission treatment tree tresp uncertainty unit university unlikely using vapnik variables variance variates verlag vlassis watkins weber welling wellner when widom wiley williams with works workshop wright yang york ziehe http://www.jmlr.org/papers/volume2/bousquet02a/bousquet02a.pdf 357 Stability and Generalization academic advances algorithmic algorithms alon analysis annals approximate april artificial bagging bartlett based bianchi bonnans bound bounded bounds bousquet breiman building cambridge carnegie cesa check classification combinatorics computation convergence convex cover crossvalidation data david december decision deleted department dependences devroye differences dimensions discrimination distribution efron elements elisseeff empirical entropy equivalent error estimate estimates estimation evgeniou exponential finite framework free function functional generalization girosi gordon guided haussler heuristics hilbert holdout ieee important independence inequalities inequality information inria instability institute intelligence introduction jaakola jebara john journal kearns kernel kluwer laboratory learnability learning leave local look lugosi machine machines madison markov massachusetts maximum mcdiarmid meila mellon memo method model more multilayer network networks neural nonparametric nonsymmetric optimization pages pattern pawlak performance perturbation poggio pontil posterior potential predictors press princeton probabilistic probability problems processes processing publishers recognition references regularization related report reproducing rockafellar rogers rules sample sanity scale science selection sensitive shapiro size solutions spaces springer stability stabilization statistics steele stein support surveys systems talagrand technical technology than that theory thesis thomas topics tour transactions unified uniform university valid vapnik vector verlag wagner wahba weights wiley wisconsin with http://www.jmlr.org/papers/volume7/begleiter06a/begleiter06a.pdf 116 Superior Guarantees for Sequential Prediction and Lossless Compression via Alphabet Decomposition academic adaptive addison aggregating alan algorithm algorithms allwein almost alphabet always amnesia analysis annual application applied approach architecture artificial asia association asymptotic asymptotically authorship automata barron based basic begleiter bejerano bell benelux best biased binary bioinformatics biometrics biometrika block bounding break burrows calculus catoni chen cheong cipher classification classifiers cleary coding coin comparisons complexity compression computation computational computer conference context convergence corporation courant cover dale data decision decomposition decompositions defense denoted determination digital discrete during edition efficient effros eindhoven elements empirical encoding encyclopedia enigma equipement essays estimates estimating estimation estimator estimators etzion experiments explanation extensions families feder frequencies frequency freund from functions gambling general generating georgia glassey good goodman grammar greedy hall haussler helmbold hierarchical hodges huffman ieee imai implementing incremental individual info information institute integer intelligence introduction intuitive israel issn jacquet john journal karp kieffer kivinen klautau krichevsky kulkarni language laplace learning lecture lempel length letters linguistics logistics loss lossless machine machines march margin markov mathematics maximal mcallester meeting memory mention merhav messages method methods minimax modeling models motivation multi multiclass national nearly neural never notes novelevents october okazaki optimal optimality optimization order orlitsky pacific page pages papadimitriou parameters parsing part performance philosophical plouffe pointwise population potapov power practicality predicting prediction prentice press principles probabilistic probabilities probability problem problems proc proceedings processing properties protein pruning published purity ranking rate realizing reducing redundancy references regret report research reviews rifkin rosenberg rule sadakane santhanam savari schapire science september sequence sequences sequential shkarin shtarkov siam simple simultaneously singapore singer single sloane smoothing society sons sorting source sources species springer statistical step stochastic strategies study suffix superior support symp szpankowski teahan tech technical technion technique techniques technische technology text that theory thesis thirteenth this thomas tishby tjalkens transactions transform translated transmission tree trees trofimov tsui turing types under unfortunately unifying universal universiteit university used using variable variations vector verdu verlag visweswariah volf vovk walker wang warmuth washington weighting well wesley wheeler wilcoxon wiley willems with without witten workshop world yang yaniv yona zero zhang http://www.jmlr.org/papers/volume7/yanover06a/yanover06a.pdf 171 Linear Programming Relaxations and Belief Propagation ­ An Empirical Study accuracy advanced advances agreement algorithm algorithms allerton alternative analysis analyzing application approaches approximate aritifical artificial athena back baker based bayes bayesian belief bertismas bioinformatics biol biology birchfield bixby bonding bone boykov canutescu chain chazelle chivian close combinatorial communication complexes complexity compter comput computational computer computers computing conf conference control convergent correspondence cowell cuts cvpr decade dechter decoding dense dependent depth design dimacs dissimilarity distinguished dunbrack early eccv editor effective efficient energy estimation evaluation explanations fast feldman felzenszwalb finding folding fraenkel frame freeman function functions generalizations generation genetics glass globally graph graphical hard held high homologous huttenlocher hydrogen hyper iccv identical ieee ijcai image implications improves inference information insensitive integer intelligence intelligent international introduction jaakkola jordan journal june karger karplus kask kaufmann kingsford kolmogorov kortemme kuhlman kurplus lazaridis learning lecture library light linear machine maps marinescu matching matlab mayo measure meltzer message minimization minimized modeling models molecular more morgan morozov murphy native networks neural operations optimal optimality optimization orientation pages parameters passing pattern pearl pennsylvania pierce pixel plausible pnas polynomial positioning potential predicrtion prediction press probabilistic problems proc proceedings processing product programming programs progress propagation protein proteins rapid real reasoning recog recognition references regions research revision reweighted rohl rosetta rotamer sampling santos scharstein science scientific sequences shelenkov shimony shum side sidechain singh society solution solutions solve solving specificity spin statistics stats stereo strauss street strongly structurally structure structured structures systematic systems szeliski tappen tardos task taxonomy that their theory tomasi toolbox track transactions tree trees ttsitskikilis uncertainty understanding university using variable veksler vision volume wainwright weiss what willsky winfree with workshop world yanover yedidia zabih zheng http://www.jmlr.org/papers/volume6/markatou05a/markatou05a.pdf 69 Analysis of Variance of Cross-Validation Estimators of the Generalization Error absolute accuracy advances algorithmic algorithms american analysis annals apparent applications approximate approximation artificial association asymptotic asymptotics australian bengio between bias bickel biometrika bootstrap bound bounds breiman calculation calculus chapman check classification classifier comparing computation computing conditional conference consequences cook covariance cramer cross data deviations dietterich discriminant distributions doksum dynamic efron elements error errors estimated estimating estimation estimator expansion expectation experimentally fold friedman function functions general generalization grandvalet hall hastie heuristics higher hitomi improvement inference information instability intelligence international introduction ioffe james japan joint journal kagihara kearns khan kohavi large learning least leave lehmann letters limiting linear loss machine markatou mathematical mclachlan mean measured methods mining misclassification model models nadeau neural nonlinear order pattern penalties penot picard piper point predication prediction prentice press princeton printing processing rate rates recognition references regression research risk ronchetti rule sample sanity selection singer society sons split springer stability stabilization statist statistical statistics study subhessians subjets supervised systems test tests their theory tibshirani training transactions unbiased university validation variability variance ventura verlag wiley with http://www.jmlr.org/papers/volume6/murphy05a/murphy05a.pdf 67 A Generalization Error for Q-Learning achim acute advances advantage algorithm alternative altfeld altshuler america american analysis annals annual anthony antipsychotic approximate approximation artificial association athena australia automatic available background baird bartlett barto base based baxter behavioral bellman belmont bertsekas biggs boosting bound brooner brunswick cambridge carmody catalog catalogrev catie chronic claude clinical clinics college colt complexity computational conference control convergence courses crismon dagerman davis death decision delayed depression design difference douglas dynamic effectiveness efficacy efficient electronically empirical estey estimation evaluating expected experience explanation fava feature fiechter fisher force foundations fourteenth freund from function generalization geriatric gradient group health hoffmann horizon hsiao http icml ieee improve infection infinite information institute intelligence international intervention introduction jeste journal kakade kashner katz kearns kernel kidorf kupfer large lavori learning lebowitz less lieberman line london lyketsos machine mailing mansour manuscript margin markov mass medication medicine mental methadone methods miller millikan mistake model more multicourse multiple murphy nashville national nature near network neural neuro nierenberg nineteenth north olin optimal ormoneit pages papers participation patterson perspectives peshkin phase planning policies policy pollock pomdps practice press princeton proceedings processes processing programming project psychiatric psychiatry quitkin rabins rago rationale references reinforcement relieve report research reusable rewards rosenbaum rosenheck rush sackeim sammut sample sampling scale scarce schapire schneider science scientific selecting sequenced seventh shelton shon shores simester small sparse springer star statistical statistics strategies study sung suppes sutton sydney systems tariot technical temporal tennessee texas thall thase theoretical theory therapeutic thesis tmap toprac trajectories transactions treatment trials trivedi tsitsiklis university unpublished updating using vaart voting walker watkins weak wellner wilson wisniewski with witte wright york http://www.jmlr.org/papers/volume7/mukherjee06a/mukherjee06a.pdf 121 Learning Coordinate Covariances via Gradients acad acosta advances algorithm algorithms amer american anal analysis annual appl applications approach approx approximation approximations aronszajn atomic banach barnhill basis bayesian belkin bernardo beyond biology bloomfield bounds bousquet bull caligiuri cambridge cancer capacity caponnetto carrel chapelle chen choosing chromosome class classification classifiers coglin coller comput computation computational computing conference connections constr cordero correction cortes cottle cross cucker cytogenet data decomposition discovery disteche distributions donoho downing duke durham ebert editor enrichment equivalent error escape estimates estimating estimation evgeniou expression factor first flippova foundat foundations from function functions gaasenbeek gene generation genetics genome gillette girosi golub guyon harmonic huard human identified ieee inactivation incativation index inform integral international interpreting jacks journal kernel kernels knowledgebased kras ladd lander large lasso learning least leave liang liao linear machine machines manifolds martingales math mathematical mathematics mesirov methods micchelli microarray model models molecular monitoring monthly mootha mukherjee multicategory multilayer multiple natl nature networks neural niyogi objective oncogenic operators optimization optimum oxford pages paradigm parameters paulovich pinelis poggio point pomeroy pontil prediction preprint press probab proc profile profiles programming pursuit quadratic radial radiance recomb reconstruction references regression regularization regularized reproducing riemannian roix rosasco royal sampling satellite saunders schoelkopf science scientific selection semi shannon shrinkage siam signature slonim smale small smola society some spaces species splines squares stat statistical statistics subramanian supervised support sweet tamayo that their theory thesis tibshirani trans transactions tsuchiya university using validation valued values vapnik variational vector versus vito wahba weather wendelberger west weston wide wiley willard with york zhang zhou http://www.jmlr.org/papers/volume2/lodhi02a/lodhi02a.pdf 354 Text Classification using String Kernels acquaintance adatron adavances advances algorithm alignment annual appear approximation automatic available bartlett based berlin boser burges california cambridge campbell categorization cavnar claire classification classifiers communications computational computer conference connection convolution cristianini cruz dellec department dietterich discrete document doktorarbeit dynamic editor editors elisseef equations euroe fast feature features friess from functions gaithersburg gram grams greedy guyon harman haussler holloway html http huffman ieee independent indexing information infromation input institute integral international introduction joachims july kauffman kaufmann kernel kernels knirsch langley language large learning leen line lodhi london machine machines making many margin maryland matrix mercer method methods mika model morgan muller munchen national nature negative networks neural nips nist nystrom oldenbourg optimal optimizing pages pean philosophical pittsburgh positive practical press procedure proceedings processing pubs ratsch references relevant report representation retrieval rouveirol royal salton santa scale scholkopf schuurmans science seeger seventeenth shavlik shawe simple smola society space sparse speed springer standards statistical string structures support system systems target taylor technical technische technology text their theory training transactions trec tresp tuebingen type ucsc universitat university using vapnik vector verlag voorhees watkins williams with wong workshop yang york http://www.jmlr.org/papers/volume5/ziehe04a/ziehe04a.pdf 2 A Fast Algorithm for Joint Diagonalization with Non-orthogonal Transformations and its Application to Blind Source Separation abed adaptive advances airoldi akad akuzawa algebra algorithm algorithms alive amari anal analysis angle angles appl application applications applied approach approaches approximate architecture array articles artifact artifacts artificial august aussois australian azimuth based beamforming bell belouchrani benster berlin biological biomed biomedical blind bunse byers californicus cambridge cardoso certain chaos characteristic chen choi cichocki circuits citeseer cliffs cocktail colored combining common comon component components computation computing concept conf contrasts convolutive correlation correlations cortical cranial criteria curio cybernetics daniel data david decorrelation definite delay delayed deprettere diagonal diagonalization diagonalizers differential discussion domain dynamics editor editors efficient eigenproblems elecom electroencephalographic elevation elsevier engineering englewood equations equivariant estimation eusipco fast fengler fields filtering finland fitting flannery flury form fractals france frequency function gaussian gautschi generalization gerstner golub hall hamalainen handbook hardle hari harmeling hasselmo haykin helsinki herault high hopkins hori horn html huang humans humboldt hyvarinen icann icassp identifiability identification ieee ieice ikeda implementation implied independent indeterminacy information inst instantaneous intern invariance invasively jacobi january japan john johns johnson joho joint jordan journal jousmaki july jung jutten kantorovich karhunen kawamoto kawanabe kearns kernel laheld laskov learning least lecture letters levenberg like linear loan london lutkepohl mackert magnetic magnetoencephalographic magnetoneurography makeig malsburg mammalia march marquardt mathematical mathematics matrices matrix matsuoka mehrmann meraim method methods michael microtus mixture mixtures models molgedey more moreau morphometry moulines mozer muller multichannel multiple multiplicative multivariate murata nara nauk nearly network networks neural neurocomputing neuromimetic newton nips nnsp noble nolta nolte nonholonomic nonlinear nonstationary notes numerical ober ochrogaster ohya optimal order orglmeister orthogonal osborne ovde oxford pages paper parameters paris parra part party paulraj performance perturbation pham physical positive prentice press principal principe problem problems proc proceedings processing processor quarterly quasi rahbar recipies recorded recordings reduction references related relative report resolution review revisited rodentia schneider schuster scientific sdif seattle second sejnowski selected sense sensor separation series several shift siam signal signals simultaneous singapore society solitons solla solution solving sons soon souloumiac source sources speech spence springer squares sssr state stationary statistical statistics steklov still structure subspace sweden symmetric symp systems tdsep technical technique techniques temporal temporally teukolsky theory time tong touretzky towards trahms trans transactions transformation translation trudy universitat university unsupervised using vanderveen vandervorst veen verlag vetterling vigario villa volatilities volume vorst watson wiley with workshop wubbeler years yeredor york zibulevsky ziehe zoology http://www.jmlr.org/papers/volume7/kim06a/kim06a.pdf 136 Segmental Hidden Markov Models with Random Effects for Waveform Modeling aaai aaron achan acoustics advances agrawal aided aksmanovic algorithm algorithms american analysis anemometry antti application applications approach arbitrary arthur artificial artur arun association auai audio australia automated automatic banff based bases bayesian beach bennett berkeley biing bioelectromagnetism biometrics biometrika boston brendan bruun bull byoung cairo california cambridge canada carl casella chan charles chee chicago christos chuanhai classification cliffs complexity components computational computations computer conference continuously convergence correspondence covariance cued daniel data databases datamining deformable dempster deng department digalakis discovery discrimination donald duration dynamic eamonn ecme editor effects efficient egypt engineering englewood ensembles estimation events explicit extension faloutsos fast faster feature ferguson foundations framework frey from functions fundamentals gales general generative george graphical hall hans harper hertzmann hidden hmms holmes hughes hwan hwang ieee incomplete indexing infeng inference information integrating intelligence international interval intrator james jamieson jankowski jeff john jordan journal juang kannan kaufmann keogh kevin kimball king knowledge koski laird lange language large lawrence leah learning letters levinson likelihood lionel longitudinal luther machines mari mark markov martin mary matching maximum mcculloch measures medicine meng michael mike mining minneapolis mitchell model modeling modelling models monotone moore morgan murphy nathan networks neural newport nicholas nonstationary norms occupancy oreziak organization ostendorf owen oxford padhraic pages pattern pazzani personal philadelphia polynomial pong prentice press princeton principles probabilistic proc processing rabiner rakesh random recognition references regional regression repeated report representation robert roberts roger roweis royal rubin russell scaling search searle segment segmental seis seismic sequence series seyoung shapes shayle shimshoni sigkdd signal signals similarity simple smyth society speech springer stanislaw state states statistical stefan stephen steve stochastic stram support swami sydney symposium system systems tampa tarassenko technical templates text theory theron thesis time trajectory trans tsutakawa uncertainty unified united university using variable variance vassilios vector verlag very view volume ware warping waveform wavelets wendy western wiley wire with xianping xiao xiaodong yair york young yung http://www.jmlr.org/papers/volume7/bhatnagar06a/bhatnagar06a.pdf 173 A Simulation-Based Algorithm for Ergodic Control of Markov Chains Conditioned on Rare Events actor actorcritic adaptive admission ahamed algorithm algorithms among analysis annals appl applied approach approximation approximations areas artificial asynchronous athena auto automatic automatica average avoidance balaji bartlett barto based baxter belmont bertsekas bhatnagar borkar boston brandiere bucklew california call cambridge cases chain chains chong communications computer conditioned conference cont contr control cost countable criteria critic cybernetics data dayan decision deterministic deviations diego discounted discrete distributed dynamic dynamical engg ergodic ergodicity estimation european event events experiments fakir form formula gallager geometrically gradient hall hernandez hidden hill horizon ieee importance infinite infinitesimal info information integrated intelligence inter international introduction irreducible jersey john journal juneja kelton kherani kluwer konda kontoyiannis kumar large learning letters like limit load machine marbach marcus markov mathematics mcgraw measurement meyn mihatsch modeling modelling models monotone multichain multiplicative multiscale multivariate networks neuro nowe operations optim optimal optimization parametric pathological performance perturbation policy potentials prentice press prob probability problems proceedings processes programming puterman ramadge ramakrishnan rare reachable references regenerative reinforcement reinforcing relations research reward risk routes routing rubinstein sampling scale scheme science scientific selected sensitive sensitivity sequences services siam simulation simultaneous some space spall spectral state steenhaut stochastic sutton system systems technique techniques their theorems theory thirunavukkarasu three time timescale timescales trans transactions traps tsitsiklis unified using varadarajan veerbeck wang watkins weaver wiley with york http://www.jmlr.org/papers/volume7/schraudolph06a/schraudolph06a.pdf 143 Step Size Adaptation in Reproducing Kernel Hilbert Space aacf aberdeen acceleration accuracy active adaptation adaptive adatron additive advances advantage aeronautical afwal algorithm algorithmic algorithms almeida altun amaral analysis anderson annual appear appendix applied approximate approximation articulated artificial assessment auer automata azoury back backpropagation baird barto base based beats bennett berlin bernhard bertinoro bianchi bishop block boosting bordes boser bottou bounds brain braun bray brooklyn budget burges california cambridge campbell canada categorization cesa chapter classification classifiers code codes component components computation computational computer computing conditional conf conference control convergence cortes crammer cristianini cruz curvature david december dekel denker dense density derivatives descent design differentiation dimensional direct discrimination distribution distributions document edinburgh editor editors eidgenos ertekin estimating estimation eurasip evaluating even exact exponential exponentiated families family fast faster fields fixed flow force forgetron francisco franz freund frie frontiers function functioning gain game gaussian general gentile geometry giannakopoulos goal goldman good gool gradient griewank handwritten harmon helmbold henderson herbster hessian hierarchical high hochschule hofmann howard http hubbard ieee image improving increased independent information initial inseparable institute intelligence interdependent international intl introduction isbn issue italy iterative jackel jacobs january joachims john jordan journal june kandola kaufmann kernel kernels kevin kivinen knowledge koller laboratories laboratory langlois large lawrence learnability learning least lecture lecun leemon leen line linear linearly littlestone local london long loss machine machines mackay management mangasarian maps margin mark mathematical mathematics matrix maximal maximum meier meir method methods milano minsky model modeling models modern morgan mozer muller multi multiclass multiplication murphy networks neural neurocomputing newton nicol norm notes novikoff number offline ohio online optimization order organization output pages papers papert parameter parameters pattern patterson pearlmutter perceptron perceptrons perspective petsche philadelphia physics plakhov platt player policy polytechnic predictors press principal principles probabilistic problems proc procedure proceedings processes processing products programming proofs propagation properties psychological publications publishers random range rate rates recognition references relative relaxed report reprinted research residual review reviews riedmiller robust robustness rosenblatt rosenfeld rprop saad santa saul scaling schapire schmidt scholkopf schraudolph science sciences scotland sebastian second seeking september sets shalev shavlik shawe shwartz siam signal silva simple singer sische smola software solla spaces special speed springer squares states statistics step stochastic storage structure structured supersab support sutton switzerland symposium system systems taylor tech technical techniques technische theoretical theory thesis thirteenth through thrun tighter tollenaere tracking training transactions tsochantaridis ualberta ultraconservative uncertainty united university using vancouver vapnik vector verlag versus vishwanathan vision volume warmuth weak weiss wellekens weston williamson with workshop wright wrightpatterson yair yale york zurich http://www.jmlr.org/papers/volume5/dy04a/dy04a.pdf 5 Feature Selection for Unsupervised Learning aaai aberdeen abstract academic acquisition advances agrawal aisen akaike alberta algorithm algorithms almuallim alto american amherst anaheim analysers analysis analyzers annals annual appear application applications applied approach artificial association atkins attention attribute attributes august autoclass automatic banff bankert based basford bayesian beal before behavioral berkeley biometrics bishop black blake bled book boston bouman bound bradley branch breiman broderick brodley brunswick california cambridge canada cardie carlo caruana case cases cbir cdsp chang cheeseman cheng chicago chichester classification classifications classifiers cliffs cloud cluster clustering clusterings clusters cognitive collins company comparison component components computation computer computers computing concept conceptual conference content control convergence cook criterion cross customized data databases davis december decision dekker dempster density dependencies description devaney diego dietterich dimension dimensional dimensionality discovery discretization discriminant distributions doak document dubes duda dynamo edition editors effectiveness efficiency efficient eleventh empirical englewood erlbaum error estimating estimation european evaluation evolutionary expectation exploratory extending extensions factor fayyad feature features february fifteenth fifth figueiredo finite fisher florida forgy formation fort fourteenth fourth fraley francisco freitag friedman from fukunaga gaussian gehrke gennari ghahramani greedy green gunopulos hall hart hartigan hierarchical high hill hinton html http huber hyvarinen identification ieee images improve incomplete incremental independent induction inference information initial initialization inside institute integers intelligence intelligent interactive internal international interpretability introduction irrelevant iterative jain john johnson journal june kambhatla kaufmann kira kittler knowledge kohavi kononenko krishnan laird langley lauderdale lawrence learning lecture leen length likelihood local look looking lung machine magazine makov management many marcel march marill mathematical maximization maximum mcgill mcgraw mclachlan means measures meeting menczer merz methods meulman milligan minimum mining mirkin mixture mixtures mlearn mlrepository model modeling models modern monte moon moore morgan muller multivariate nakano narendra nashville national neural ninth normal northeastern norvig notes november number objects oblivious october optimization order pages palo pattern pazzani pelleg pfleger points portland practical prentice preprocessing press principal prior probabilistic problem proceedings processing projection properties provan psychometrika purdue pursuit queries raftery raghavan reasoning receptors recognition reduction references refinement refining reina relief rendell report repository research results retrieval rissanen river royal rubin russell saddle sage salton scene schwarz science scotland search searching seattle second security selection selective separating september series setiono seventeenth shapiro shyu sigkdd sigmod signal simoudis simplification singh sixteenth sixth slovenia smem smith smyth society software solla sons square stanford statistical statistics statwww step stork street study stutz subset subsets subspace survey surveys systems talavera technical tenth their theory thirteenth thirty tipping titterington toronto transactions trees twelfth types ueda universal university unsupervised users using vaithyanathan validation vancouver variational vision visualization volume wald washington wichern wiley with wolfe working workshop wrappers york http://www.jmlr.org/papers/volume4/botta03a/botta03a.pdf 264 Classes of Kernels for Machine Learning: A Statistics Persp ective american analysis angeles annals appear appl applications applied aronsza assessment association atoires based bochner brockwell brownian brownien calculus california cambridge christakos classes communications compactly completely composition connection consequences correlation covariance cressie cristianini data davis deformation dependent design dimensional ellis entifique environmental equations estimation exponentielle fields fonctions fractional functions gauthier genton geostatistics gilbert gneiting graham guttorp harmonic homogeneous horwood huang implementation information integral interpolation intrinsic introduction isotropy john journal kernel kernels kriging kronecker learning letters limited local locally london machines mandelbrot matching mathematical mathematics matheron matlab matrices matrix mercer methods metric models modern moler monotone motions mouvement multivariate nature ncsu negative ness noises nonparametric nonseparable nonstationary norm ordre orthogonale other oxford papanicolaou paris permissibility permissible perrin philos positive press probab probability problem processes processus products pure random reducing references related report reproducing research resources review revue risk sampson schoenberg schreiber scie second senoussi series shawe siam silverman some sons space spaces sparse spatial spatio spatiotemporal springer stationarity stationary statistical statistics stein stochastic stochastiques structure support supported taylor technical temporal their theorem theory time trans transactions type university using vapnik variation variogram vector villars water weakly wiley with yaglom york http://www.jmlr.org/papers/volume7/glasmachers06a/glasmachers06a.pdf 156 Maximum-Gain Working Set Selection for SVMs aaai active advances algorithm algorithms annual applied auer based bengio beyond bhattacharyya blake bordes bottou burges cambridge chang chapter chen cheung cjlin classifier classifiers colt conference convergence core cristianini csie data databases decomposition design document editors ertekin fast fawcett freund general generalized gilbert giles girosi glasmachers gradient haffner html http hush icml ieee improved information international introduction iterative joachims john journal keerthi kernel kernelbased kernels kwok large learning lecun leon library libsvm list lncs machine machines making meir merz method methods minimal mishra mlearn mlrepository morgan murthy murty nearest networks neural nishi online optimization order osuna other pages peter platt point polynomial practical press principe proceedings processing proof recognition references regularization repository research rigorous scale scholkopf scovel second selection sequential sets shawe shevade signal simon simplesvm singer smola springer statistical support takahashi taylor termination theorem theory time training transaction transactions tsang twentieth university using vapnik vector verlag very vishwanathan volume weston wiley wilson with working yoram york http://www.jmlr.org/papers/volume3/chickering02b/chickering02b.pdf 374 Optimal Structure Identification With Greedy Search academy acyclic algorithm andersson annals artificial automation background ball bayes bayesian belief besnard bouckaert breese buntine carnegie causal causation characterization characterizing chickering choice classes cognitive combination compelled complete computer conference consistent construct cooper cpdag czech darwiche data department determining diagrams digraph digraphs each edge edges editors eighteenth either eleventh else engineering enumerating equivalence every exists explanation exponential extension families family figure finding fisher fourteenth friedman from geiger gillispie glymour goldszmidt goto graph graphical greedy guide hanks haughton heckerman henrion herskovits identification ieee immediate implementation incident inclusion induction inference influence information input institute intel into irrelevance jeffreys journal kanal kaufmann king know knowledge koka koller label labeled laboratory leads learning ledge lemmer lenz ligence ligent literature lowest machine madigan markov mateo meek mellon method microsoft model models moral morgan msrtr network networks optimal order ordered oriented output oxford pages parent partially pastime pearl perlman plausible prediction press probabilistic probability problem proceedings rational reasoning references report republic requisite research reversible scheines science sciences search selecting selection seventeenth shachter simple sixth spirtes springer statistical statistics stratified structure structures studeny such synthesis systems tarsi technical that theory there thesis transactions transformational tutorial ucla uncertainty university unknown using verlag verma which while with york http://www.jmlr.org/papers/volume7/jonsson06a/jonsson06a.pdf 184 Causal Graph Based Decomposition of Factored MDPs about abstraction abstractions actions active advances agency algebraic algorithm amherst analysis animals animats application approach artificial aubin automated automatic barto based bayes bayesian bellman berkeley between boutilier bradtke california cambridge causal causation chentanez clustering complex computational computer conference construction continuous control creating dean dearden decision decomposition density department diagrams dietterich digney discovering discovery diverse domains duff dynamic efficient emergent environments estimation european exploiting factored feng fikes finding formalism framework from function generalization ghavamzadeh givan goals goldszmidt graph guestrin hansen harel hauskrecht helmert hengst heuristic hexq hierarchical hierarchies hierarchy hoey hoze identify identifying information intelligence interaction international interscience intrinsically introduction jaakkola joint jonsson journal kaelbling kanazawa kearns klein koller large learning line local machine machines macro mahadevan mannor markov massachusetts mathematical maxq mcgovern mdps mechanics menache methods meuleau minimization model motivated murphy national networks neural nilsson norm novelty options parameter parr partitioning persistence pickett planning policy policyblocks precup press problem problems proceedings process processes processing programming projections proving puterman ravindran reactive real reasoning references reinforcement relationships relative report research russell scheduling schwartz science search semi shimkin simsek singh solution solving special spudd state statecharts steck stochastic strips structure structures subgoals sutton symbolic systems technical techniques temporal theorem thesis thrun time tong tree uncertainty university unsupervised useful using value visual volume wiley with wolfe york zilberstein http://www.jmlr.org/papers/volume6/binev05a/binev05a.pdf 74 Universal Algorithms for Learning Theory Part I : Piecewise Constant Functions academic acta adapting adaptive alternative amer anal annals appear application approximation artificial assoc asymptopia baraud barron basis belmont berlin best binev birge bounds breiman bulletin carolina cart classification cohen comp complexity computation connection cucker dahmen daubechies density design devore distribution donoho encoding entropy error esaim estimates estimation estimators fast foundations free friedman functional gaussian gyorfy harm international johnstone journal kerkyacharian kluwer kohler konyagin krzyzak learning lectures likelihood lower massart math mathematical maximum methods minimax model network neural nonlinear nonparametric numerica numerische olshen optimal ortho pages penalized philadelphia picard preprint preprints prob publishers random references regression regularization related roussas royal selection shrinkage siam smale smoothness society some south springer statist statistical statistics stats stone structured temlyakov testing theory thresholding topics tree trees universal university unknown wadsworth walk wavelet wavelets with http://www.jmlr.org/papers/volume3/weston03a/weston03a.pdf 287 Learning Precise Timing with LSTM Recurrent Networks adaptive advances agtm algorithm algorithms applications applied architecture architectures artificial back backprop backpropagation based bengio berlin blues bottou brauer bulsari calculations cambridge carnegie carrasco chapter chauvin complexity computability computation computational computer computing conf context continual continually continuous control cottrell cued cummins department dependencies descent difficult difficulty diploma doya driven during dynamic dynamical dynamischen editors efficient engineering erlbaum error eurospeech existence experiments explicit fachbereich fallside features felix field fixed flow forcada forget frasconi fred free from fully garrison genevieve germany gers gradient guide hillsdale hinton hochreit hochreiter identification ieee ilchmann infeng inference informatik information initial input institut intl jectories jurgen kaiserslautern kaufmann klaus kolen kremer language languages learn learning lecture lecun lehrstuhl letters line long lstm mateo mathematics mathematik mcclements mellon memory mikel modelling morgan muenchen muller munchen nets network networks netzen neural neuroe neuronalen noisy notes nowlan order oscillations oscillator output owens pages pearlmutter peng phase pittsburgh plaut praetzelwolters precise prediction press proc proceedings processing prof propagation prosody rafael recurrent references regular report robert robinson ruiz rumelhart running schmidhuber science second sensitive sequences sheng short siegelmann simard simple size sontag space springer state storage structure survey systems technical technische temporal term their theory thesis time timing townley trade training transactions tricks tsung turing universitaet universitat university untersuchungen using utility verlag volume weiss williams with without works yoshizawa zipser http://www.jmlr.org/papers/volume4/blanchard03a/blanchard03a.pdf 260 A New Approximate Maximal Margin Classification Algorithm abound adaptive adatron additive advances aizerman algorithm algorithms allwein alto analysis angluin annals annu anthony applied applying approach approximate artificial association attributes auer automata automation average backpropagation bakiri bartlett based basis bengio bhattacharyya bianchi binary blake block books boosting boser bottou bounds brain braverman brunot burges california cambridge campbell categorization cesa classification classifier classifiers code codes comparing comparison comput computation computational computer computing concept conf conference confident context control convergence correcting correction cortes cowan cristianini dagan dags data databases denker dependent dept design dietterich digit discovery discriminant distance driven drucker duda editors effectiveness efficient empirical error explanation exponentiated fast feature forthcoming foundations freund friess from function functioning gaussian general generalized gentile giles girosi golding gradient grove guyon handwritten hanson hart helmbold henderson hierarchies hinge howard html http hubbard icann ieee improved indian information input institute intel international introduction irrelevant irvine iscas iterative jackel jersey joachims jority journal karov kaufmann keerthi keogh kernel kernels kivinen knirsch know kowalczyk language large learning lecun ledge leen ligence like line linear linguistics littlestone logarithmic long loss machine machines making mangasarian margin mateo mathematical maximal maximum mechanisms merz method methods mika minimal minimization mining mistake mlearn mlrepository model modern morgan muller multi multiclass multidimensional murthy nachbar national natural nearest network networks neural neurodynamics niyogi nnsp norm nossek novikov online optimization osuna other output over pages palo pattern perceptron perceptrons physics platt poggio point potential practical prediction preliminary press principles problems proc procedure proceedings processing programming proofs queries quickly radial ratsch recognition reducing references regression relative relaxed relevant remote report repository research results reviews risk robustness rosenblatt roth rozonoer sackinger scale scene schapire scholkopf school schuurmans schwenk science sciences self sensitive separation sequential servedio shawe shevade signal simard simple singapore singer smola solving somerset sons space spartan spelling statistical statistics strobl structural sung support surface symposium system systems taylor technical text theoretical theory thesis threshold training trans transformation tresp unifying university unpublished updates using vapnik variables vector version versus volume voting warmuth washington weak weighted when wiley williamson winnow with york http://www.jmlr.org/papers/volume8/dasgupta07a/dasgupta07a.pdf 210 A Probabilistic Analysis of EM for Mixtures of Separated, Spherical Gaussians achlioptas advances algorithm analysis annals annual applications applied arora artificial assignment asymptotics balls belmont bishop california cambridge classification clustering computation computational computer conference convergence dasgupta data decision dempster densities diaconis distance distributions duda dudley durrett duxbury examples feller finite foundations freedman from gaussian gaussians general generalizations gonzalez graphical hard hart haussler ieee incomplete information intelligence intercluster introduction john jordan journal kannan kearns laird learning likelihood london makov mansour massachusetts mathematics maximum mcsherry method methods minimize mixture mixtures model models networks neural nonspherical other oxford pattern points press probability proceedings projection properties pursuit recognition redner references review royal rubin salmasian scene science sciences separated siam smith soft spectral statist statistical statistics subsets symposium systems theoretic theoretical theory titterington uncertainty university vazirani vempala walker wang wiley york http://www.jmlr.org/papers/volume3/marx02a/marx02a.pdf 300 Limitations of Learning Via Emb eddings in Euclidean Half Spaces academic adaptive algorithm algorithmic alon annual arriaga automata bartlett based baxter bollob books bound bounds brain cambrigde classification communication communications complexity computational computer concepts conference control convergence cornell david edition embeddings error estimating euclidean explicit extremal forster foundations frankl freund generalization geometrical graph gyorgy half hans improved information janos jection john jonathan journal jurgen kernel large learning limitations line linear llew lower maass machine margin margins marvin mason math mathematical mechanisms methods minsky model models neurodynamics niels novikoff optimal optimization organization pages papert paturi perceptron perceptrons peter plexity press principles priori probabilistic problem proceedings processing proofs psych ramamohan random realization references report results robert robust rodl rosa rosenblatt santosh save schapire schmitt science sciences separation series seymour shai signal simon sons spaces sparsity spartan springer statistical storage symposium system systems technical theory third through turan ulrich unbounded university using vapnik vempala verlag vladimir washington wiley wolfgang workshop yoav zarankiewicz http://www.jmlr.org/papers/volume7/vert06a/vert06a.pdf 132 Consistency and Convergence Rates of One-Class SVMs and Related Algorithms advanced alamos algorithm amer analysis annual anomaly applications approximation aronszajn arsenin assoc asymptotic bartlett based bayes behavior books boosting boser bounds bousquet brooks classification classifiers cole combinatorial compare complexities complexity comput computational computer concentration conditional consistency consistent constructive contour control convex convexity covariance density description detection devore devroye dimensional dimensions discrimination distribution estimating estimation explicit fast folland fourier framework function functions gaussian grove grundlehren guyon gyorfi hartigan high hilbert hush ieee induced inequalities inform information jordan june kernel kernels koltchinskii laboratory learn learning lecture level likelihood local localized lorentz loss lugosi mach machines mammen manuscript margin massart matache math mathematics mathematischen maximum mcauliffe mendelson method methods minimization national neural nonparametric notes optimal other pacific pages pattern penalized platt posed press probabilistic probabilities probability problems proceedings rademacher rates recognition references regularized report reproducing results risk risks scholkopf science sciences scovel september series sets shawe silverman smola smooth software solutions some spaces sparseness springer stat statist statistical statistics steinwart support taylor technical tewari their theory tikhonov toeplitz toulouse training trans tsybakov universally using vapnik vayatis vector verlag vert volume wadsworth williamson winston wissenschaften workshop zhang http://www.jmlr.org/papers/volume7/schmitt06a/schmitt06a.pdf 106 On the Complexity of Learning Lexicographic Strategies academic accumulation accuracy advances agnostic algorithm algorithms amaldi among anja annual anthony appear applications approximability approximately approximation arndt artificial assessing atri ausiello bamberg barcelona bartlett based behavior behavioral bellare berlin bernhard best binary bottou bounded brain brake broder brunswikian bulletin bullock cambridge case characterization checkable chervonenkis choice classification cognition cognitive cohen combinatorial comparison compensatory completeness complexity computational computer computers computing conditions conference confidence constant continuous convergence convex cover covering crescenzi cues cummins daniel david dawes decision degree design detje dieckmann different directions does doherty dorit dorner ecological editor editors effects efficient empirical environments eric error even events everyone evidence experimental fabra factors fast feasible feature fifth finding format foundations francisco freeman frequencies from frugal gambosi garey general gerd gigerenzer goldstein goldwasser greiner group guide hans heinz herbert heuristic heuristics hochbaum hoffgen hoffrage hogarth horn human inference inferences influencing information intelligence interna intractability john johnson joseph journal kadane kann karelaia kearns kevin klaus kleinbolting laura lawrence learning leon lexicographic limited linda linear lists logic look lund machine machines made make making manual marchetti martignon martin matching maximum measure memory mental method michael minds minimizing model modeled modeling models natalia nellen network neural neurons newell nicola none nonzero october onereason optimal optimization order ordering organizational orponen other oxford pages paired paper pekka peter platt pompeu precis predictors preprint preserving press probabilistic probabilistically probabilities probability probably problem problems proceedings processes processing programming programs proofs properties protasi psychological psychology psychonomic quality quantity rational rationality reason reasoning recent reductions references regression relations relative representation research rest results review revising right rivest robert robin robust robustness ronald rule russell safra satisficing saul schapire schaub schiffer schmitt scholkopf science sciences search sellie seth shanks shmuel siam simon simple simplicity simultaneous singer single skiena slegers smart solutions solving spaccamela springer stefani stefanie steven strategies structure structured stuart studies study subsystems symposium system systems take takes tarrant test tests that their theoretical theory they thing things timal time tional todd toward trainability traub ulrich under uniform unifying universitat universitatsverlag university unsatisfied validity vapnik variables verlag versus vertex volume wefald weighted weiss well weston when william with work working yair yehuda yoram york http://www.jmlr.org/papers/volume6/langford05a/langford05a.pdf 40 Tutorial on Practical Prediction Theory for Classification additive advances algorithms always american annals anthony applications association assumption asymptotic averaging bartlett based batch bayes bayesian beating binomial biometrika blum blumer bound bounded bounding bounds bousquet carnegie case changing chernoff chervonenkis classification classifiers clopper colt combining communications complexities compression computable computation computational computer conference confidence convergence conversions covering crossvalidation data decomposition department dependent described devroye dimension efficiency ehrenfueucht error essentially events exist exponentiated fiducial floyd fold frequencies gaussian generalization gradient graepel gyorfi haussler herbrich here hierarchies hoeffding holdout hypothesis icml ieee illustrated inequalities information international intervals joachims journal kalai kivinen langford large learnability learnable learning letters limits linear littlestone local lugosi machine machines marchand margins mathematical mcallester measure mellon mendelson microchoice minimization model neural nips note notes observations occam occur online over pages pattern pearson point prediction probabilistic probabilities probability process processing program progressive rademacher random razor recognition references relating relative report research risk sample samples scale science seeger self setting seung shawe shell springerverlag statistical statistics structural sums svmlight system systems taylor technical test tests that their theory there these train transactions under uniform unpublished updates upon valiant vapnik variables versus warmuth williamson york http://www.jmlr.org/papers/volume7/spratling06a/spratling06a.pdf 131 Learning Image Components for Object Recognition academic acoustics adapted adaptive advances ahalt algorithm algorithms american analysis anti arbib architecture association auto barlow based benefits berlin beyond biological blind brain cambridge categorisation cause cells chapter charles chen cheng codes coding cognition cognitive competition competitive component components computation computational computer computing conditions conference constraints cortex cortical current cvpr cybernetics davies dayan dendritic development dietterich discovering discovers distributed doctrine editor editors effects efficient ellis enhances entropy exploring extraction face factorization feature feedback feng field foldiak forming frey functional fyfe gauthier gazzaniga generalization generative ghahramani handbook harpur harris hebbian helmholtz herault hinton hochreiter hoyer icassp icdl icsc ieee independence independent inference information inhibition inhibitory inns inputs integration interactive international iwata jenkin johnson joint jordan journal jutten know kohonen krishnamurthy lateral learning leen letters linear local localized lococode london lucke machine macrocolumn malsburg maps matrix mechanisms meila melton methods minor mixture mixtures model modelling models multiple naiso nature neal negative network networks neural neurobiology neurocomputing neuromimetic neuron neuroscience neurosciences object objects olshausen opinion optimal organizing pages palmeri part parts pattern pearson perception perceptual philosophical plumbley poggio prager press primate principal proceedings processing psychology psychophysical pyramidal quantization rapid recognition rectification recurrent references reilly representation representations research reviews royal saund schmidhuber science seeking self sensory separation series seung shepherd shum signal significance similarity simple sinha sleep society soco soft sources sparse sparseness spatially speech spratling springer strategies structure symposium systems task that theory think through transactions trees tresp unconscious understanding university unsupervised using vector verlag versatile vision visual volume wake with workshop york young zemel zhang zheng http://www.jmlr.org/papers/volume5/cortes04a/cortes04a.pdf 11 Rational Kernels: Theory and Algorithms academic acids advanced advances alberta alex algorithm algorithms alignment allauzen analysis anders annual artificial arto aspects asru august automata automatic banff based berg berlin bernhard berstel biennial binary biological boostexter boosting boser brian budapest bunch california call cambridge canada capable categorization chichester chris christensen christian christina christophe ciaa classification classifiers cleve codes colt combinatorial combinatorics communication complex composition computational computer conference context convolution corinna correcting cortes counts cristianini cruz cyril david definite definitions deletions design dialog dietterich discrete distance distances distribution doklady dongarra durbin dynamic eatcs ecai eddy edit editors eiji eilenberg eleazar eskin eugene european eurospeech extended fernando fifth finite first formal foundations framework frameworks free general geneva germany gokhan graeme grammar guide guyon haffner harmonic haussler heidelberg holloway holt hong http huma hungary icassp icml ieee information insertions intelligence international isabelle islands izhak jack january jason jean jens jeremy jmlr john journal july kernel kernels kingston kong krogh kuich language languages lattice lawler learning lecture leen leslie levenshtein library linpack lodhi london machine machines manfred march margin massachusetts matroids matti mehryar michael mismatch mitchison models mohri moler moments monographs multiplicative nello networks neural ninth nips noble notes november nucleic number ontario optimal optimization optimizing pages path patrick paul pereira peter physics pittsburg positive power press principles probabilistic problems proceedings processing protein proteins publications rational recognition references report research ressel reus reutenauer reversals richard riley rinehart roark robert royal salomaa samuel santa schapire scholkopf science sean semigroups semiring semirings september sequence series session shafran shawe shortest siam signatures singer smola soittola sons soviet special speech spoken springer springerverlag stafford state statistical stewart string structures studienbucher stuttgart support svms switzerland system systems takimoto taylor technical technology teubner text their theoretic theoretical theory thomas todd tools training transducer transductions tresp trick twenty ucsc understanding university updates user using vancouver vapnik vector verlag virgin vladimir voice volker volume warmuth washington watkins weighted werner weston wiley william winston with workshop wright yoram york http://www.jmlr.org/papers/volume2/horn01a/horn01a.pdf 323 A Family of Additive Online Algorithms for Category Ranking advances algorithm algorithms analysis annual appear applied association automatic boosting brain buckley butterworths cambridge case categorization classification cliffs collins combining computational computer conference confidence crammer cristianini development developments discrete document duda duffy editor efficient eleventh elisseeff englewood experiments feature feedback fifteenth filtering fourteenth freund gerard hall hart helmbold images improved information international introduction isri ittner iyer journal kernel kernels labeled large learning length lewis linguistics london machine machines margin meeting method mitra model multi multiclass neural neurocomputing nevada normalization online organization over pages parsing pattern perceptron pivoted pranking predictions preferences prentice press probabilistic problems proceedings processing psychological quality ranking rated references relevance reprinted research retrieval review rijsbergen rocchio rosenblatt salton scene schapire science sciences selection shawe sigir singer singhal smart statistical storage structures study support symposium system systems tagging taylor text theory ultraconservative univ university usability using vapnik vector vegas voted warmuth weak weston wiley with http://www.jmlr.org/papers/volume7/bergkvist06a/bergkvist06a.pdf 152 Linear Programs for Hypotheses Selection in Probabilistic Inference Models advances algorithm algorithms analysis angle annals approaches artificial backbone based beger bennett bioinformatics biomolecular biopolymers bolton bradley case chalmers chemical christin cognitive combinatorial combined computer computing conference construction cornilescu correlations cost damaschke data database decision delaglio density determination determined diagnosis dihedral discriminant discrimination effects flow from functional generalized glover goteborg holland homology hypersurface identification implementations improved inseparable integer intelligence internal international jardetzky journal knapsack learning lecture leta linear linearly machine mangasarian martello master mathematical maximum medical medicine methods midwest minimum mining models multicategory multidimensional multiple network neural north notes operations optimization pages parallel partitioning patil peptides polynomial probability probing problems procedures proceedings programme programming protein references research restraints robust scatterplot scheduling schwartz science sciences search searching secondary separation sequence sets shift shifts simple sociecty software structure structures sweden szolovits tardos tertiary their theory thesis toth training tree university using volume wang wayne wiley wisconsin http://www.jmlr.org/papers/volume5/leslie04a/leslie04a.pdf 26 Fast String Kernels using Inexact Matching for Protein Sequences aaai acid akutsu alignment altschul amino analysis annual approach approximate atlas avoid based basic biocomputing bioinformatics biological biology biomedical blocks brenner cambridge chapter characteristic chemistry chothia classification classifiers cluster cohen colt combinations combining comparing computational computer computers conference convolution cortes cristianini cruz data database dayhoff detecing detect detection diekhans discovery discrete discriminative distant distantly durbin dynamic eddy eggert eisenberg elisseeff eskin evaluate extraction fast fisher foundation freund from gish gribskov haffner haussler hebrew henikoff holloway homologies homology http hubbard inexact information intelligent international introduction investigation jaakkola journal joyce kandola kernel kernels knowledge krogh kuang learning lecture leslie liao lipman local lodhi machine machines matching matrices mclachlan method methods miller mining mismatch mitchison models mohri molecular motif motifs murzin myers national neorocolttechnicalreport neural neurocolt noble notes operating optimizing over oxford pacific pages pairwise pattern phylogenetic pitfalls pnas prediction press proceedings processing profile protein proteins rational receiver recommended references related relationships remote repeated report research robinson royal sagot saigo salzberg santa saunders schwartz science scop search semi sequence sequences seventh shawe siddiqi silver similarity singer sixth smola snir sparse spectrum spelling spring string structural structure structures substitution suffix supervised support symposium systems target taylor technical text tool tree unified university using vector vert vishwanathan volume wang waterman watkins weston workshop zhou http://www.jmlr.org/papers/volume8/teboulle07a/teboulle07a.pdf 206 A Unified Continuous Optimization Framework for Center-Based Clustering Methods abstract acad academic action advances algorithm algorithms allerton also alternatives alto american analysis annealing annual application applications applied approach approximation association asymptotic athena auslender axiomatic bagirov banerjee barrier based basis bell belmont berger berkeley berkhin berlin bertsekas better bezdek bialek biometrics bottleneck bregman brimberg bull buzo cambridge censor chap chen cikm class classification classifications cliffs cluster clustering clusterings common communication communications comp company compound compression computation computers computing cones conference conic conjunction constrained control convergence convex corps cover csiszar data dayal design deterministic dhillon difference dimensional dirichlet distances distortion distributed distribution distributions divergence divergences division documents donn duality duda edition efficiency elements eleventh elkan englewood entropic entropy euclidean expected exponential extremal family feature fifth find finding flynn forgy formerly framework francisco freeman function functions fuzzy garey general generalized ghosh global gordon gradient gray grouping guan guide gurewitz hall hamerly hansen hardy harmonic hart held henderson heuristic high hill hungarica icml ieee indirect inequalities information interior international interpretability interscience interval intractability inves iterative jain jersey john johnson journal kannan karayiannis kluwer knowledge kogan laboratories labs lagrange learning least lecture lent lequel letters like linde littlewood lloyd location love machine macqueen management mappings massachusetts masson materiels math mathematica mathematical mathematics means measures merugu method methodes methods minimization minimizing minimum mining minisum mixture mladenovic models modha multidimensional multinomial multiplier multivariate murray murty nakano networks neuhoff neural nicholas nondifferentiable nonlinear nonsmooth notes npcompleteness numerical numeriques objective observations operations operativa optimisation optimization oxford pages palo paper parallel paris parties pattern penalty pereira phillips phys plasticity plenum point points polon polya pour prentice press princeton probability problem problems proc procedure proceedings programming proximal quantization quantizer rate recent recognition references regression related relaxation report research review rockafellar rose rousseau rubinov salmasian science sciences scientific second sets siam smoothing soft solution some somme sons soukhoroukova spangler spectral springer squares stat statistical steinhaus stochastic stork studia supervised surveys symposium systems teboulle technical technique telephone that theory thomas tigacion tishby tohoku trabajos transaction transactions tsitsiklis type ueda ugon unified university unsupervised using ussr utility variational vector vempala verlag volume weighting weiszfeld wiley with workshop yang yearwood york zenios zhang http://www.jmlr.org/papers/volume6/wolf05a/wolf05a.pdf 92 Feature Selection for Unsupervised and Supervised Inference: The Emergence of Sparsity in a Weight-Based Approach academic aguiar alavi algorithm allen allerton almuallim american amsterdam analysis angelo angles annual applications approach artificial association aster average baltimore barnhill based bernards bialek biegel bioinformatics biology black bloomfield blum boolean boosting bottleneck bounded bousquet boyer bradley brand breast califano caligiuri cambridge canadian cancer case categorization cell central certain change chapelle characterizations choosing chung class classification clinical clique clustering code coller combinatorica combinatorics communication compaq complexity computation computations computer computing concave concepts conf conference continuous control curran cuts cviu cvmb cvpr data descent detection dietterich diffuse dimensional discovery distinct distribution diversity downing eccv editor editors eigenvalues eigenvectors elissef embedding embryonal emergence european examples exponentiated expression extraction factorization feature features field freeman friedman friend furedi gaasenbeek gene genes genome gibbons golub goumnerova gradient graph graphs grouping guyon hall hart hearn herrmann hoeffding hopkins http huang huard iccv icml identifying ieee image images implications inequalities information intelligence interpretation introduction irrelevant isbn issn issue john johns joint jones jordan journal kerkhoven kernel kernels kira kivinen kohavi komlos kooy koval kutok lander langley language laplacian large last learning least lecture lewis linear linsker linsley lister loan long louis lymphoma machine machines malik management mangasarian many marginal marton matas math matrices matrix maxima maximum mclaughlin measures medicine mehta mesirov method methods microarray minimization mohar molecular monitoring motzkin mukherjee multiple nato natural nature nearest negative neighbor nervous network neuberg ninth nips normalized norton notes object objects olshausen olson oper optimality organization outcome pages pajdla pami parameters pardalos parts pattern patterns pavan pelillo perceptual pereira perona personal peterse pinkus placement poggio pomeroy pontil post practical prediction predictors predicts presence press principal probability problem proceedings profiling proof properties quadratic quantitative ramana ramaswamy random real receptive recognition recomb references regularized reich relevant rendell report research rifkin roberts robust roots ross sarkar scholkopf schreiber science sciences segmentation selection self sequence series sets seung shashua shaving shipp side similar simple slonim smola society sparse special spectral spectrum speech springer squares statistical statistics stolovitzky straus sturla subset sums supervised support svms symmetric system systems tamayo technical text theorem theoretic theory time tishby tumors turan unifying university using vapnik variable variables vasconcelos vector veer vega verlag versus view vijver viola vision visual volume warmuth weiss weng weston wetmore wigner wiley with witteveen wolf workshop wrappers yakhini york zagzag http://www.jmlr.org/papers/volume5/dash04a/dash04a.pdf 17 Model Averaging for Prediction with Discrete Bayesian Networks aaai about academic accounting adrian advances american analysis andrew annual application approach artificial association augmented averaging bayes bayesian becker being belief bias bishop blake bonissone buntine cached california cambridge catalonia causal causation censored cerquides chickering citeseer clark classification classifier classifiers combination company comparison computations conference cooper curse daphne dash data databases datasets david decision denver detection development dietterich dimensionality discovery discrete discriminative dissertation domingos druzdzel duda editors edward efficient egan elsevier engine environment equivalence exact expert feature florida francisco frey friedman from geiger generative genie george ghahramani glymour goldszmidt graphical gregory hart heckerman henrion herskovits html http icml improving induction inference information intelligence intelligent international jaakkola january jerome john jordan journal judea july kanal kaufmann knowledge kohavi koller langley large lauritzen learning lemmer local logistic loss machine madigan march marek marina mary mateo meila merz method methodological michael mining mlearn mlrepository model modeling models moises moore morgan naive national network networks neural nineteenth ninth ntaras occam optimality orlando overfitting pages pattern pazzani pearl pedro peter plausible prediction press probabilistic probabilities problem proceedings processing publishers publishing raftery reasoning references refinement regression repository research review richard royal sage scene scheines science search selection selective series seventeenth seventh signal simple sixteenth smile society sons soon spiegelhalter spirtes springer statistical statistics steffen stephanie structural structure structures subset sufficient survival synthesis systems technical tenth their theoretic theory thesis tommi tractable tree twentieth uncertainty under university using variance verlag verma volinsky washington west wiley window with workshop wrappers york zero http://www.jmlr.org/papers/volume3/zelenko03a/zelenko03a.pdf 291 R-max ­ A General Polynomial Time Algorithm for Near-Optimal Reinforcement Learning agent algorithm algorithms annals apllications approximating architectures artificial atkenson aumann average banos barto based bayesian brafman building center certain classes colell conf conference confirming control convera correlated csaba curious data decision dynamic econometrica efficient embedded equilibrium factored framework fudenberg game games gence general generalized hart hebrew hoffman ijcai incomplete information integrated intel international intl introduction jerusalem joint journal kaelbling karp kaufmann kearns koller leading learning less levine ligence littman machine making management markov maschler mathematical mdps megiddo model monderer moore morgan multi near networks neural nonterminating optimal pages planning played players polynomial press prioratized proc procedure programming pseudo rationality reacting real references reinforcement repeated report research reward schmidhuber science self singh statistics stochastic survey sutton sweeping systems szepesv tadepalli technical tennenholtz theoretical theory time university wellman with http://www.jmlr.org/papers/volume4/cussens03a/cussens03a.pdf 262 On the Algorithmic Implementation of Multiclass Kernel-based Vector Machines advances algorithm algorithms allwein annual application applications approach april artificial august bagging bakiri based basis bengio bennet bernhard binary boosting boser bredensteiner bregman breiman brooks burges cambridge censor charles class classification classifier classifiers codes collobert common comparison computational computer conference confidence constructing control convergence convex corinna cortes crammer cristianini criteria dags data decision decomposition decoste dennis depratment dept design dietterich discovery division editor editors elisseeff engineering ensembles errorcorrecting european experimental fast fifth finding first fourteenth freund friedman function gaussian generalization generalized ghulum gilbert girosi guermeur guyon hans hoffgen ijcnn improved information institute intel international introduction invariant isabelle january jerome joachims john journal july justification kaufmann keerthi kernel kernels klaus know koby large learnability learning ledge ligence line machine machines making margin massachusetts mathematical mathematics mechanical memo method methods minimal mining moisy morgan multi multicategory multiclass national nello networks neural neurons niyogi olshen online optimal optimization optimizations output oxford pages paral paring pattern paugam pennsylvania physics pittsburgh piuri platt poggio point practical predictions press problems proceedings processing production programming programs quinlan radial randomization rated recognition reducing references regression relaxation report research result richard robert robust ronan ross samy scale schapire scholkopf science sciences september sequential sets seventeenth seventh shawe simon singapore singer single smola solution solving statistical stavros stone stopping sung support svmtorch symposium system systems taiwan taylor technical techology theoretic theoretical theory thesis thirteenth thomas thorsten three trainability training trees tutorial ultraconservative uniform unifying university using ussr vapnik vector vladimir wadsworth watkins weston wiley with workshop yair yoav yoram zenios http://www.jmlr.org/papers/volume8/garcia-pedrajas07a/garcia-pedrajas07a.pdf 204 Nonlinear Boosting Projections for Ensemble Construction aaai accuracy adaboost adaptive additice advances against aggressive ahuja aladjem algorithm algorithmic algorithms alonso alpaydin american analysis analyzers anderson annals annual application applications approaches approximate arcing artificial asia association august automatic available back backprop bagging banfield bang bankert barcelona bari bartlett based bases bauer baxter bayesian berkeley berlin best better bhadoria bias binary blake boost boostexter boosting bottou bounded bounds bowyer boyer breiman bshouty california cambridge carney categorization cbms chang chen cherkauer cixl cjlin classical classification classifier classifiers coevolution combine combined combining comparative comparing comparison components comprehensive computation computational computer computing conditions conference confidence connection conservative consistency constructing contextual convergence cooperative correlation correspondence could creation cristianini criterion crossover csie cunningham data databases david decision denison department detection diagnosis diao dietterich different dinstein discovery discrimination distributed distributions diversity domingo douglas duin dzeroski early ecml edition editor editors effectiveness efficient eibl eleventh empirical england enhancement ensemble ensembles ensembling entity erlbaum error eschrich european evaluation evolutionary experimental experiments expert explanation explicit extraction face factor feature features fern fifteenth fifth first fisher flach forest forests foundation fourteenth fourth francisco freund friedman from games garc gavinsky generalization gesture ghosh givan gorsuch graphs grove guez guilford guterman hall hastie haykin hervas hettich higuchi hillsdale hinton holmes html http human icpr identification ieee image implementation improved independent information initial integrating intelligence internal international introduction italy john jolliffe journal july june kaufman kaufmann kegelmeyer kittler kleinberg knowledge kohavi kohonen kolen kriegman krol kuncheva kunz lallich language learned learning lecture lecun lenz lerner letters level library libsvm limit linear lippmann lnai logistic machine machines madaboost majority mallick mantaras many maps margin margineantu margins mart mason mateo mathematical maximizing maximum mcclelland measures medical medicine melbourne merz meta method methods minimize mining mitra mixtures mlearn mlrepository models modern modification moody morgan mount muller multiboosting multiclass multiple multivariate munro named national natural negative network networks neural nock nonlinear notes november number oblivious online operator opitz optimization option organizing orlando orthographic ortiz pacific pages pang paradigms parallel path patrick pattern pechinizkiy pedrajas performance pfeiffer plaza poli pollack polynomially population prediction predictions predictors prentice preparing press principal probability problems proc proceedings processing programs propagation pruning publishers pure puuronen quality quinlan raedt random randomization rated recognition reduction redundant references regression regressions regularized relationship report repository representations research richard river rodr roli rosset rotation rumelhart saddle schapire school schuurmans science sciences scientific search sebban second selecting selection self sensitive sequential series seventh shawe sinai singer sixteenth sixth skurichina smith society soft software solutions sons spaces spain speaker springer springerverlag stacked stacking stanford statistical statistics stochastic stooping stopping strategies study subsets subspace subspaces supervised support sussex symposium system systems taipei taiwan tang task taylor technical technique techniques test tests text than their theory thesis third thirteenth three through tibshirani touretzky trade transactions trees tricks tsymbal tumer university upper using utsugi variance variants vector verlag versus view volume voting wagging wang washington watanabe weak webb west whitaker wiley williams with working workshop world wrappers yand york zenko zenobi zhang zhou http://www.jmlr.org/papers/volume6/tsochantaridis05a/tsochantaridis05a.pdf 80 Large Margin Methods for Structured and Interdependent Output Variables acoustics advances algorithm algorithmic algorithms align altun analysis annual applied approach artificial association based best biological boostexter boosting boundaries cambridge categorization chapelle chow class classification classifiers clickthrough cohen collins combinatorial combinatorica computational computer conditional conference consequences constraint context control convex convolution cornell crammer cutting data department dependency discovery discrete discriminative distance document duffy durbin eddy edit efficient eighteenth elisseeff ellipsoid empirical engines estimation exact experiments exponential families fields finding first forney fortieth foundations free functions generative graepel grotschel guestrin herbrich hidden hierarchical hofmann holloway hypotheses ieee implementation industrial information intelligence interdependent international joachims johnson joint journal karmarkar kelley kernel kernels klein knowledge koller krogh labeling lafferty language languages large learning likely linear linguistic linguistics london lovatz machine machines management manning margin markov mathematics maximum mccallum measures meeting method methods mining mitchison models most multi multiclass multivariate natural networks neural obermayer optimization optimizing order ordinal output over pages parsing pcfg peled perceptron pereira performance plane polynomial pranking press probabilistic procedure proceedings processing programming programs random rank ranking recognition references regression report representations reranking research ristad roth royal schapire scholkopf schrijver schuetze schwarz science search second segmenting sequence sequences seventeenth seventieth shapire signal singer sixth smola society solving sons spaces speech statistical string structured structures support system systems tagging taskar technical text theory things thirteenth time training tree tsochantaridis twentieth twenty university using vapnik vector viterbi voted watkins weston wiley with workshop yianilos york younger zimak http://www.jmlr.org/papers/volume5/grandvalet04a/grandvalet04a.pdf 13 No Unbiased Estimator of the Variance of K-Fold Cross-Validation accuracy additive algorithmic algorithms alpaydin analysis annals anthony applied approximate artificial assessment banks beating bengio binary blum bootstrap bounds breiman chapman check choice classification combined comparing computation computational conference cross crossvalidation dawid devroye dietterich editors efron encyclopedia error estimation fold fourteenth functions generalization generalized gyorfi hall hastie heuristics hold holden inference instability intelligence international interscience introduction joint journal kalai kearns kohavi kotz langford learning leave lugosi machine model models monographs nadeau neural pages pattern predictions prequential probabilistic probability proceedings progressive read real recognition references royal sanity sciences selection society springer stability stabilization statistical statistics stone study supervised test tests theoretical theory tibshirani update validation validatory valued variance volume wiley http://www.jmlr.org/papers/volume5/mannor04b/mannor04b.pdf 237 Dep endency Networks for Inference, Collab orative Filtering, and Data Visualization academcy advances aleatories algorithm algorithms american analysis angeles applied approach approaches architecture arti aspirations association bartlett bayesian bergstrom besag between biometrika bishop breese brook buntine cambridge carlo carolina cation chain chaines chapman chickering cial claredon clarendon class classication collaborative company computation computational computations computer conditional conference contingency continuous cooperative criteria data dawid dayan dellen density department deux dieterich directed distinction distributions does doubles edition educational elds empirical encouragement estimators evaluating fast fonctions fourteenth fowlkes freeny frey friedman geman generalized gibbs gilks goldszmidt good graphical green grouplens hall hasselmo heckerman higdon hinton hofmann iacovou ieee images independence inference information intel interaction introduction jensen john joint journal kadie kaufmann kernel landwehr large larsen lattice lauritzen learning leimer ligence ligent linear local logistic ltering machine machines madison marko markov mateo mcclave mccullagh mean meek mengersen methods meyer minimal models monte morgan mozer multivariate neal nearest neighbor nelder nement netnews networks neural nonlinear north olesen open optimization oxford parental paris passage pattern pearl platt plausible portland practice predictive press probabilistic probability proceedings processes processing produce properties providence publishing quarterly reasoning recognition recursive references regression relaxation resnik restoration richardson riedl royal science second selection sensitivity sequential seventh sewell shah sleep social society sociology sons spatial speci spiegelhalter state statistical statisticals statistician statistics stochastic structure suchak support supported systems tables tech theory thirteenth times toronto touretsky training transactions tresp twelfth uncertainty university updating using variables vector wake whittaker wiley wisconsin with work york http://www.jmlr.org/papers/volume8/alaiz-rodriguez07a/alaiz-rodriguez07a.pdf 207 Minimax Regret Classifier for Imprecise Class Distributions aaai accurate adams addressing adjusting alaiz algorithm algorithms alternative analysis ankara annual application applications approach approximations artificial august back barandela barnard based bayesian benefit berger berlin bilkent blake both bounded bowyer breiman burns carcinoma caruana changing chapman chawla class classification classifier classifiers comparing competitive composite computation computational computer conference consistent constrained cost costs curieses curse cybernetics data databases decaestecker decision decisions density detection deterministic diagnosis diego discovery distribution distributions drummond duda edition editors effect eighth eldar elkan emeksiz engineering environments estimates estimation example expected explicitly fast fawcett feature feder fifth figueiras fixed francisco friedman functions garc gastric gaussian giles good group guerrero guvenir hall hand hart histogram holte html http hypothesis icml ieee ikizler images imbalance imbalanced impact imprecise improves individual information intelligence intelligent international intervals into invited isbn iterative january japkowicz john journal july june kaufmann kegelmeyer kelly knowledge kubat langford latinne lawrence learning lecture leen linear liuj machine making march mathematical matwin maximizing medicine membership merhav merz method methods minimax mining minority misallocation mizil mlearn mlrepository modulation moon morgan muller multi multiclass multivariate nemirovski network networks neural niculescu notes november olshen optimal optimization ormeci oversampling pages paper parameters part pattern performance point polak populations posterior postprocessor predict predicting prentice press prior priori probabilities probability problem problems procedure proceedings processing projections proportionate provost radar random rangel recognition references regression regret report repository representation representing research robust rodr rodriguez rule saerens sanchez satellite science scores second selection sense sensitive sets seventh sided sigkdd signal simple sixth smote sons spills springer state statistical stephen stirling stone stork strategies strategy strict structure study sueiro supervised surveys synthetic systematic systems takimoto technical technique tenth testing that theory third ting trade training trans transactions transforming trees tricks tsoi turkey uncertain uncertainties under unequal universal university unknown using vantrees varying verlag vidal warmuth webb weighting when wiley with workshop york zadrozny zhou http://www.jmlr.org/papers/volume3/bekkerman03a/bekkerman03a.pdf 274 Intro duction to Sp ecial Issue on Machine Learning Approaches to Shallow Parsing aaai abney abstracts academic acquisition advances annotated annual answering answers armstrong artificial assignment association attribute automated base based bergen between bigram brants brill brunswick buchholz building california caruana cascaded case chunk chunking chunks citeseer collins complex computational conference conll corpora corpus cruz daelemans damereau dejean dependencies dordrecht editor editors eleventh emnlp engineering english esslli exceptions extraction foundations freitag frequently fung gene generalization genes grammatical greedy hall halteren hammerton html identifying information intel interaction international john johnson jorn journal june jurafsky kaufman kluwer knowledge language large learning lexical ligence linguistic linguistics machine marcinkiewicz marcus martin material medline meeting megyesi memory michael molina morgan national natural noisy norway noun osborne pages park parser parsing part pascale penn phrases prentice principle proceedings proceeedings processing products publishers question ramshaw recognition references relation research rule rules sabine sang santa santorini seattle seen sekimizu selection shallow skut some specialized speech springer srihari stationary statistical study supported syntactic syntax system tagger taggers tagging text their tjong tospeech training transformation translation trec treebank tsujii university using veenstra verbmobil verbs very wahlster walter washington winnow with wolfgang wordclass workshop zhang zhou http://www.jmlr.org/papers/volume2/genton01a/genton01a.pdf 334 An Extensive Empirical Study of Feature Selection Metrics for Text Classification aaai adrian alex algorithms american analysis andrew artificial assumptions barnhill based bayes berlin bernhard best binormal bulletin cambridge cancer categorization centroid class classification comparative comparison conference curves data david decision detectability development discovery distribution document dumais dunja ecml edwin eibe english european event examination experimental explanation feature features fifteenth fitter fitting fourth france frank frequency gene george germany grobelnik guyon hanley heckerman hong icml implementations index inductive information intelligence international isabelle james jason java joachims john journal kamal karypis kaufmann kernels knowledge kohavi learning lkopf lyon machine machines making management many marko maryland mccallum medical mehran methods mike miller mining mladenic models morgan naive national newman nigam pages pedersen pkdd platt practical press principles proceedings psychological psychology rank references relation relevant representations research results retrieval robustness sahami second selection sigir simpson sixteenth smola statistical stephen study subset support susan techniques tenth tests text thorsten tools twenty unbalanced used using vapnik vector vladimir weston what with witten words workshop wrappers written yang yiming http://www.jmlr.org/papers/volume3/rivals03a/rivals03a.pdf 284 On the Convergence of Optimistic Policy Iteration algorithms approximation asynchronous athena barto belmont bertsekas cambridge computation computational conference convergence dayan delayed dynamic eligibility england eurocolt european fischer fourth from introduction iterative jaakkola jordan learning machine massachusetts neural neuro open pages press proceedings programming questions references reinforcement replacing rewards scientific simon singh springerverlag stochastic sutton theoretical theory thesis traces tsitsiklis umass university watkins with http://www.jmlr.org/papers/volume5/christmann04a/christmann04a.pdf 10 On Robustness Properties of Convex Risk Minimization Methods for Pattern Recognition addison additive advances aggregation akerkar alamos algorithm american analysis annals appear applications applied approach approximate approximation association asymptotic asymptotically banach barbara bartlett based bauer behaviour belmont berkeley berlin between beyond bickel biometrika blanchard boosting bounds bousquet brabanter breakdown brown burges california cambridge chen cheney christmann city classicial classification classifiers comparison complexities complexity computational computer conference consistency consistent convergence convex convexity cornell curve data dehli department depth develop diestel discovery discrete discussion dmkd doksum donoho dortmund download dudley edition editors elements elisseeff erich estimation estimators experiments exploratory fast festschrift fischer francisco freund friedman from functional functions gacv gaussian generalization grimmer gruyter guntzer habilitation hampel hand hastie hilbert hipp hodges hoffgen hong horn house http huber hubert ieee influence information ingo insurance integrationstheorie international introduction issues joachims johannes jordan journal kauffman kernel kernels knowledge kong lanl large learning least lehmann lindenstrauss local location logistic lugosi lukas machine machines making mannila margin massachusetts mathematical mathematics mcauliffe measures measuring median mendelson methods minimization minimum mining misclassifications mitarbeiter models morgan motor narosa necessity neural neurocomputing neurons nonlinear notion number olkopf operator optimal optimization other overlap pages papers parameter pearcy pedersen point positive practical press principles probability proceedings processing providence publications publishers publishing quality rademacher randomized rate rates reading real references regression regularization regularized report reports reproducing research response results rieder risk robust robustness role ronchetti rousseeuw santa scale schapire scholkopf sciences scovel sharp simon simple single smola smyth society soft some spaces sparse sparseness springer squares stability stahel stat statistical statistics statistik steinwart strata strategy structural support suykens system systems tariffs tech technical theory thesis tibshirani trainability transactions tsybakov tukey tzafriri universally university vandewalle vapnik variables various vayatis vector vehicle view virtue wadsworth wahba weighted wesley wiley with workshop ying york yosida zeidler zhang zhou http://www.jmlr.org/papers/volume8/bartlett07a/bartlett07a.pdf 231 Sparseness vs Estimating Conditional Probabilities: Some Asymptotic Results academic academy advances analysis annals anthony asymptotically bartlett based bayes behavior boosting bounds cambridge certain classes classification classifiers consistency convergence convex covering david fiacco foundations function gabor grace hard hilbert ieee information ingo introduction jordan journal kernel large learning linear loss lugosi machine machines margin martin mcauliffe methods michael minimization national network neural nicolas noise nonlinear number other peter pollard press princeton proceedings processes processing programming rates references regularized reproducing research risk rockafellar sciences sensitivity sharp soft some space sparseness springer stability statistical statistics steinwart stochastic support systems theoretical theory tong transactions tyrrell university vayatis vector verlag wahba york zhang http://www.jmlr.org/papers/volume3/lanckriet02a/lanckriet02a.pdf 376 A Robust Minimax Approach to Classification academic advances alarc algebra algorithm analysis andersen anderson annals applications approach arcing asymptotics athena available bahadur bayesian becker belmont bennett bernardo bertsekas bertsimas bhattacharyya bottou boyd bredensteiner breiman burges cambridge chapelle chebyshev class classification classifiers computation cone cones conf convex course covariance crisp cruz detection diana diego dietterich different discriminant distributions douglas duality edition editors fast fisher francisco frenk fukunaga geometric geometry ghahramani ghaoui groot guez high homogeneous http ieee implementation inequalities information insead interior international interpretation into introduction issue jordan kadane kass kaufmann kernels kluwer lanckriet larsen learning least lebret leen letters lindley linear lobo machine marshall mathematical matlab matrices methods mika minimax minimization morgan mosek muller multivariate nature navia nemirovsky nesterov networks neural nonlinear normal notes novelty olkin optimal optimization optimizer order over oxford pages pattern performance philadelphia point polynomial popescu press probability proc processing programming publishers ratsch recognition references relevance report risk robust rodr roos scholkopf scientific second sedumi siam signal single smith smola software solla special springer squares stanford statistical statistics sturm supplement support suykens symmetric systems technical terlaky theory tierney tipping toolbox training tresp university using vandenberghe vandewalle vapnik vazquez vector vicinal weston wilson with zhang http://www.jmlr.org/papers/volume3/stoppiglia03a/stoppiglia03a.pdf 285 Data-dep endent margin-based generalization b ounds for classification akademie algorithms alon american annals annual anthony antos appear applications approximate approximation artificial association banach bartlett berlin bianchi binomial blake boucheron bounded bounding bounds cambridge cesa chervonenkis classification classifiers clicque combinatorial combinatorics combined complexities complexity computational computer concentration concepts conference convergence covering data databases david density dependent devroye differences dimensions discrete distribution distributions editor efficient empirical erlebach error estimation eurocolt european fischer focs foundations fourth free gaussian generalization geometric german graphs hard hastad haussler hierarchies hoeffding html http ieee important inequalities inequality information intelligence jansen john journal kearns kegl keogh koltchinskii learnability learning lecture ledoux linder lugosi machine margin massart mcdiarmid mendelson merz method methods minimization mlearn mlrepository model more moscow nauka network networks neural notes numbers observed over pages panchenko pattern performance polynomial press probabilistic probability proceedings rademacher random recognition references repository research results risk russian sample scale schapire schemes science sciences seidel selection sensitive sharp shawe siam simon size slud soda space springer statistical statistics structural structures sums surveys symposium systems talagrand taylor terms than theoretical theorie theory time transactions translation uniform university vapnik variables verlag weights wiley williamson with within york zeichenerkennung http://www.jmlr.org/papers/volume4/meir03a/meir03a.pdf 259 A Generalized Kernel Approach to Dissimilarity-based Classification acad academy advances algebra algorithms almost analysis appear approach arkadev artificial automatic barcelona bartlett based bennett bollmann borg bradley braverman british brussels burges cambridge canada chapman character class classes classification classifiers coefficients cohn columbia compactness complexity component computation computer computers computing conference data database deformable devijver digits discriminant discrimination discussion dissimilarity distance distances dubuisson duda duin editing editors electronic elsevier empty euclidean feature features february friess fukunaga gacv garris gdalyahu geometry goldfarb gower graepel gramming graphs greub groenen hall handprinted handwritten hart hausdorff herbrich hilbert householder ieee image information informs inseparable institute intel international introduction invariance jacobs jain ject john journal kanal kearns kernel kimia klein learning letters ligence linear linearly lkopf london machine machines mangasarian matching mathematical memoriam methods metric military modern modified muller multidimensional munich mutual national nature networks neural obermayer optimization pacl page pages pairwise pattern pekalska pierre points port press principal problems processing processings programming progress proo properties proximity psychological psychometrika publishers randomized recognition references relational report representation representations reproducing retrieval review ridder robust rosenfeld royal scaling scia science scientist sdorra sebastian selection semiparametric sets shapes shock similarity smola software solla sons spaces spain springer standards statistical statisticsal stork street supo support symposium system systems technical technology templates terms their thesis thompson transactions trick tversky unified using vancouver vapnik vector verlag vision volume wahba washington weinshall wiley williamson wilson with york young zongker http://www.jmlr.org/papers/volume3/bshouty02b/bshouty02b.pdf 313 Coupled Clustering: A Method for Detecting Structural Correspondence aaai academic advances agglomerative algorithm allerton american analogies analogy analysis annual application artificial association athens attributes authoritative automatic based basic batagelj bayesian belkin berlin between bialek biology bipartite books boolean bottleneck brodley buhmann calculations california cambridge canada chakrabarti chalmers champaign chemical church clustering clusters cognitive cohn coling college collocations columbus combinatorial commonsense communication competitive computational computer computers computing concept concepts conceptual conference constructing contextual control cooccurrence copycat corpora corpus correspondence coupled cover creative critique dagan danyluk data decompositions deerwester designs detecting detection development dhillon digital dimensions disambiguating discovery distributional distributions document dumais dyadic editor editors eighteenth elements engine engineering english environment equation estimation example expression external factorization falkenhainer fast ferligoj finland fluid fluidity forbus forthcoming fourth framework francisco french from furnas gale gaul geman gene gentner germany gibbs gorodetsky greece group harshman hartmann hebrew helsinki high hofmann hofstadter huang humanities hyperlinked icml ieee illinois images indexing information ingwersen intelligence international israel issue jerusalem john jordan journal kaufmann kearns kehler keyword kleinberg knowledge laboratory landauer language large latent learning leen leong level lexicographical lexicography linguistics ller local london machine machines making mannila mapping marcus markovitch mars martin marx maryland massachusetts matrix meaning meeting mehrotra mental method methods metropolis mind mining minneapolis minnesota mitchell model modeling modha montreal morgan natural nature negative neural nips nurmela objects ohio opitz optimization ortega pages park parts pattern patterns perception pereira physics practical practice press principles probes proceedings processing project proximity public publishers puzicha queries ranked reasoning recognition references relational relations relaxation report representation research restoration retrieval retrieving revealing review ronkainen rosenbluth schader schutze science scientific search semantic senses seung shamir sigir similar similarity slonim smadja society solla sons sources sparse special speech springer state sterkenburg stochastic stolke structural structure supercomputing supporting systems technology teller term text textual theoretical theory thesis thomas through tishby topic transactions university unsupervised urbana using wiley word words workshop xtract yakhini yarowsky york zanasi http://www.jmlr.org/papers/volume2/pekalska01a/pekalska01a.pdf 328 An Introduction to Variable and Feature Selection aberdeen adaptive advances algorithm algorithms allerton alto amaldi analysis analytic annual appear application applications applied approach approximate approximation april artificial association barnhill based bayesian bednarski bekkerman benefitting bengio bennett between bialek bioinformatics biology blum boser bottleneck brain breiman breneman brooks brownstein cambridge cancer canu caruana categorization causality chapados chapelle class classification classifier classifiers cliff clustering clusters communication comparing comparisons component computation computational computer computing conference control convergence cristianini criteria cuts damage data december denker dependencies descent detecting dhillon dietterich dimensional dimensionality discards discovery discrimination distributional divisive dreyfus dualities dubois duda duffy edition editor editors edwards effect elements elisseff embrechts empirical english entropy error estimation example examples expected exponentially expression extensions extensive extraction fast feature features fifth filtering forman francisco friedman from function furey gene generalization gibbs globerson golub gradient grafting grandvalet guyon hart hastie haussler high howard humana incremental inference information innovative input intelligence international ionizing irrelevant issue iterative jaakkola jackel jebara jmlr john jordan july june kann karp kaufmann kernel kernels kira kohavi kohodursky koller kumar lacker lampinen langley lasso layer learning lecun letters linear linguistics machine machines making mallela many margin mateo maximization maximum meeting method methods metric metrics microarray microarrays minimizing mlps model modeling models molecular monari monitoring mono morgan mukherjee multi mutual nadeau neural neurocomputing nips nonlinear norm normalized olshen optimal oussar overfitting pages palo parameterized parametric pattern pearl perceptrons pereira perkins personnaz pittsburgh pnas poggio polynomials pontil posterior practical prediction press principal probabilities proc processing radiation rakotomamonjy random ranking rates reduction references regression relations relevant rendell report response reunanen rivals sahami samples scaling schoelkopf schummer schuurmans science selection series shrinkage significance sleeman smola solla song sons space sparse springer stable stanford statistical statistics stone stoppiglia stork study sufficient supervised support svms systems technical test text that theiler theoretic theoretical theory this tibshirani tipping tishby tissue torkkola touretzky toward training trees tusher uncertainty university unsatisfied using utilities validation vapnik variable variables vector vehtari voting wadsworth weston wiley winter with withdrawing word words workshop wrappers xing yaniv york zero http://www.jmlr.org/papers/volume2/tong01a/tong01a.pdf 348 Supp ort Vector Machine Active Learning with Applications to Text Classification academic action active advances algorithm algorithms annual artificial automated automatic averaging based bayes bayesian breese burges campbell categorisation categorization catlett cauwenberghs classification classifiers claude clustering cohn committee computational conference cortes cristianini dagan data decremental dell dependences dependent development discovery distribution document dumais editor editors elaborazione eleventh empirical employing engelson estimation european experiments fabrizio feedback fifteenth filter freund further gale generalization graepel hall heckerman herbrich heterogeneous horvitz incremental inductive inference information informazione intel international istituto jean joachims journal kaufmann kernel kluwer know koller lang language large latombe learning ledge less lewis ligence machine machines making management margin mcallester mccallum methods microsoft mining mitchell model modeling more morgan motion netnews networks neural newsweeder nigam opper pages pattern planning platt poggio point pool practical prentice press probabilistic proceedings processing publishers query recognition references relevance report representations research results retrieval robot rocchio rommelse rutledge sahami salton sampling scale schohn scholkopf search sebastiani selective sequential seung seventeenth seventh shamir shawe sigir sixteenth smart smola sompolinsky springer statistical supervised support system systems taylor technical text theory time tishby tong toolkit training transductive troubleshooting tutorial twelfth uncertainty under using utility vapnik vector verlag volume wiley with http://www.jmlr.org/papers/volume1/tipping01a/tipping01a.pdf 342 Feature Extraction by Non-Parametric Mutual Information Maximization academic adaptive advances aladjem algorithms alpha amari analysis andreou annals antos application applications approach april arbor artificial august bari based battiti bayes bell bengio berkeley bhattacharyya binary bishop bollacker bonnlander bound bounds california campbell canada chapelle chernoff cichocki classification clustering collobert common communication communications component computation conf conference continuous correct cspl dasgupta data december delhi density devijver devroye diego dimensions discriminant discrimination discriminative distance divergence edition editor eigenvalue entropy equivocation error estimation experiments exploratory extraction extractors fano feature features filtering fisher fourth fukunaga function ghosh girolami gorman guorong gurewitz gyorfi hall haykin hellman hero heteroscedastic hillion hmms icassp icpr ieee ijcnn improved independent indexing india inequalities information input intelligence international introduction italy journal july june kangas kapur kernel kesavan kittler kohonen koller kumar laaksonen laboratory large learning lett linear london lower machine machines masson mathematical maximization maximum measures mechanics michel michigan minhui minimum mode model moody mukherjee muller mutual network networks neural nongaussian nonlinear nonparametric november october okada optimal optimization orlando orthonormal oxford package padmanabhan pages pami parametric parzen patrick pattern peiqi phase phys piscataway poggio pontil prentice press principe principles probability problem problems proc proceedings processing program projection quantization random rank raviv recognition recursive reduced references regression related renyi report research retrieval rome rose roux sahami saon scale scholkopf selecting selection shannon signal simon smola some speech stanford statistical statistics supervised support svms svmtorch symposium synthetic system systems tainan taiwan technical textures their theoretic theory tomita tops torkkola toward transactions transformations transforms transitions transmission uncertainty university unsupervised using vancouver vapnik variables vector visualization volume washington weigend weston wiley with yang york http://www.jmlr.org/papers/volume7/raghavan06a/raghavan06a.pdf 164 Active Learning with Feedback on Both Features and Instances abound academic accuracy acquisition active agreement algorithm algorithms allan alternative angluin anick annual applications artificial association assumption atlas attributes automated automatic available baram based baum bayes beineke belkin bibliography biometrics boosting brank buckland budgeted cabelli carballo carnegie categorical categorization catlett chakrabarti chang choice cjlin classification classifiers clustering coefficient cohen cohn commerce computation computational computer computing conference construction cool corpus croft csie data databases dejong dept design detection development diaz discovery document domeniconi ecml editors educational electronically entities european evaluation expansion experiments explanation exploration extract factor falls feature features feedback filter forty francisco frayling from generalization godbole greiner grobelnik gunopulos gupta hard harpale hastie heterogeneous hill http hughes human icdm icml ieee ijcai improving incorporating incremental independence inductive inference information intelligence interactive international into irrelevant isbn issn iterative joachims joint jones journal kaufmann kedar keller kelly kibler kluwer knowledge koch koller labeled labels ladner landis lang language learning less lewis library libsvm linear linguistics literature littlestone lizotte machine machines madani madison main management many margin mccallum mcgraw mean measurement meeting mellon microsoft milic mining mitchell mladenic modeling mooney more morgan naive netnews netwroks neural news newsweeder nist nominal notebook observer online oracle organization pages park pattern pazzani perez pittsburgh pkdd poorly porter practice principles prior proceedings processing provided psychological publishers query quickly raghavan rahim recognizers references refinement reformulation relevance relevant report research resources retrieval reuters review rochery role rose salton sampling sarawagi scales schapire schohn sciences search sebastiani selected selection semi sentimental short sigir sigkdd sikora sixteenth srihari statistical stevenson stripping study succeeds suffix supervised supervision support survey surveys symposium systems technical tenth term terminological text theory thesis threshold through tomorrow tong toolkit topic tracking transactions transductive trec uncertainty unifying university unlabeled used using vaithyanathan vector view volume voorhees weighted when whitehead wisconsin with work yaniv yesterday http://www.jmlr.org/papers/volume2/molina02a/molina02a.pdf 316 Finding the Most Interesting Patterns in a Database Quickly by Using Sequential Sampling aaai accelerating adaptive addison advances agrawal algorithm algorithms american analysis annals approximating artificial assistant assisting association asymptotic based bayesian bell berlin between bounded bounding bounds breiman catalunya chernoff class classification climbing colt complexity computable computational computer conference cross curve cybernetics data databases decomposition dept discovery dodge domingo domingos editor efficiency efficient eleventh error estimation european explora extraction fayyad finding first framework freund friedman from function functions gavelda general generalization generation ghosh greiner grove haas haussler hidden high hill hoeffding hoschka hulten hypothesis ieee imielinski inconclusive inequalities information inspection intelligence intelligent international interpreter introduction isukapalli items journal july kearns klosgen knowledge landmarks langford large learning line machine magdeburg management markov maron mathematical mathematics mcallester measure mechanics method methods minimizing mining model models moore mukhopadhyay multi multipattern multistrategy nature neural observations olshen pacific pages palo part physics piatetski pkdd politecnica practical presentation press principles probabilistic probability problems proc procedures proceedings processing quesy races random references regression relational relationship report research rigorous romig rules russell sampling scaling scheffer science sciences search select selection self sequential sets seung shapiro shell sigmod size smyth spangler speed springer statistical statistics stefan stone streams strong studies subgroups sums support swami symposion system systems technical tests text their theory tishby toivonen towards transactions treatment trees unifying university useful using uthurusamy utility validation vapnik variables vazirani vldb wald watanabe wesley wiley wolpert workshop wrobel http://www.jmlr.org/papers/volume3/forman03a/forman03a_full.pdf 380 An Extensive Empirical Study of Feature Selection Metrics for Text Classification aaai accuracy achieved additional adrian alex algorithm algorithms alpha also american analysis andrew appendix artificial assumptions attained average averaged axis barnhill based bayes bayesian beat below benchmark berlin bernhard best between binary binormal both bottom bulletin cambridge cancer categorization causal centroid class classes classification classifier code collecting color colored common comparable comparative compare comparison conference containing contains contours cora corners cross curves cutoff data dataset david decision details detectability determine development dfreq different dimensional discovery distribution document documents down dumais dunja each ecml edwin eibe empirical english entire european event examination except experimental experimenting explanation extensive feature features fifteenth figure figures fitter fitting fmeasure fold following forman fourth france frank frequency frequent from gain gene george germany gives goal graph graphical graphs grobelnik guyon half hanley heckerman high highest hong hundred icml ignoring illustrating implementations includes index indicate inductive inference information intelligence international invnorm invocations isabelle isocline isoclines james jason java joachims john journal kamal karypis kaufmann kernels knowledge kohavi learning left level like lkopf lyon machine machines macro making management many marko maryland maximum mccallum measure measures medical mehran methods metric metrics mike miller million mining mladenic model models more morgan most naive national nearly negative networks newman nigam normal note number numerator oddn odds only other others over pages paper pedersen percentage perfect performance performed pkdd plane platt plot plots positive power practical precision predictor preference prefers press principles probabilistic probably problems procedure proceedings processing projected pseudo psychological psychology rand random rank ratio recall record reference references relation relevant repeated representations research resulted results retrieval right robustness sahami same scale score scored scores second section select selected selection selects separation several shape shows side sigir simpson sixteenth skew smaller smola split splits squared statistical stephen stratified studies study subset successes superior support surface susan systematically tasks techniques tenth testing tests text that these this thorsten three ties tolerance tools topographic train training trials twenty unable unbalanced under unique unweighted used using validation value values vapnik variations varies vary vector version vladimir weight were weston what which while with within witten word words workshop would wrappers written yang yields yiming zoomed http://www.jmlr.org/papers/volume6/daume05a/daume05a.pdf 83 A Bayesian Model for Supervised Clustering with the Dirichlet Process Prior academic advanced advances aharon alberta algorithm alon altun alvaro american analysis andrew anhai annals antoniak application applications approach approaches approximately artificial association august available based basu bayesian beal bhaskara bilenko blackwell blei breaking brian carbonetto cardie carl carlo chain charles chung citation claire cluster clustering clusterings code cohen colt committee comparing computational conditional conference constraints continuum converence coreference criteria daniel daphna data database daum david density department detecting developments digits dimensional dirichlet discussion distance distribution distributions doan domain domingos duke duplicate editors edward efficient elkan equivalence eric escobar estimating estimation evaluation examples expectation ferguson finley freitas from fumio functions game ghahramani gibbs graphical halevy handbook handwritten hanna haplotype hermant hertz hidden high hillel hofmann honor hwee hyperparameter icml identity ilya improving independent inference infinite information integration intelligence interdependent international ioannis isds ishwaran jacek jacob jain james jasa jayant jcgs joachims jordan journal july kamal kamishima kisynski labeled lancelot large learning lecture linguistics logic lyle maceachern machine machines macqueen made madhavan march marina markov marthi match matching matt mccallum means measures meila meng merge methods metric michael mike mikhail milch mining minka mixture mixtures model models monge monograph monte mooney motoyoshi muller multiclass nando neal neural nigam nips noam nonparametric normal notes noun november objective office ontologies ontology output pages pantel paper papers pasula patrick pedro peter phrases pitman polya poole postal press prior priors probability problems procedure proceedings process processes processing propagation radford rand rasmussen raymond recent records reference references relation relations report resolution richman rizvi roded russell rustagi sampling scheme schemes search semi september series service sets sharan shental shpitser side similarity some sonia soon spaces split springer staab states statistical statistics steven stick structured stuart studer sugato supervised support systems technical technology theory thesis thomas thorsten tomer toronto toshihiro tsochantaridis uncertainty ungar unifying united university unlabeled using usps variational vector velag vincent weinshall wellner west william with workshop xing yasmine yong zoubin http://www.jmlr.org/papers/volume4/cardoso03a/cardoso03a.pdf 241 Lagrangian Supp ort Vector Machines academic active advances advantage algorithm algorithms applications arising athena baltimore bartlett belmont bertsekas bradley burges california cambridge cessors checkerb cherkassky classification classifiers clustered complementarity computational computations computer concave concepts conference constrained cplex cristianini data databases dataset datasets department discrimination distributed division dordrecht edition editor editors error extensions fast feature ferris fifteenth first francisco from generalized golub guide hopkins html http icml ieee ilog improved incline information interior international introduction january joachims john jplatt kaufman kaufmann kernel kleinb kluwer lagrangian large learning light linear loan machine machines madison making mangasarian manual margin maryland massive math mathematical mathworks matlab matrix metho methods microsoft minimal minimization mining mlearn mlrep morgan mpml mulier multiple munson murphy musicant natick nature networks neural nevada nips nonlinear normally numb oard oint optimization ository ounds overrelaxation pages pang philadelphia platt practical press problem proceedings processing prog programming publishers quadratic reduced reference references reports research rsvm same scale scholkopf schuurmans sciences scientific second selection sequential series session shavlik shawe shtml siam smola software solo solution solutions solving sons springer ssvm statistical successive supp support svmlight systems take taylor tech technical theory thorsten training transactions unconstrained university unix user vapnik vector village wiley wisc wisconsin york http://www.jmlr.org/papers/volume3/marchand02a/marchand02a.pdf 299 Efficient Algorithms for Universal Portfolios accoustics algorithm algorithms annals annual applegate applied behavior blum bodies calculation chemical chen combining comparison complexity compression computational computer computing concave conference convex costs cover data decision diameter distributions economic efficient equation estimation fast finance foster foundation foundations frieze from functions games helmbold ieee inequalities information integration international journal kalai kannan language learning line lovasz machine machines math mathematical metropolis mixture models multiplicative near online ordentlich physics portfolio portfolios probability problem proceedings processing random randomized references regret rosenberg rosenbluth rosenfeld sampling schapire science selection side signal simonovits singer sobolev speech state structures symp symposium teller theory third transaction transactions twenty universal updates using vempala vohra volume vovk walks warmuth watkins with without http://www.jmlr.org/papers/volume6/evgeniou05a/evgeniou05a.pdf 51 Learning Multiple Tasks with Kernel Methods academic adaptive advances aggregation algorithm allenby amer american analysis analytical ando angelo annals application applied aronszajn arora artificial association august bakker bartlett based baxter bayes bayesian berkeley bias boussios breiman brief brown california cambridge canada cancer caruana categorization cbms center chung class classics classification clustering colt combining computation computational computer conf conference conjoint connection consumer cristianini cross data datamining david definite demand demographic dept description discovery disparate division domain duin dwivedi dynamic econometric econometrics edition efficient empirical equations estimating estimation evgeniou experts exploiting feature fifth forecasting forthcoming framework friedman from functions fusion gating gehrke generalized genomic ghaoui ginter girosi golub graph greene hall harvard heisele heskes heterogeneity hierarchical icml inductive information intelligence international jacobs jebara jordan journal july kernel kernels kluwer knowledge lanckriet lander langley learn learning letters linear machine machines mangasarian marketing math mathematics matrix measure membrane mercer mesirov method methods micchelli mining mixtures model models molecular mukherjee multi multiple multivariate networks neural noble nonlinear note november object observational parallel parts pattern philadelphia poggio pontil pool pratt predicting prediction predictive prentice press primary proc proceedings processing programming protein providence publishers ramaswamy rates recognition references regression regularization regularized reich relatedness report reproducing research responses review ridge rifkin robust rossi royal sampling scholkopf schuller science seattle secondary seemingly selection selective semi september series serre shawe siam silver smola society sources spectral splines srivastava statistical statistics structures sullivan support survey svms systems tamayo task tasks taylor technical tests theoretic theoretical theory thesis thrun trans transfer university unlabeled unrelated using valued vancouver vapnik vector vetter wahba watson wiley with yeang york zacharia zellner zhang zidek http://www.jmlr.org/papers/volume7/liu06a/liu06a.pdf 144 New Algorithms for Efficient High-Dimensional Nonparametric Classification aaai accelerating access accuracy adaptive advances algorithm algorithms american analysis anchors annals anomaly applied approach approximate approximation april apte archive artificial arya ashley assn attacks barber bartlett based baxter behaviour bentley best binary birch body book bootstrap bouktache bound boundary bowyer branch bumptrees california canonical card cardie carnegie case cell chae chan chang chem ciaccia citeseer clarkson class classification classifier classifiers cluster clustering combination communications company complex complexity comput computational computer computers computing condensed conf conference constraint content contents cost cover credit cybernetics data database databases datasets december decision deerwester deng department dept design detection devroye dietterich digital dimensional dimensions discovery discriminant discrimination distances distortion distribution djouadi draper duda dumais dynamic editors effective efficient electronic equitz estimates estimation exact expected experimental expo faloutsos fast fawcett feature features fifteenth fifth filtering filters finding finkel fisher fixed flickner francisco fraud friedman fukunaga function furnas gates general geometric geometry gionis gorkani gray guttman hafner hamamoto hammersley handling harshman hart hashing hastie hauptman hierarchy high hill holland howe html http huang hypersphere ieee image imbalanced improving index indexing indyk inequality information informedia initial instance insurance intelligence intelligent international introduction investigation irvine isenhour issues jain japan john joint journal june kanal kassam kaufmann kegelmeyer knowledge koivune komarek krishnaiah kruegel labeled landauer large latent learning leen letters library lippmann livny local logarithmic logistic lowry machine machinery maneewongvatana manipulation many march matches mathematical mcgill mcgraw means measure media mellon metalearning method methods metric mining minority modeling modern moody moore morgan motwani mount multi multimedia multiple multiresolution multivariate narendra nearest neighbor neighbors neighbour netanyahu network neural niblack nist nonlinear norm north novel november oard omachi omohundro optimal outputs pages palau parametric patella patrick pattern pednault pelleg pentland petkovic photobook picard pods practical prediction preparata preprocessing press principle principles probabilistic problems proc proceedings processing prodromidis project prototypes provost proximity publishing qbic qsar quantity queries query querying ramakrishnan reasoning recognition recognizing reduced references regression report results retrieval risk ritter robust rosen rule rules salton salzberg satisfying sawhney scene science sciences scientific sclaroff search searching security segmentation selection selective semantic september sethi sets shamos shot sigact sigart sigmod signal silverman similarity smeulders smith snapp society software sons space spaces sparse spatial specific springer srihari statistical statistics steele stolfo structure supervised survey survive symbolic symposium system systems table task technical technique theory third thomas tibshirani time todd tomita touretzky trackorganized trans transactions trec tree trees tresp triangle triangular tropsha twelfth uchimura uhlmann uncertainty under university using variable verlag very video vigna vldb volker volume wads wagner weighted weights wiley william with woodruff woods workshop world yang yanker york zezula zhang zheng http://www.jmlr.org/papers/volume5/chen04b/chen04b.pdf 16 Support Vector Machine Soft Margin Classifiers: Error Analysis advances algorithm amer anal analysis annals annual anthony appear appl applications approximation aronszajn arsenin artificial barron bartlett based basis bayes behavior between book boser bounds bousquet bull burges cambridge capacity chen classification classifiers comp compact complexities complexity comput computational consistency consistent convergence convex convexity cortes covering cristianini cucker data denison density dependences dependent devroye different discovery dortrecht editor elisseeff empirical entropy error estimating estimation evgeniou fifth foundations from function functional functions gacv generalization girosi global guyon gyorfi hansen hierarchies hilbert holmes hypothesis ieee importance important improving influence information introduction john jordan journal kernel kernels kluwer knowledge learning least lecture local loss lugosi machine machines mallick manuscript margin math mathematical mcauliffe mendelson methods minimization mining models monograph more mukherjee network networks neural neurodynamics niyogi nonlinear nonparametric note notes number numbers observational operators optimal over pages pattern performance pittsburgh poggio point pontil posed preparation preprint press principles probabilistic problem problems proceedings processes rademacher radial randomized recognition reconstruction references regression regularization relationship reproducing research rifkin risk rosenblatt roussa rule sample sampling scholkopf shannon shawe siam size smale smola solutions sons spaces spartan spline springer springerverlag square stability statis statistical statistics steinwart structural support taylor than theory tikhonov training trans transactions transations universally university using vaart values vapnik vector verlag viewpoint wahba weak weights wellner wiley williamson winston with workshop york zhang zhou http://www.jmlr.org/papers/volume7/zhou06a/zhou06a.pdf 170 Streamwise Feature Selection abramovich adaptation adapting adaptive akad akaike algorithms alphainvesting amer american analysis annals applications approach artificial assoc association asymptotic bachrach bankruptcy based bayes bayesian benjamini bickel binary biometrika blum budapest building calibration challenge classification clopinet cluster coding competes computer concept conditional conference controlling criterion csail csaki data defining dept diffusion dimension dimensional dirichlet discovery distributions doksum donoho dzeroski editors empirical estimates estimating examples experts extension extraction false fast feature features fleuret foster foundations from gedatasets george getoor gilad gunn guyon hall haystacks high hochberg html http hypothesis icml ideal ieee ijcai inflation info information intelligence international introduction invention isabelle isbn issue jasa jensen jmlr john johnstone journal kiado knowledge kohavi langley larsen lavrac learning lecture likelihood link local logistic machine margin marx mathematical maximum medical methods mining model models multi multiple mutual navot neal needles networks neural nikravesh nips nipsfsc notes number pages petrov popescul possibly powerful practical predictive prentice principle priors problem problems proc projects publication radford raedt rate references regression relational relevant report research results risk rissanen royal schwartz sdmc selection sequences series shrinkage sigkdd silverman slides society sociological soton sparse sparsity spatial special springer stanford stat statist statistical statistics stine straw streaming streamwise structural submitted subset symposium technical testing theory tishby toronto trans trees ungar university unknown using variable verlag wavelet with workshop wrappers wrobel zadeh zhang zhou http://www.jmlr.org/papers/volume8/kalisch07a/kalisch07a.pdf 225 Estimating High-Dimensional Directed Acyclic Graphs with the PC-Algorithm academic acyclic algorithm aliferis allen analysis anderson annals approximating arbor artificial background bayes bayesian besnard biometrika brown buhlmann cambridge causal causality causation chickering chow classes climbing coefficient combination completeness conf conference consistency correct correlation counting cowell data dawid dependence digraph digraphs dimensional directions discrete discussion distribution distributions domany edition editor editors edwards enumerating equivalence examples expertsystems explanation exponentially faithfulness feature features fikes first fisher francisco from geiger gillispie glymour goldenberg graph graphical graphs greedy hall hanks haray heckerman henrion high hill hotelling icml identification ieee inference inferred information intelligence international introduction irrelevant journal kalisch kanal kaufmann knowledge labeled large lasso lauritzen learn learning lemmer light machine many margel markov meek meinshausen metron model modelling models moore morgan multivariate neapolitan needed network networks number optimal oxford pages partial pearl pearson perlman prediction prenctice press probability proc proceedings reasoning references representation research robins robinson royal samples sandewall scheines science search second selection series shachter sixth society sparse spiegelhalter spirtes springer sprites statistical statistics strong structure structures synthesis theory third tractable training transactions transforms trees tsamardinos twenty uncertainty uniform university variable verlag verma volume wasserman wiley with york zhang zhao http://www.jmlr.org/papers/volume6/tsang05a/tsang05a.pdf 43 Core Vector Machines: Fast SVM Training on Very Large Data Sets achlioptas active adaptive adatron advances algorithm algorithmics algorithms amelia annual annulus application applications approximate approximating approximation april area arlington artificial august australia available badoiu bakir balls barbados barcelona based bayesian becker bengio bhattacharyya biological birch bischof boley boosted boosting boston bottou bousquet bradley breaking budapest buena burges cambridge campbell canada cascade cauwenberghs center chan chang chapelle cheung choosing chudova chung cjlin clarkson classification classifier classifying clear clustering clusters cohn collobert committee complexity computation computational computer computers computing conference core coresets cristianini cross csie curve cylinder data databases davis dealing december decomposition decremental description design detection diagonals diameter dietterich dimacs dimensional dimensions directsvm discovery disks distribution domain dorffner duin duraiswami editor editors efficient ellipsoids enclosing eskin estimating evaluation example experimental face fast features fifteenth fine finite first fitting francisco freeman freund friess fung garey gauss geometry ghahramani girosi gray greedy grossman guide heisele hierarchical high hong hornik http hungary ieee images improved incremental indyk information intelligence international intractability introduction island iterative jagadish january joachims johnson joint jones journal juan july june kaufmann keerthi kernel kernels knowledge koller kong kumar kwok laboratory lagrangian lake large learning lecture leen leslie less letters library libsvm linear livny machine machines madison making management mangasarian mannila massachusetts mathematics matrices matrix maurer mcsherry megiddo memo method methods minimal minimum mining mitchell mixture montreal more morgan motwani mukherjee multiple mumick murthy murty musicant nearest networks neural neurocomputing newton nielsen noble nock notes npcompleteness nystrom object optimal optimization osuna outliers pages parallel parameters pattern pavlov peled perceptron platt poggio point pontil practical preparata press principles problems procedure proceeding proceedings processing programming puerto quarterly quebec question ramakrishnan rank rapid recognition reduced references regression related representations research results rico roobaert rsvm sampling scalable scale scaling scheinberg schohn scholkopf schwaighofer science second seeger selection september sequential sets seventeenth shape shawe shevade siam sigkdd sigmod signal simple simplesvm situation sixteenth sixth smallest smola smyth software spain sparse speed springer squashing stanford statistical statistics still sung support svms sydney sylvester symposium systems taylor techniques tenth text theory thesis time tong towards training transactions transform trends tresp tsang tutorial twentieth under using vapnik vazirani vector verlag very viola virginia vishwanathan vision vista volume wang washington water welzl weston width wiley williams williamson wisconsin with workshop yang yildirim york zhang http://www.jmlr.org/papers/volume2/zhang02c/zhang02c.pdf 317 Stopping Criterion for Boosting-Based Data Reduction Techniques: from Binary to Multiclass Problems aaai accuracy adaptive advances algorithm algorithmic algorithms allwein annals annual application approach artificial austral average bakiri balance bartlett based best better bigpole binary bold boosting bound bounds breast breiman brodley cancer carla christopher classification classifier classifiers climbing codes colic computational computed computer condensed conference confidence correcting coupling cover daphne data databases dataset datasets david decision dependences depicts dietterich distance each echocardio echocardiogram effectiveness eighteenth eleventh eliminating empirical erin error estimation exemplar experiments explanation feature features figure filter finite florida font freund friedl friedman functions gates generalization geoffrey george german ghulum glass godfried graph hardness hart hastie heart hepatitis heterogeneous hill horse house identifying ieee ight improved information instance instances intelligence international iris irrelevant jerome john journal karl kaufmann kohavi koller lallich learning letters line machine marc margin mark martinez means mehran merz methods mislabeled model morgan multiclass murphy mutation national nearest neighbor neighborhood nock olshen online optimal output over pages pairwise pattern peter pfleger philip pima planar predictions preserving press problem problems proceedings prototype pruning randall random rated recognition reduced reducing reduction references regression relative repository research richard risk robert rule sahami sample sampling scatterplot schapire sciences sebban selection seventeenth sharper singer skalak society solving springer standard statistics stephane stone stopping subset system table techniques than that theoretic theory thirteen thirteenth this thomas three tibshirani tony tools toussaint toward transactions trees trevor twelfth twenty under unifying using value vapnik vehicle verlag vladimir voting wadsworth weighted weighting where which white whitehouse wilson with yoav yoram http://www.jmlr.org/papers/volume6/wong05a/wong05a.pdf 96 What's Strange About Recent Events (WSARE): An Algorithm for the Early Detection of Disease Outbreaks aaai aberration aberrations academy accelerated accurate activity adaptive advances aerosol aggarwal aharonowitz ajit alahakoon alan alberta alert algorithm algorithms allen alternative alto american amia analysis andrew anna annual anomaly anthrax application applications approach approaches april artificial association astronomical astrophysical attacks auai august autonlab autonweb available averages bakken balakrishnan banff barak bars based bayesian behavior belfus belief benjamini bernhard biomedcentral biomedical biometrika biosurv biosurveillance bioterrorism birkhauser bishop bonferroni boston bottou bovismorbificans breast brodley bromberg brossette bruno buckeridge burkom burst cabuk cached calcolo calculations california calls cambridge campbell canada cancer carbonell carla carlo carnegie caruana categorical chang change changes changing channels charles chart charu chaudhuri chiang chickering chief choi chou chris christina christopher chung citeseer classi classification clay clifton clustering clusters coded cohen combinedregtext commerciali communication communications complaints comput computer computing conference connolly construction continuous contrast control controlling cooper counter covert current damminda daniel dash data datasets dato david deaths decision default delle dennis denver dept design detecting detection detector development diagnosing dichtiar diego discovery disease disk distributions document domingos douglas draft drives early ears economiche edition editor editors efficient eighteenth elastic eleazar elkan elliptic embedded epidemics epidemiology erica error eskin evaluating evaluation event evolving excess expected explorations failure fall false farzad fast fawcett fienberg fifth firenze fleetwood forecasting forecasts forrest foster framework france francisco fraud from gali galit garrick genovese geoff geographical geometric george ghahramani ginner giovanni glaz goldenberg good green greg gregory greiner grocery guide gwilym halpern hamerly handling hanley hanna hardin hartman have havkin health hochberg hogan holdenday hooper hopkins hospital howard html http huang hubert hudson hulten hutwagner hypotheses icml ieee image imbalanced infection inference influenza informatics information inspection intelligence interesting international introduction intrusion intrusions issn issue issues istituto jeff jenkins jessica jiame jian john jones joseph joshua journal july junshui kaufman kaufmann keen keewhan kenneth knowledge kruegel kulldorff kymie laboratory lane large larry laurence laurie lavi lawrence lazy learned learners learning leon levander leviatan linda line local localized lori machine madigan making management mandl manfred march marom maron martin mary master matthew maxion medical medication medicine mellon menlo method methods michael michal miller mining minority mishra mitchell model modeling models monitoring montgomery moore more morgan mortality moser mostashari moving much multi multidimensional multimedia multiple murray national neill network networks neural newsletter nichol nina ninth noisy northeastern noticing novelty october oded ofra omni online operator optimal ormatics outbreak outbreaks over page pages palo parameter park parts pattern pazzani pearlmutter pedro peled pennsylvania pereira periodic perkins permutation perspective peter phillip phua pickle pierce pitt pittsburgh pnas pneumonia poster powerful practical prediction preparedness presentation press privacy probabilistic probabilita probability proc proceedings processing prospective provost pubblicazioni public quality rachel racing rapid rate real realtime reduction references reichart reinsertion reis related release report reporting reports resampling research resolution response retrieval retrospective review rich richard richardson rita robert roberts rods royal rule rules russell sales salvatore sample saul scan schemes schneider scholkopf school science sciences scienze search sebastian security seeman selection sequence sequences serdar serfling series sets seventeenth seventh shalev shasha shemer shields shmueli showpaper sigir sigkdd sigmod signal significant simon singh skewed small society sons soon sosin spatial special spencer sprague springer ssue states statistic statistica statistical statistics stephanie stephen stolfo streams structure study submitted sufficient sumandrew summer superiore supplement surveillance sylvia symp symposium syndromic synthesis system systems tamar technical technologies technometrics temporal teoria terran terrorist testing tests thacker theory thesis third thompson threshold through thrun time timing tracee tracking trans transactions transform treadwell tsui twentieth uncertainty united university urban using validation value varda verlag vigna vincent virginia vision wagner waites wallstrom warren warrender washington wasserman watier wavelet weatherby weiss weng wenke what when wiley william williamson with wong yael yair yang yiming yoav york yosef yunyue zalman zhang zoubin http://www.jmlr.org/papers/volume7/hush06a/hush06a.pdf 129 QP Algorithms with Guaranteed Accuracy and Run Time for Support Vector Machines accepts according accuracy advances alamos algorithm algorithmic algorithms also analysis annals annual anomaly appear applications approximate archive arranged assumes assumptions asymptotic auer australia balcazar bhattacharyya blake bound burges cambridge canbridge canonical chang chen chih chung citeseer cjlin classification classifier colt complexity component composite compositeupdate computation compute computer computes computing conference control convergence corresponding cristianini criteria csie data databases decomposition delete density dept design desomposition detection determines direction distribution division dual duality edition editors electrical else engineering eprints estimating fast faster feasible first formal framework freund from gaussian general generalized gilbert girosi gmax gmin gradient guaranteed highdimensional html http hush icdm ieee imax imin improvement improvements index inew information ingo initialize input insert international introduction iterative joachims john jose journal keerthi kernel kernelbased kernels kingdom laboratory lagrangian lanl large largest laskov learning lecture level liao library libsvm linear list machine machines making mangasarian mani mechanical meir merz mesce method methods minimal mining mlearn mlrepository monash murthy musicant national nearest network networks neural nonincreasing note notes offset optimization order osamu osuna other overrelaxation pages pair palaniswami papers parameter pascal platt point polynomial practical press procedure proceedings production provably publication publications pubs rates references regression repeat report repository research return returns role routine scale scholkopf science scovel second select selection sequential sets shawe shawetaylor shevade shtml simon simple singapore singer smola solution solutions sons sorted sortincreasing stability statistical statistics steinwart stepwise stopping study submitted successive support systems taylor technical that then theorem theory this threshold time training transactions trivial tuples type uine united university unstable until updatemlist updates uses using value vapnik vector violating volume watanabe while wiley with without working wwwc yang york http://www.jmlr.org/papers/volume6/ihler05a/ihler05a.pdf 61 Loopy Belief Propagation: Convergence and Effects of Message Errors algorithm algorithms analysis angelov application approximate approximating approximations artificial association based bayes bayesian belief berlin between boston bounding boyen cetin chan change chen clifford communication complex compression computation computer conference constrained constructing control correctness coughlan darwiche data decision deformable dept digital disorder distance distributed distributions editors energy errors estimation european expecatation factor february ferreira fields finding fisher fixed free freeman frey general generalizations generalized georgii gersho gibbs graphical graphs gray grimmett gruyter guestrin heskes hybrid ieee ihler inference information intelligence intelligent international isard jaakkola jordan journal julier kaufman kluwer koller kschischang laboratory lauritzen lerner local loeliger loops loopy markov mateo mathematical mathematics measure measures merl message method minka modeling models morgan nets networks neural nonlinear nonparametric optimization oxford pages pampas paskin pattern pearl phase physical points press probabilistic probability proceedings processes processing product propagation quantization random real reasoning recognition references relationships reparameterization report robust science sensor september shapes signal statistics stochastic studies submitted sudderth systems tatikonda technical theory tractable transactions transformations transitions tree uhlmann uncertainty uniqueness univ university using valued vector vision wainwright weiss welsh willsky with yedidia york http://www.jmlr.org/papers/volume2/chickering02a/chickering02a.pdf 355 Learning Equivalence Classes of Bayesian-Network Structures aaai able about above absence acyclic acyclicity added adding addition additional additions adjace adjacencies adjacency adjacent admit admits after again akaike algorithm algorithms allen allow allows along already also although always ambrosio among analogous ancestors andersson annals anytime append application applications applied applies apply applying approach approximating arbitrarily argue argument arguments arjan artificial assume assumption automatic averaging away background base based basis bayesian because become been before behavioural being belief belong belongs below bendou besnard between blair bonissone boston both bouckaert break breaking breese broken buntine cannot cardinality carnegie case causal causality causation cercone chain change changed characterization characterizing chickering child chooses choosing chord chordal chordality chosen chow class classes clearly clique closest cognitive combination combining common communications comp compel compelled complements complete completed completely completeness completes complexity component components computations computer computing conclude condition conditions conference configuration conjunction connect connected connecting connects consecutive consequence consequently consider considered considering consist consistent consisting consists constitute constitutes construct constructed constructing construction contain contained containing contains continue contradicting contradiction contradicts control convenient converging convert cooper corollaries corollary correct correctly correspond corresponding could covered create created criterion current cycle cyclic dags dash data databases decision decomposable decreasing deduce defined definition definitions delete deleted deletes deleteu deleting deletion demonstrate demonstrating denote department depend dependence depends derives descendant describe described description desired detailed detecting determine determining developed difference differs difficult digraph digraphs dimension direct directed directing direction directions directly directs disambiguates discovery discrete disjoint distributions does druzdzel each easily easy ecall edge edges edition editor editors efficient either element elements eleventh eliminating elled else emma emphasize empty endix ends engineering enote enough enter entirely enumerating equal equivalence equivalent erator erators essentially estimating every exactly example except exisl exist existed existence exists exit experimental explanation extension extensions extra extract extracting extracts fact favor fayyad fewest fifteenth fifth figure fikes final finally find first fisher flexibility follow followed following follows form forms found foundations four framework frequently from furthermore geiger general generality gets gillispie given glymour goldszmidt graph graphical graphs guarantee guaranteed guarantees guide hanks have heckerman hence henrion heorem herskovits higher hold holds horvitz however hybrid hypergraphs hypothesis identical identification identified identifies identify identifying ieee immediately implies imply important impossible incident include including inclusion increase increases independently individually induction inference inferred information initial initially input insert insertd inserted inserting insertion insertions insertu instead intel intermediate international into introduction invalid inverse involve involves iteratively itself iven jensen jordan journal kanal kaufmann kluwer know knowledge known kocka koller label laboratory largest laskey last later leads learning least lecture ledge lemma lemmas lemmer length lenz less lest ligence ligent linear literature local locally long longer look loss lower machine made madigan main makev mamdani mark markov match mateo matrix maximal maximum mean means meek mellon member members method methods microsoft middle minimum mining model modeling models modification more morgan most msrtr munteanu must nato necessarily necessary need neighbor neighbors neither network networks never next ninth node nodes none notation note notes noting number numbered obtained once only operation operator operators opposite order ordered ordering orderings organized orient orientation orientations oriented orienting orients orollary osition other otherwise output over pages pair paper parent parents part partially participate participated participates participating particular pass passes path paths pdag pdags pearl perform perlman peyton phase place plausible point possibility possible potentially prade precedes precisely prediction preliminary presence presented press previous previously principle probabilistic probability procedure proceed proceedings proceeds process processed processing produced proof proofs prop properties property proposition propositions prove proved proven proves provide proving prow purpose purposes reachable reason reasoning recall reduce refer references refinement regard related relation relies remain remainder remaining remains removed repeated repeatedly replacing report representation requirement requires research respectively restriction result resulting results returns reversal reversed reversible reversing rissanen root rule rules same sandewall satisfying scheines schwarz science sciences score scored scoring search second section sections segment segments select selected selecting selection selectively semi separate sequence series sets seventeenth seventh shachter share short shorter shortest show showing shown siam similar similarly simon simple simultaneously single sition sixth size skeleton smets social some sound sparse specifying spirtes springer staring state statement statistical statistics step still stochastic strict structure structures studeny subgraph subset subtracting such sufficient supp suppose supposition suzuki symmetric synthesis systems table tail takes tarjan tarsi technical terminates test tests than that them themselves then theorem theoretic theory there therefore these thesis they third this those three through throughout thus ties time together topology total transactions transform transformational transforms transition transitivity trees triangle trivial trivially true turn turns tutorial twelfth ucla uivalent unaffected uncertainty understood undirected undirecting unfortunately university unprocessed until used useful uses using uthurusamy valid validity variables various verlag verma vertices violated violating volinsky volume vstructure vstructures well were when whenever where whether which while whose will with within without work workshop would yannakakis yielding yields ylthat york zero http://www.jmlr.org/papers/volume6/ye05a/ye05a.pdf 47 Characterization of a Family of Algorithms for Generalized Discriminant Analysis on Undersampled Problems academic academy adding advances algebra algorithms american analysis annals annual anouar application applications applied approach arias aspremont association automation baltimore based baudat belhumeur benavente berry beyond bioinformatics brie buja cancer categorization cell center chellappa class classification classifier classifiers clustered collection comparison component computation computational computations computer computers conference constraints covariance criterion data database decomposition deerwester diagnosis dimension direct discriminant discrimination distribution diswcovery duchene duda dudoit duin dumais edelman edition eigenfaces eigenfeatures eighteenth elements error expected expression extraction face feature features first fisher fisherfaces foley formulation fridlyand friedman fukunaga furnas gaussian gene generalized geometry ghaoui golub graphical hand harshman hart hastie hespanha hopkins howland http ieee image indexing industrial inference information intelligence intelligent international introduction inverse janardan jeon johns jolliffe jonathan jordan journal kernel kernels kriegman krzanowski lanckriet landauer lasso latent learning leclerq letters leukemia lewis linear loan lymphoblastic machine machines martinez maryland mathematics matrices matrix mccarthy methods mining mixtures models modified molecular multiclass multiple national networks neural nist observational optimal optimization orthogonality outcome pages park pattern pediatric penalized pentland phillips plataniotis prediction preserving press principal problems proc proceedings processing profiling programming projection pseudo ramaswamy raudys recognition reduction redundant references regularization regularized report research retrieval reuters review royal sammon sample schokopf scienc science semantic semidefinite series siam signatures singular size skurichina small smith smola society sparse specific spectroscopic speed spline springer stabilizing statistical statistics stork structure studies subspace subtype support swets systems tang technical technique test text theorem theory third thomas tibshirani transactions transformation trec tumor tumors turk twenty types uddin uncorrelated undersampled university using value vapnik vector vectors venetsanopoulos very vision wahba weng wiley with yang yeang yeoh york yuen zhao http://www.jmlr.org/papers/volume7/rousu06a/rousu06a.pdf 162 Kernel-Based Learning of Hierarchical Multilabel Classification Models addison algorithm algorithms alphabets altun analysis associative athena automatic based benchmark berkeley berlin bertsekas bianchi biocomputing california cambridge cancedda categorization cesa chatalbashev chen ciaramita cikm classes classification classifications classifying clique collection computation conditional conference content cristianini data dekel department distributive document documents dumais efficient eskin european explorations exponential extensions factor families fast features february fields framework frey gapped gartner gaussier generalized gentile goutte graphical graphs guestrin hidden hierarchical hierarchically hierarchy hofmann http icml ieee improving incremental inference information intellectual interdependent international introduction jaakkola jmlr joachims jordan journal kernel kernels keshet koller kschischang lafferty large learning leslie lewis lodhi loeliger machine machines many margin markov massachusetts mccallum mceliece methods mitchell models networks neural nips noble nonlinear organization other output pacific pages papers pattern press proc proceedings processing product programming property protein random references related relevant renders reparameterization report representation research rose rosenfeld rousu rsise sahami salton saunders scientific selection semantics sequence shawe shrinkage sigir sigkdd singer smola spaces spectrum springer statistics string strings structured substring support survey syllables symposium syntax systems taskar taxonomies taylor technical text theory tironi transactions tree trees tschach tsochantaridis university users using variational vector very vishwanathan vishy vissmo wainwright watkins wesley willsky wipo with word words workshop world yang zaniboni http://www.jmlr.org/papers/volume6/sigletos05a/sigletos05a.pdf 89 Combining Information Extraction Systems Using Voting and Stacked Generalization aaai accuracy accurate active adaptive additive advanced advances agency agent agents algorithm algorithms applications approach approximate artificial austria automated automatic autonomous bagging bari based bases bayesian best better bled boosted boosting bottom breiman califf categorization chan chang chawathe ciravegna class classifier classifiers cohen columbia combination combining comparing computational computing conditional conference conferences continuous craven crescenzi csur cucerzan current daelemans darpa data davulcu defense demo demokritos department descriptions dietterich dipasquo directions disambiguation discovery distributions diversity documents domains domingos dzeroski eighteenth ekaw eleventh elkan engineering enhancing enrichment ensemble ensembles entropy error estimating experiments extended extensive extract extraction fast fields fifteenth flexible florian four fourteenth frank free freitag freund friedman from generalization halteren hammer hastie hatzopoulos hawaii heterogeneous hidden hierarchical hmms html http hurst icdm icml ieee iepad ijcai implementations improving induction inductive info informal informatics information instance integration intelligence intelligent international ipsj ireland issues italy jair japan java jensen jiis jmlr john journal karkaletsis kauchak kaufmann kernel knoblock knowledge kuncheva kushmerick labeling lafferty langley language large lavelli learned learning learningpinochio linguistics lists lnai logistic machine machines magazine management markov matching maximum mccallum measures mecca meeting merialdo message meta methods michalski minimal mining minton mitchell model models molina mooney morgan mukherjee multi multiple multistrategy muslea national natural ncsr networks neural nigam online ontological ontology optimization paliouras papakonstantinou parsing pattern pazzani pereira platt population practical predictors press probabilistic proceedings processing programs project projects quinlan rabiner ramakrishman random real recognition reduction references regression regressions relational relationship report repository research results rise roadrunner rome rosenfeld rule rules scalable schafer schapire seattle sebastiani seewald segmentation segmenting selected selecting semi semistructured sense sequence sequential services seventeenth seventh seymore shrinkage sigletos sites sixteenth sixth slattery slovenia smarr society sonderland sources speech springer spyropoulos stacked stacking stanford statistical stochastic structure structured structures success supervised support surveys symbolic system systems tables tagging tasks technical techniques tecuci tenth tests text than their thesis thirteenth thompson through tibshirani ting tools towards training tsimmis tutorial twenty ullman uncertainty understanding unifying university used using valarakos vector very view vldb voting vouros washington whitaker wide widom wien williamstown with witten wolpert word workshop world wrapper wrapping yarowsky york zavrel zenko http://www.jmlr.org/papers/volume6/boulle05a/boulle05a.pdf 79 A Bayes Optimal Approach for Partitioning the Values of Categorical Attributes aaai after algorithms altos american analysis annals answertree applied approach approximate artificial asseraf assistant association assumption attributes australian bayes bayesian berckman bernardo binary blake boull boulle bratko brebbia breiman california caps categorical cestnik chicago chimerge chou classification classifiers colonnes comparing computation computer conference connaissances continuous contrainte crois data databases decision department dicteurs dietterich dimension dirichlet discretization dougherty dover ebecken efficient elicitation espace estimating exploratory extraction factors features finding francisco friedman fulton gestion graphes grouping guide hand hermes html http huang humaines human idiot ieee implications induction information inria intelligence international investigating irvine joint journal kasif kass kaufmann kerber khiops knowledge kohavi kononenko kullback langley large lavrac learning lechevallier lignes machine massachusetts math matiques maximisation merz method methods metric mining mlearn mlrepository model morgan multi naive national neural nicoloyannis nouvelles numeric olshen optimal optimale ordre paris partition partitioning pattern preparation press problem proc proceedings programs progress publications pyle quantities quinlan raftery rakotomalala recherche references regression regroupement report repository republished review revue ritschard rnti robust sahami salzberg schwarz science sciences sigma smith sophisticated sous splits spss statistical statistics stone stupid supervised system tableau technical technique technologies tests theory thirteenth thompson tool total transactions trees twelf university unsupervised user users value values variables volume wadsworth webb wiley wilmslow wong works yang york zanasi zighed http://www.jmlr.org/papers/volume2/horn01a/rev1/horn01ar1.pdf 324 Word-Sequence Kernels advances alex algorithm algorithms alignment american analysis annual appl automatic automatique ballerini bernhard boser burges california cambridge cancedda carballo categorization cbsm chris classication classification classifiers computational computer conf conference convolution cortes cristianini cross cruz data deerwester department development devlopment discrete dumais dynamic ecml editors elisseeff european examination experiments features fifth franais furnas gaussier generalized grefenstette guyon harshman haussler herbrich hermes hill holloway http hull indexing information international introduction joachims journal kandola kernel kernels landauer language langues large latent learning lecture linguistic littman lodhi london machine machines making management many margin mass math mcgill mcgraw methods model models modern multilingual natural nature networks neural neurocolt notes number observational optimal pages perez practical press proceedings processing progress recherche references regional relevant report reps research retrieval roux royal salton santa saunders scale scholkopf science semantic shawe sheridan siam sigir smola society space spider spline springer statistical string structures strzalkowski support system systems target taylor tech technical text theory training traitement ucsc university using vapnik vector verlag wahba watkins with wong workshop yang york ziarko http://www.jmlr.org/papers/volume7/braun06a/braun06a.pdf 185 Accurate Error Bounds for the Eigenvalues of the Kernel Matrix abramowitz alexander analysis appear applications approximation asymptotic available berlin bernhard bernoulli blanchard bonn bousquet braun cambridge chapter charles christianini christopher compoment component computation convergence dauxois december discriminants diss dissimilarity dover editors eigenspectrum eigenvalue electronically empirical error evarist fisher formulas function functions generalization germany gilles gine gram graphs handbook horn http ieee inference information integral irene john johnson journal july kandola kernel klaus koltchinskii laurent learning legendre luxburg machine march math mathematical matrix methods mika mikio milton muller multivariate nello neural nonlinear november olivier online operators orthogonal pages polynomials pousse press principal printing problem processes properties published random references relation robert roger romain scholkopf sebastian shawe similarity smola some spectra spectral springerverlag statistical stegun tables taylor technische their theory thesis transactions ulrike universit universitat university vaart vector vladimir weak wellner williams with york zwald http://www.jmlr.org/papers/volume2/tax01a/tax01a.pdf 326 Matching Words and Pictures aaai access accompanying advances algorithm allocation ambiguity analysis annotated application applied approach archive archives armitage arpa assignment association athitsos augmenting barnard based belongie berkeley bernard blei blobworld bouman bregler brown browse browsing buffalo burhans buxton cambridge captions cars carson cascia celeux challenges chauveau chen chicago chopra cipolla classification clustering collateral combining computational computer computing conference conjunction content context cues cuts cvpr dalton data database databases della dempster dense department detection diebolt digital dirichlet dissertation dividing division document documentation documents drabenstott duygulu editors electronic empirical engine enser estimation european expectation exploiting extracting face faces fifteenth finding first fixed fleck florida forsyth foundations frankel freitas from frost gargi govindaraju greenspan hall hawaii hierarchical hofmann human ichi ieee image images impact incomplete indexing information inria instance institute intelligence intelligent international interpretation introduction jonker jordan journal june jurafsky kanade keister kobus label laird language large latent learned learning lexicon libraries library likelihood linear linguistics machine malik management manning march markel markkula maron martin massachusetts matching mathematics maximization maximum melamed memo mercer methodological methods mixture modal model modeling models modern monterey mori multi multimedia multiple naked name natural need neural newspaper niles noakes normalized object occurrence oren orlando ornager osuna page pages papageorgiou parallel parameter path pattern patterns pedestrian people photo photographs pictorial picture pictures pietra ponce practices prentice press problems proceedings processing progress puzicha quantizing queries query querying ratan recognition references report representing research retrieval roberto royal rubin satoh scene schneiderman schtze schutze science sclaroff search searching seattle segmentation semantics sequences series sethi seventh shin shortest sinha society sormunen sparse speech spie springer srihari statistical stochastic storage studies suny swain swedis systems takahashi taylor technical technology templates text texts textual theoretical thesis tools topic torres transactions transformation translation trends types understanding university user using vector venkataraman versions video view vision visual vocabulary volgenant wavelet webseer wide with word words workshop world http://www.jmlr.org/papers/volume5/rudin04a/rudin04a.pdf 30 The Dynamics of AdaBoost: Cyclic Behavior and Convergence of Margins abraham adaboost adam additive advanced advances algorithms amer analysis annals annual application approach april arcing artificial august ayhan bagging bartlett based baxter bennett berkeley boosting bounding bregman breiman bruno california caprile cesare chaos classification classifier classifiers collins column combined computation computational computer conference cynthia dale daubechies decision demiriz department descent distances distributions dmitry duffy dynamics edge editors effectiveness efficient empirical ensembles error estimation eurocolt european evolution explanation exponential february fifteenth fourth frean freund friedman function furlanello games generalization generation geometric gradient grove gunnar hard hastie helmbold highlighting implies information ingrid intelligence introduction jerome john jonathan journal kittler koltchinskii kristin learned learners learning lecture lectures leveraging limit line linear llew lncs logistic loss machine manfred manuscript marcus margin margins mason math maximizing maximum meir mendelson merler methods michael monthly msri muller multiple national neural ninth nonlinear notes october onoda overview pages panchenko path patterns period peter prediction princeton proceedings processing programming quinlan ratsch references regression regularized report research robert roli rosset rudin saharon schapire schuurmans science sciences shawe singer sixteenth smola smooth soft springer statistical statistics stefano study submitted system systems taylor technical theoretic theory thesis thirteenth three tibshirani trevor university unpublished using verlag view vladimir voting warmuth weak weights with workshop wyner yoav yoram yorke http://www.jmlr.org/papers/volume7/munos06b/munos06b.pdf 130 Policy Gradient in Continuous Time algorithms american analysis annals application applications applied approximate approximation arbib artificial baltimore bartlett barto based baxter bensoussan berlin bogdanov book bradford calculus cambridge carlo cart chichester computations concentration conference connectionist control csee differential discrete double dynamical editors engineering equations estimation event falb following french from function gauthier glynn gobet golub gradient grant henriksen hill hopkins horizon independence infinite information intelligence introduction inverted john johns journal kalman kelton kloeden kushner lavalle learning ledoux likelihood loan look machine malliavin mansour marbach markov martingales mathematical mathematics matrix mcallester mcgraw measure method methods modern monte munos neural nonlinear numerical ohsu optim optimal optimization overview pages parametric pendulum perturbation phenomenon planning platen policy polyak press probability proceedings processes processing providence ratio ratios references reiman reinforcement report research reward roberts school science search sensitivity series siam simple simulation singh society software solutions sons space springer springerverlag statistical stochastic sutton system systems talagrand technical theory thesen tomson topics translated tsitsiklis university using verlag villars weiss wiley williams wilson winter with yang york http://www.jmlr.org/papers/volume3/antos02a/antos02a.pdf 363 Data-dep endent margin-based generalization b ounds for classification akademie algorithms alon american annals annual anthony antos appear applications approximate approximation artificial association banach bartlett berlin bianchi binomial blake boucheron bounded bounding bounds cambridge cesa chervonenkis classification classifiers clicque combinatorial combinatorics combined complexities complexity computational computer concentration concepts conference convergence covering data databases david density dependent devroye differences dimensions discrete distribution distributions editor efficient empirical erlebach error estimation eurocolt european fischer focs foundations fourth free gaussian generalization geometric german graphs hard hastad haussler hierarchies hoeffding html http ieee important inequalities inequality information intelligence jansen john journal kearns kegl keogh koltchinskii learnability learning lecture ledoux linder lugosi machine margin massart mcdiarmid mendelson merz method methods minimization mlearn mlrepository model more moscow nauka network networks neural notes numbers observed over pages panchenko pattern performance polynomial press probabilistic probability proceedings rademacher random recognition references repository research results risk russian sample scale schapire schemes science sciences seidel selection sensitive sharp shawe siam simon size slud soda space springer statistical statistics structural structures sums surveys symposium systems talagrand taylor terms than theoretical theorie theory time transactions translation uniform university vapnik variables verlag weights wiley williamson with within york zeichenerkennung http://www.jmlr.org/papers/volume7/cesa-bianchi06a/cesa-bianchi06a.pdf 105 Incremental Algorithms for Hierarchical Classification about academy advances algorithm algorithms american analysis annals annual applications artificial association automata automatic available azoury bartlett bianchi bounded bounds brain cambridge categorization cesa chang chen ciaramita cjlin classes classification classifier classifiers classifying clustering collection competitive computation computational computer computing conconi conference consistency content control convergence corpus csie data dekel density development distributions documents dumais dynamic efficient electronically estimates estimation european evaluation exponential familiy from functions gentile granitzer graz hersh hierarchical hierarchically hierarchy hoeffding hofmann horn http identification ieee improving inequalities information intelligence interaction international into johnson journal kaufmann kernels keshet koller language large learning least library libsvm line linear lnai loss machine machines margin mathematical matrix mccallum medir methods mining mitchell mladenic model modeling models morgan multimodal multiple national nato networks neural nips novikov ohsu ohsumed omnipress online order organization page pages perceptron perceptrons phoneme poggio press probabilistic probability proceedings proofs psychological publishers random references regression regret regularized relative research researchandstandards retrieval reuters review rifkin robbins rosenblatt rosenfeld ruiz sahami sampling scholkopf science sciences second selective semantics series shrinkage siam singer smola springer squares srinivasan statistical statistics stochastic storage strong sums support symposium syntax systems taxonomies technology test text theory thesis threshold toolkit turning university using vapnik variables vector very volume vovk warmuth wiley williamson with words workshop yahoo http://www.jmlr.org/papers/volume8/gadat07a/gadat07a.pdf 222 A Stochastic Algorithm for Feature Selection in Pattern Recognition adaptative adaptive additive algorithm algorithmes algorithms amit analysis annals applications apprentissage approach approaches approximation approximations architecture arcing artificial averi azuaje bachrach back bakiri barnhill based bases bayesian bena benaim benveniste berlin beyond bias bienenstock binary bins biosystems blake blind bloomfield blum boosting bousquet breiman buche burges cachan caligiuri cambridge cancer challenge chapelle character chen choosing class classification classifiers clickthrough coarse codes cohen college coller combining communications complete computation computational computer concept conditional conference constrained continuity control convergence convex cooperative correcting courses cover cunningham dans data databases department detecting detection dietterich dilemma dimensional discovery doursat downing draper drift dror duality dublin duflo dune dupuis dynamical ecole edition efficient eighteenth eighth elements elisseeff ematiques engineering engines error etection evolutionary examples expression extracting extraction face fast feature fields filter fine first fleuret forests foundations framework france friedman from gaasenbeek gadat geman gene genomic gilad given golub gunn guyon hastie herault hettich high html http huard huge image inference information instability intelligence internation international invariant ishii iterative joachims john joint jordan journal jutten kernel kernels klinkenberg knowledge kushner lander learning lecun linear lipschitz logistic machine machines mackay mader mapping margin math mathematics memory merz mesirov method methods metivier metric microarray mining mlearn mlrepository model models molecular monitoring mukherjee multiclass multiple mutual nature navot network networks neural neuromimetic newman nips node nonlinearity norm normale objets optimization optimizing output pages parameters part pattern pletscher poggio pontil pour practical prediction press priouret probability problem problems proceedings processing propagation quantization ramanan random randomized rate recognition recursive reddy redundancy references regression regularization related relevance reliable relief render report reports repository research result retrieval rivest ruppin russel schlkopf schneider scholkopf science search second selection separation sequential sets seventeenth shape shapley siam signal simard singhi skorokhod slonim smola solution solutions solving sons sources sparse springer statistical statistics stochastic stochastics stochastiques strategies subset superieure supervised support svms symbolique system systems tamazyo task teaching technical theory thesis thomas tibshirani tipping tishby training transductive transformation trees trinity under unpublished unsupervised using value vapnik variance various vector verlag video view vision visual vocabulaire volume weighting weston whose wiley with wrapperbased xing york zero zhang http://www.jmlr.org/papers/volume7/ross06a/ross06a.pdf 187 Learning Parts-Based Representations of Data academic active adams adaptive advances algorithm algorithms allocation alspector among analysers analysis analyzers andrieu annual appearance application artificial autoencoders automatic based bayesian beal becker benavente bernhard biclustering biederman biological biology bishop blei bottou boutilier brand cambridge canada capturing caspi categories cause cbcl center cheng chical church classes classification clustering clusters cohn collaborative collections component components computation computer conditional conf conference correct cowan cvpr data database decomposi dept description detection dietterich dirichlet discovering distinctive does donoho doucet dynamic editor editors efros eighth einhauser energy entropic example experts expression extinction face factor factorial factorization features fergus filtering find flexible framework free freeman freitas frey from gesture ghahramani give google graphical gray groups heads heisele helmholtz hierar hierarchical hinton hofmann http human iccv ieee image images incomplete incremental index inference information institute intelligence intelligent international intl into introduction invariant ismb jacobs jojic jordan journal justifies kannan kaufmann kearns keypoints kluwer latent lawrence layers learning leen length local localized locus lowe machine manifolds maps marlin martinez massachusetts master mateo mathematical matrix mcmc memo michael minimum mirkin mixtures model models mohan molecular morgan muller multiple natural nature neal negative neural nowlan obermayer object objects occluded october olkopf other pages papageorgiou parameter parse parts pattern perona perspective poggio pontil positive press principal prior probabilistic probability proc proceedings processes processing psychological publishers quantization recognition references related report representation review ross roweis russell sara saul scale scene scenes science scoring search sebastian segmentation semantic seung sharing sivic solla spatially spca sprites statistics still stockholm stodden structure structures sudderth systems technical technology tesauro that theory thesis think thrun tion tipping toronto torralba touretzky transactions transformation transparency trees tresp uncertainty understanding university unsupervised using variational vector video view viewpoint vision weber weiss welling when williams wilsky winn with yair zemel zhang zisserman http://www.jmlr.org/papers/volume7/huang06a/huang06a.pdf 107 Generalized Bradley-Terry Mo dels and Multi-Class Probability Estimates advances agresti algorithm algorithms allwein amer american analysis annals annual applied approach artificial assoc association asymptotics available bakiri bartlett based becker behavior bengio berechnung bibliography binary biometrics biometrika blake block boser bottou bradley brier california cambridge categorical chang citeseer cjlin class classification classifiers cliffs codes comparison comparisons competition computational computer consistency contributions convex correcting corrigendum cortes coupling csie data database databases datasets david davidson department designs dietterich document edition editor editors englewood ergebnisse error estimates exdb experiments expressed farquhar fifth ford forecasts from generalization generalized ghahramani gradient group guyon haffner hall handwritten hastie html http huang hull hunter ieee incomplete individual infinity information ingram intel irvine issn journal kupper large learning lecun liacc library libsvm ligence likelihood linear lkopf machine machines margin mathematical mathematische maximumproblem mccullagh merz method methods michie minimization mlearn mlrepository mnist model models monthly multi multiclass national nelder network neural note november number olkin optimal output outputs oxford pages paired pairwise papers parameters pattern pendergrass permutations placket platt plattprob prentice press probabilistic probabilities probability problem problems proceedings processing rank ranking recognition reducing references regularized report repository research review risk schapire schuurmans science second simons singer skill smola software solution solving spiegelhalter stanford statist statistical statistics statlog support svmprob systems taiwan taylor technical tends terry text theory tibshirani ties training transactions triple turnier unifying university vapnik vector verification wahrscheinlichkeitsrechnung weather weng when wiley workshop yann yoram york zadrozny zeitschrift zermelo zhang http://www.jmlr.org/papers/volume1/meila00a/meila00a.pdf 335 An Extensive Empirical Study of Feature Selection Metrics for Text Classification aaai accuracy achieved additional adrian alex algorithm algorithms alpha also american analysis andrew appendix artificial assumptions attained average averaged axis barnhill based bayes bayesian beat below benchmark berlin bernhard best between binary binormal both bottom bulletin cambridge cancer categorization causal centroid class classes classification classifier code collecting color colored common comparable comparative compare comparison conference containing contains contours cora corners cross curves cutoff data dataset david decision details detectability determine development dfreq different dimensional discovery distribution document documents down dumais dunja each ecml edwin eibe empirical english entire european event examination except experimental experimenting explanation extensive feature features fifteenth figure figures fitter fitting fmeasure fold following forman fourth france frank frequency frequent from gain gene george germany gives goal graph graphical graphs grobelnik guyon half hanley heckerman high highest hong hundred icml ignoring illustrating implementations includes index indicate inductive inference information intelligence international invnorm invocations isabelle isocline isoclines james jason java joachims john journal kamal karypis kaufmann kernels knowledge kohavi learning left level like lkopf lyon machine machines macro making management many marko maryland maximum mccallum measure measures medical mehran methods metric metrics mike miller million mining mladenic model models more morgan most naive national nearly negative networks newman nigam normal note number numerator oddn odds only other others over pages paper pedersen percentage perfect performance performed pkdd plane platt plot plots positive power practical precision predictor preference prefers press principles probabilistic probably problems procedure proceedings processing projected pseudo psychological psychology rand random rank ratio recall record reference references relation relevant repeated representations research resulted results retrieval right robustness sahami same scale score scored scores second section select selected selection selects separation several shape shows side sigir simpson sixteenth skew smaller smola split splits squared statistical stephen stratified studies study subset successes superior support surface susan systematically tasks techniques tenth testing tests text that these this thorsten three ties tolerance tools topographic train training trials twenty unable unbalanced under unique unweighted used using validation value values vapnik variations varies vary vector version vladimir weight were weston what which while with within witten word words workshop would wrappers written yang yields yiming zoomed http://www.jmlr.org/papers/volume4/jang03a/jang03a.pdf 249 Supp ort Vector Machine Active Learning with Applications to Text Classification academic action active advances algorithm algorithms annual artificial automated automatic averaging based bayes bayesian breese burges campbell categorisation categorization catlett cauwenberghs classification classifiers claude clustering cohn committee computational conference cortes cristianini dagan data decremental dell dependences dependent development discovery distribution document dumais editor editors elaborazione eleventh empirical employing engelson estimation european experiments fabrizio feedback fifteenth filter freund further gale generalization graepel hall heckerman herbrich heterogeneous horvitz incremental inductive inference information informazione intel international istituto jean joachims journal kaufmann kernel kluwer know koller lang language large latombe learning ledge less lewis ligence machine machines making management margin mcallester mccallum methods microsoft mining mitchell model modeling more morgan motion netnews networks neural newsweeder nigam opper pages pattern planning platt poggio point pool practical prentice press probabilistic proceedings processing publishers query recognition references relevance report representations research results retrieval robot rocchio rommelse rutledge sahami salton sampling scale schohn scholkopf search sebastiani selective sequential seung seventeenth seventh shamir shawe sigir sixteenth smart smola sompolinsky springer statistical supervised support system systems taylor technical text theory time tishby tong toolkit training transductive troubleshooting tutorial twelfth uncertainty under using utility vapnik vector verlag volume wiley with http://www.jmlr.org/papers/volume7/micchelli06a/micchelli06a.pdf 198 Universal Kernels academic acta advances algebra algorithm also amer american amsterdam analysis annals annual appl applications applied approximation argyriou aronszajn author bach banach bartlett based basic basu bertinoro besov beurling bishop bochner bounds bulthoff burges cambridge characters chebyshev clarendon closure colloquium colt combinations company completely completeness complex complexity comput computations computer conference conic connection consistency continuously convex cristianini cucker dagm data definite density department direct duality duke edition editors efficient elements entire equations error euclidean evgeniou exponentials families feature fitzgerald forthcoming fourier friedman function functional functions general geometric ghaoui giese harmonic hastie hauser hill horvath hyperkernels icml influence information integral integrals international introduction italy jersey john jordan journal june kernel kernels kluwer lanckriet learning lecture lectures linear london machine machines macmillan malliavin mass math mathematical mathematics matrices matrix mcgraw mercer methods metric micchelli minimiza models monotone monotonic mukherjee multiple negative netherlands networks neumann neural niranjan noncompact note notes objective observational orthogonal oxford parameterized pattern perspectives philadelphia philos pinkus platt poggio polynomials pontil positive preprint preserve press princeton problem proceeding processing programming providence publications publishers publishing raklin rasmussen ratsch real recognition redheffer references regularization representation reproducing research revised rifkin rivlin rkhs royal royden rudin schafer schnorr schoenberg scholkopf science selection semi semidefinite series sets shawe siam smale smola society sonnenburg space spaces spheres splines springer statistical steidl stein steinwart stieltjes studies supplement support suykens symposium systems szego taylor that their theorem theory tibshirani tion trans type uncertainty universal university vandewalle vector verri wahba weiss wiley williamson winkler with york zeros zhou http://www.jmlr.org/papers/volume3/perkins03a/perkins03a.pdf 281 Round Robin Classification aaai acoustics advances algorithm algorithms alicante allwein alpaydin amsterdam anaheim analysis anand andr angulo another appear application applications approach approximate architectures artifi artificial assistant atlanta australia back bagged bagging bakiri based bauer belgium berthold between binary blake boost boosting boswell breiman brodley brooks bruges brunswick building burges california cambridge case catala chapter chemnitz cial clark class classification classifier classifiers clouse codes cohen combination comparative comparing comparison computation computer conditions conference confidence connectionist conquer constructing correcting correlated cortes coupling cristianini current dags danyluk data databases decision decomposition dellec department design dichotomies dietterich difference digit directions dreyfus ecml effective efficient empirical engineering ensemble ensembles error esann european evaluation ewsl experimental expert exploiting explorations fast feature feelders first fisher florida fogelman fort four francisco frank freund friedman from function furnkranz fusion generalization germany gish grove hand handwritten hastie hirsh html http hyperlink hypertext icassp icml identification identifying ieee implementation implementations improved improvements incremental induction information initial inspired intel intelligent interface international into irvine issue issues iwann java jordan journal kalousis kaufmann kearns kernel kinds kittler knerr kohavi kolen krieger lake large lauderdale layer learner learning learns leen ligence ligent line lists long lopez machine machines maclin magazine mantaras margin mateo mayoraz mcnemar mehrotra menlo merz method methodological methods mining mira mlearn mlrepository modular module mohan moreira morgan most muller multi multiclass multiple nashville national nato natural nearest neighbor netherlands network networks neural neurocomputing niblett nips noemon noisy note olshen opitz output outputs pacific pairwise park pattern percentages performance personnaz pfahringer platt plaza pollack polychotomies polychotomous popular porto portugal practical predictions predictors preferences preparation press price prieditis probabilistic problems procedure proceedings processing programs propagation proportions pruning psychometrika publishers pyle quinlan randomization ranka rated rault recent recognition reduced reducing references regression relations report repository research results review revisited rivest robin roli round rouveirol rule russell sampling sanchez schapire schmidt scholkopf science sciences selection sensitive separate series session shawe sigkdd signal simple singer single smola solla solving some souli spain speaker speakers special speech springer stanford statistical statistics stepwise stone structural studies study subsets supervised support svcr sydney symposium system systems tahoe task taylor technical techniques tesauro tests text theoharis theoretic three tibshirani tools touretzky training transactions trees unifying university using utgoff vapnik variants vector verkooijen verlag verleysen volume voting wadsworth watkins weston which widmer williamstown winning winter with witten work working workshop wyner http://www.jmlr.org/papers/volume8/neville07a/neville07a.pdf 227 Relational Dependency Networks abbeel adaptive adopters aggregating aggregation algorithm amherst analyses analysis annals applications approach approximate architecture artificial autocorrelation avoiding based basic bayesian belief berger bernstein besag bias blau building business calibrated cambridge carlo case casella categorization cause ceder chain chakrabarti chickering class classes classifers classification classifier classifiers clearwater clustering collaborative collective comets communities computer concept conditional conference congress consistency consumer control cortes craven crgtr customers data decision degree department dependency dept design detection differential dipasquo discovery discriminative disparity distributions document domain domingos downgrading duxbury dynamic efficiently electrical elkan empirical engineering engines enhanced estimates estimation estimators exponential extract families fawcett feature features field fields filtering first flach fraud freiburg freitag friedland friedman from gallagher geman genetic getoor gibbs gidas goldberg graffine graphical heckerman hill hyperlinks hypertext icml identifying ieee ijcai image immerman improves inducing inductive industry indyk inference institute intelligence intelligent interaction interest international invention iterative jaeger jensen joint jordan journal kadie kersting knowledge koller komoroske labeling lachiche lafferty language lattice lauritzen lawrence learned learning lehmann lessons likelihood likely link linkage logic logical loopy machine macskassy management marketing markov massachusetts mathematicians maximum mccallum meek methods microsoft mining mitchell model models monte multi murphy naive national neal network networked networks neville nigam obtaining operations order pages palmer pennock pereira perlich pfeffer point popescul power pregibon press prevent principles probabilistic probability proceedings programming programs propagation provost pseudolikelihood query raedt random references relational rennie report representational research richardson rounthwaite royal sanghai school science search securities segal segmenting selection sequence series seymore sheehan sigkdd sigmod simple simsek slattery society space spatial specific specification springer statistical statistician statistics stern stochastic structure study supporting symbolic symposium systems taskar technical their theory toolkit toronto trees uncertainty ungar univariate university upgrading using value variance vector verlag vision visual visualization volinsky weiss weld when white wide with workshop world york zadrozny http://www.jmlr.org/papers/volume7/bratko06a/bratko06a.pdf 200 Spam Filtering Using Statistical Data Compression Models aaai academic adaptive advances alamitos algoet algorithms along alto american analysis androutsopoulos annals annual anti applications applied approach april assis assistee association attributes augmenting automatic automatica barron based basic bayes bayesian benedetto boosting bound bratko breiman breyer brodley brown bulk buntine burges caglioti carreras categorization ceas chandrinos chapter character chen chhabra chui chung class classification classifiers cleary clustering coding commercial communications comparison complexity compression computational computer computing conference context contexts cormack cost cover croft cross data dbacl della demokritos description design discovery discrete document dynamic ecml editors efficient email emnlp empirical english entropy estimate estimation european evaluating evaluation event experiments exploiting feature filipic filter filtering filtron fisher framework frank free from gaithersburg goodman graham grunwald hackers harper heckerman hidalgo horspool hovold howard huynh ideas identification ieee information institute international introduction island joachims journal july karkaletsis keogh kernel kluwer knowledge kontkanen koutsias kwei lafferty language large learning length letters level levene linguistics linux lists lonardi loreto lossless lynam machine machines madrid mail mailing making management march markov marquez matching mccallum mcmillan memory mercer messages method methods metric meyer michelakis minimum mining mirkin modeling modelling models mountain myllymki myung naive natural ncsr nigam notes november october olkopf ordinateur overview pages painters paliouras palo pampapathi parameter paris partial pattern peng perspective physical pietra pitt pittsburgh position practical prediction press principle probability problem proc proceeding proceedings processing proof properties providence publishers ranlp ratanamahatana recent recherche references reilly report retrieval review rhode riao rigoutsos rissanen robinson rounthwaite sakkis sandwich scale schneider schuurmans science scientific sculley seattle segmentation semi sensitive shannon shortest shtarkov siam siefkes sigkdd similarity smola snowbird society sources space spam spambayes spyropoulos stacking stamatopoulos statistical stochastic stopping string strings structural structured suffix support symposium system systems tartu teahan technical techniques text theorem theory thesis tirri tjalkens towards track transactions trec tree trees tretyakov tutorial unbounded universal university unsolicited unwald upper using vector vectors versus view vitanyi wang weighting willems with witten word workshop yerazunis zipping http://www.jmlr.org/papers/volume3/blockeel02a/blockeel02a.pdf 295 The Subspace Information Criterion for Infinite Dimensional Hyp othesis Spaces academic active adaptation advances akaike albert algebraic algorithm algorithms amari american analysis annals application applications applied approximation aronsza artificial asymptotic atomic ators automatic automatica available bartlett based basis bayes bayesan bayesian berlin bernardo between beverly beyond bias bienenstock biometrika bishop bootstrap bounds bunke burges cambridge camp center central chapelle chen cherkassky choosing clarendon classification clustering cognitive cohn combination comments communications completely complexity component computation computing conference connection continuous control correct corrections correspondence craven cristianini criteria criterion cross curves data daubechies decision decomposition decompositions degree degroot delve dependencies description design determining devroye dilemma dimension dimensional discovery distribution donoho doursat droge dual dubes edinburgh editor editors efficient electronically empirical equivalence error estimates estimating estimation estimators evaluation expectation experimental experiments extension fedorov filters finite fisher four from fujita fukumizu function functions fundamentals further gammerman gaussian geman generalization generalized ghahramani gibbs girosi gpros graphical greedy hall hasselmo henkel heskes hidden hills hinton http huber hunter hypothesis icml ideal ideals identifiable identification ieee ieice image imai imaizumi implementation incremental industrial inference infinite information institute intel international interpolation intro introduction inverses issue jain japanese jective john johnstone jolliffe jordan journal kernel kernels kimeldorf kitagawa know kohonen konishi kustra laplace lasso learning lectures ledge letters ligence likelihood lindley linear london look lugosi machine machines mackay mallows manual maps mathematical mathematics mathematik matrix meeting method methods mika mining miyakoshi model modeling models modified moore mozer mulier muller multilayer multiple murata nature neal network networks neural nishii noising noisy norm number numerische numo observational ogawa oper operators optimal optimizing organizing other oxford pages papers parameter parameters pattern pennsylvania penrose perceptrons philadelphia prediction prentice presented press principal prior probabilistic procedure proceedings processes processing properties pruning pseudoinverse publication pursuit quadratic radial rasmussen ratsch recognition references regional regression regressors regularization regularizers report reproducing research restoration revow ridge rissanen robust royal sage saunders schatten scholkopf schwarz science sciences scientific selection self series shao shawe shibata shinomoto shirinkage shortest shrinkage siam signal significance smith smola smoothing society soft some sons spaces sparse spatial special spline splines springer statistical statistics stochastic subset subspace sugiura sugiyama support survey systems takeuchi tanaka taylor technical technometrics term tests their theoretic theoretical theory thesis thresholding tibshirani toronto touretzky transactions tsuda tutorial types unified units university using valencia validation validity vapnik variables variance vector verlag volume vovk wahba watanabe wavelet wavelets wiley williams williamson with yamanishi york yoshizawa http://www.jmlr.org/papers/volume6/fan05a/fan05a.pdf 93 Working Set Selection Using Second Order Information for Training Supp ort Vector Machines advances aerospace alexander algoo algorithm algorithmic algorithms annual appear application arbitrary assumptions asymptotic august australia austria automatic available bailey bartlett bellingham bernhard bhattacharyya blake borochoff boser building burges california cambridge categories chang chen chih christopher chung cjlin classification classifier classifiers cliffs clint competition complexity computation computational computer conference conv convergence cortes cover csie cvpr danil data databases datasets decomp decomposition department dept design detection dhush editors electrical engineering englewood eugene face fast faster fifth ford freund generalsmo geocities girosi guyon hall hans hsuen html http hush ieee ijcnn improvements increasing information international irvine jectable jiang joachims keerthi kernel kleinberg laboratory land lanl large laura learning liacc library libsvm linear linearconv lkopf machine machines making mani manry marco margin mecse merz mesce method methods michie minimal mlearn mlrepository modified monash murthy national network networks neural newtonian optimal optimization osuna pages palagi palaniswami papers pattern pettit platt polynomial practical prentice presentation press proceedings prokhorov recognition references remote report repository research rithms rong scale sciandrone science scovel select selection sensing sequential sets shevade simon slide smola software spie spiegelhalter statistical statlog step study support svmlight svms systems taiwan taylor technical theory thorsten time training transactions type ulrich university usgs using vapnik vector version vienna williamson without working workshop york http://www.jmlr.org/papers/volume7/climer06a/climer06a.pdf 135 Rearrangement Clustering: Pitfalls, Remedies, and Applications acad academic advanced algorithm algorithms alpert amico among analysis anders angeles appl applegate applications approach approximation arabie arora array artificial asia asymmetric atsp baldi banff base based bases biology bixby block bond boolean botstein brown budding cambridge canada carpaneto castro cell cellular cerevisiae ceri challenge chan chhugani chtsp chvatal ciliax civera classification climer cluster clustering combinatorial comparison complexity component comprehensive compressing computational computations computer computing conf conference conform control cook cuts cybernetics cycle dasigi data database databases daws decision decomposition dell densis derisi deriving design diameter diego dimacs dingledine direct display distributed document dunham edition editors effective eisen energy engineering england essex estivill evaluation exact experimental expert exploring expression fischetti flexible flow formation formulations functional futcher gatech gaul gene genes genetic genomewide genomic geometric gorla graph group grouping gutin hall hamiltonian hatfield herskowitz heuristic heuristics hoffer home html http hubert hybridization hypergraph icml identification ieee image implementation informs integer intelligence international into introductory iterative iyer john johnson jonker journal july june junger kahng karp karypis kernighan keyword king kluwer knowledge krishnan kumar kusiak large lawler learning lenstra letters local lodi machine manufacturing march marcotorchino marketing mathematical matrices mccormick mcgeoch metabolic metaheuristics methods microarray microarrays milner minimize mining minnesota mode models molecular moscato mulholland multi naddef natl navathe nonsymmetric norwell operational operations optimal optimization order ordering ozsu pacific pages paper paradigm partition partitioning path paths patterns physical pivoshenko plenum poster prentice press principles problem problems proc proceedings production program programming publishers punnen quarterly rank rearrangement recomb records reducibility references regulated reordering reorganization report representation research retail revisited rinnooy river rodrigues saccharomyces saddle salesman scale schader schleutermann schweitzer science sciences search september sequencing seriation severance shelfspace sherlock shmoys simple society software solution solve some sons spellman splitting sporulation springer steinbach stochastic strategy structural structures structuring stsp support surveys symmetric syst systems take technical technique techniques template thesis topics torres toth trans transcriptional transforming traveling travelling tspbib ucla unicamp unified university upper using valduriez variations velzquez venkatasubramanian verlag vertical very vision visitation vldb volgenant walk which white wiederhold wiley with yeast york zhang zverovich http://www.jmlr.org/papers/volume3/bengio03a/bengio03a.pdf 293 On Online Learning of Decision Lists abound aditi algorithm algorithms andrzej angluin annual anselm anthony attribute attributes available avrim biggs blum blumer boolean bounds bshouty california cambridge classes closed communications complexity computational computer concept cruz dagstuhl dana david decision dhagat differences efficient ehrenfeucht eiter electronically finitely foundation functions giessen guijarro haussler hellerstein helmbold http ibaraki ieee infintely informatics information institute intersection introduction irrelevant jection journal june kaufmann kazuhisa lavin learnable learning leslie letters line linear lisa lists littlestone littlstone logarithmic machine makino manfred many march martin mexico mistake mistakebound models mona monotone morgan nader nested nicholas norman occam online pages presence press proceedings processing queries query raghavan rank razor references related report research rivest robert rocco ronald sample santa science sciences servadio singh sloan subclass symposium system term terms theoretical theory thesis third thomas threshold toshihide trees university valiant victor vijay warmuth when with workshop http://www.jmlr.org/papers/volume6/drukh05a/drukh05a.pdf 72 Concentration Bounds for Unigram Language Models acoustics algorithmic algorithms always american analysis angluin annals annual anticipation artificial association asymptotically asymptotics balls based bayes bernoulli bigrams bins biometrika bounded bounds building cambridge check chen chicago church circuits combinatorics comparison complexity component computation computational computer concentration conference connection convergence corpus cross crossvalidation cryptanalysis curran data deleted dependence differences distribution does drukh dubhashi empirical english enhanced enigma ensemble error estimates estimating estimation estimators fast flajolet frequencies from gale good goodman hamiltonian harvard histogram hoeffding holden ieee independent inequalities intelligence international issue joint journal katz kearns kutin language large learning leave lectures like london machine mass matchings math mathematical mcallester mcdiarmid method methods missing model modeling models natural naval negative neural ninth notes number optimal orlitsky ortiz osborne pages parameters population press probabilistic probabilities probability proceedings proceesings processing random ranjan rate recognizer references reliable report research rule sample sanity santhanam schapire science sciences seventeenth signal simulation singularity sixth smoothing sparse special species speech stability statistical statistics structures study successes sums surveys system technical techniques theoretical theory thesis thirteenth transactions trials turing university upper valiant validation variables very with workshop yield zhang http://www.jmlr.org/papers/volume7/klivans06a/klivans06a.pdf 123 Toward Attribute Efficient Learning of Decision Lists and Parities abound algorithm algorithms angluin annual approach approximation artificial aspects attribute attributes auer available avrim barzdin based batch beigel blum boolean bounds calendar california cambridge cheney circuits communication complexity computation computational computer computers computing concept correction cruz decision degree dhagat disjunction doklady domain donnell efficient ehrenfeucht electronic empirical examples features finitely foundations freivald from functions gates general generalization golding goldmann halfspaces hastad haussler hellerstein helmbold hierarchy hill html http identification infinite infinitely information input integer intelligence intersections introduction irrelevant journal kautz kearns kivinen klivans krause langley lattices learning letters line linear lists littlestone logarithmic machine majority many mathematics mcgraw mistake mitchell myhill nevo online pages papers parity partial perceptron perceptrons personal polynomial polynomials power prediction presence press proceedings processing projection pubs pudlak queries quickly random rank razborov recursive references relevant report required research results rivest roth sample santa scheduling science sciences search second selection servedio siam size sloan soviet space spelling spielman subclass support switching symposium system technical theoretical theory thesis threshold thresholds time toward trans trees tsuchida ucsc uehara university valiant variables vazirani warmuth wegener weighted weights when winnow with workshop yaniv york http://www.jmlr.org/papers/volume6/opper05a/opper05a.pdf 103 Exp ectation Consistent Approximate Inference adaptive advanced advances albers algorithms alternatives analysis analytical anderson anomaly applications approach approaches approximate approximations artificial attias averages averaging bayesian becker belief berkeley berlin bernhard bethe beyond bias bishop boltzmann bootstrap boyd brain breese cambridge cancellation candela cccp cdma classification classifiers coherent cohn component computation concave condition consistent constrained constraints convergence convergent convex cornford correcting csat cycles darwiche determinant dietterich directions division dynamic editor editors energies energy engine equation evans expectation fabricius family field fluctuations framework free freeman friedman from gaussian generalized ghahramani gibbs glass graphical graphs hansen haykin heidelberg heskes ieee incremental independent inequality inference infinite information integral intelligence interference introduction inverse ising jaakkola jordan journal kappen kaufmann kearns kikuchi koller lawrence learn learning lecture leen lett linear lkopf mach machine machines mackay malzahn matrix maximization mcwhirter mean media method methods michael minimize minka modeling models morgan networks neural notes obermayer opper optimization pages palmer parisi path phys physics plefka practice press principe principle probabilistic problem procedure proceedings processes processing propagation quantum quinonero rangara range references report resampling research response retrieval roepstorff royal saad saul scatterometer scientific sebastian sejnowski self semidefinite siam signal society solla some sorensen sparsity spiegelhalter spin springer statistical structured submitted subtractive support suzuki symmetries systems technical theory thesis thouless thrun tommi tractable trans tree tresp university vandenberghe variational vector verlag vibes virasoro volume wainwright weiss welling wind winn winther with world yedidia york yuille zard zoeter zoubin http://www.jmlr.org/papers/volume7/zhang06a/zhang06a.pdf 151 Ensemble Pruning Via Semi-definite Programming aaai accurate active adaptive advances algorithm algorithms analysis annals anomalies application approach approaches approximation arcing arising artificial aspremont association athens attributes bagging banking bartlett based baxter bennett bites blake boosting boutell bowyer breiman brown burer card caruana chan chapman chawla chen citeseer classification classifier classifiers column combinations combined comparison concept conf conference constructing cooper credit crew cristianini cross currency data database databases december decision definite demiriz detection dietterich direct disagree discovery distributed dollar drifting editor editors electronic empirical engel ensemble ensembles evolutionary experimental experiments extending extremely factorization feige first fleiss forests formulation francisco fraud freund from generalization generation genetic ghaoui goemans graph greece grove hall hansen hashem html http hybrid iaai ieee image imaging improved improves information integration intel international issn issues january jiang john joint jordan jose journal kaufman kaufmann kegelmeyer kernel know krogh ksikes label lanckriet langberg large learned learning ledge leen libraries ligence ligent limit linear london machine maclin mammone manteo margin margineantu margins mason mathematical matrix maximization maximizing maximum mccallum menczer merz meta method methods mining mixture mizil mlearn mlrepository model models money monteiro morgan multi multitask national network networks neural niculescu nonlinear november october opitz optimal optimization pages partition partitioning pattern perrone philosophical popular predictors press prgramming problems proc proceedings processing prodromidis programming programs proportions pruning quinlan random randomization rank rates references relaxation repository research rose rounding royal salamon satisfiability sburer scalable scale scene schapire schuurmans sdplr selection selective semantic semi semidefinite series seventh shawe shen sigkdd skewed society software solving sons sparse speech stacked statistical statistics stolfo streaming streams street study subspace symposium system systems taylor technologies technology tesauro text three touretzky trading trained transactions trees uiowa unions using validation vedelsby volume wang website when wiley williamson with wolpert workshop yule zhang zhou http://www.jmlr.org/papers/volume7/kolter06a/kolter06a.pdf 202 Learning to Detect and Classify Malicious Executables in the Wild aaai academic advanced advances aegis against aiken alamitos albert algorithm algorithms alternative analysis annals anonymous approach arcing arnold artificial association attacking auditing august author authors authorship bacon bagging bartlett based bauer bayesian berkeley berlin biannual biologically birmingham boosting boser breiman burges california cambridge categorization chess chicago christodorescu cigital citeseer classification classifier classifiers code cohen columbia company comparative comparison computational computer computers computing conference constructing cost council data decision defenses department detect detecting detection diagnostic dietterich discovery discrimination domingos drummond duin dumais durning edition editors effective eighteenth email empirical ensemble ensembles environments eskin eugenics european evaluation executables expected experimental experiments expert explicitly extending fast fawcett feature features filter filtering fisher forensic forensics fourteenth fourth francisco frank freund frieder from gaithersburg gray guyon hand heckerman hexdump hill holte horvitz html http identifying ieee image imprecise index indexing indianapolis induction inductive information infosec inspired instance institute intelligence international jain jair jankowitz java jiang joachims john joint journal junk kaufmann kephart kernel kibler kjell knowledge kohavi kolter krintz krsul kuhns kurt laboratories lafayette large lawrence learning levitt linguists loss macdonell machine machinery machines maclin macmahon malcode malicious maloof management mannila many margin maron master maximum mcbride mcgraw measurements menlo methods metz mika miller minimal mining mitchell morgan morisett moss multiple national networks neural nishikawa october olsson opitz optimal optimality optimization page pages papers park pascal pattern patterns pazzani pederson pickett plagiarism platt popular practical press principles privacy probabilistic probabilities problems proceedings processing program programs provost publishing purdue quinlan radiologic radiology randomization recognition references relevance relevant report representation representations representing research retrieval review robust rossmann rule sahami sallis sams schapire scholkopf schultz schuurmans science security selection september sequential seventh shake sigkdd signal simple sixth smola smyth soft softeng software soman sorkin spafford speare springer standards static statistical statistics stolfo student study support swets symposium system systems taxonomic technical techniques technology tenth tesauro text theory thesis third thirteenth three tools track training transactions trees tuwien twelfth uchicago under university usenix using vapnik variants vector vigna virtual virus viruses visualization voting waikato weeber weka west white wild with witten woods workshop yang york zadok zero http://www.jmlr.org/papers/volume7/whiteson06a/whiteson06a.pdf 134 Evolutionary Function Approximation for Reinforcement Learning aaai achieves ackley acquired action actor adaptation adaptive adding advances adverserial agents ailamaki ales algorithm algorithmic algorithms allocation alspector amarel american analysis anderson annals annual application applications applied approach approximating approximation approximators arbor arita artificial association auer augmenting august automating autonomic averaging averse back backgammon baird baldwin baldwinian balter bandit barto based basis beaver bees behavior behaviors beielstein bellman berlin between bianchi biology block boers borst boyan bryant building burger busy butz cambridge carrier cascade case cellular cesa changing chapter characters chess chip choice classification classifier clear cmacs coarse coding coevolution coil college combining comparison competitive complex complexification complexity computation computational computer computing conference configuring congress congresss continuous control convex corrected correlation costa costs cowan crites critic cross culling dahlin dalvi darwinian data decision delay delayed descent description design detection development dietterich differences different direct discovery distinctions distributed dokas domingos downing dynamic dynamically editor editors effect efficient efficiently eilerston elevator elman encoding engineering entropy environment environments error ertoz european evolution evolutionary evolvable evolving examples experiences experiments explanation exploitation exploration factor fast fifth finite first fischer fitted fogarty foraging foundations france french from function functions fusion game games gammon gecco general generalization generalized generated genes genetic giraud global goldberg gomez gordon gradient graphical group gruau guide harchol hard hardware hinton holland howe hypothesis ieee improving incremental information inheritance inns instance intelligence intelligent interactions internal international introduction introductory intrustion iteration january joel joint jordan journal july june kaelbling kaufmann kephart king knowledge kohl konda kretchmar kumar kuyper lagoudakis lamarckian lamarckism lanzi large layland lazarevic learnable learning least lecture level life littman local locomotion machine machinery machines macready mahadevan mannor mansour markon master mathias mausam mcallester mccallum mcgovern mcquesten mcwherter mdps mechanisms meets meilijson messinger method methodologies methods michigan mieghem miikkulainen minds mining minnesota models mooney moore morgan moss multiarmed multimodal multiple multiprocessor multiprogramming national natural naturalist nature nero nets network networks neural neuro neurocomputing neuroevolution nineteenth noise nolfi notes nowlan oltp optimization organizing outstanding packet pages parallel parameter parisi parr pegasus pereira phenes planning play policy pomdps population ppsn practical precup predict presence press priority probabilistic probability problem problems proceedings proceedinngs proceeedings processing program programming propagation publishers pyeatt quadrupedal radcliffe radial rates reactive real recombination references reidmiller reinforced reinforcement representations research residual resource rewards richardson risk rivest robotics role rollouts routing roving rubenstein rule rumelhart ruppin safely saitta samuel sanghai sankhya santamaria sara sasaki scheduler scheduling schroeder sciences search second sekiguchi selection self sequential seventeeth seventh sharing sherstov shop sigkdd simple simulated singh sixteenth sixth smart smith soft solving space spaces sparse springer sprinkhuizen squares srivastava stagge stanley state stolzmann stone strategy studies study successful sutton suzuki symposium synapses system systems teaching temporal tesauro tests thesis third thomas threshold through tile time tokoro topologies topology towards tradeoff transactional transactions tree tsitsiklis twelfth twentieth uncertainty under understanding unifying university using utile utility value various verlag verma version video vision volume walsh watkins whiteson whitley wildstrom williams wilson witchel with without wolpert workshop yamasaki zhang zucker zweben http://www.jmlr.org/papers/volume8/tatti07a/tatti07a.pdf 208 Distances between Data Sets Based on Summary Statistics aaai acta advances agrawal aino algorithm american analysis andrew annals antwerp april artificial arun association axiomatization baldi baseville bayesian belgium belief best between binary bouldin buneman calders calinski capable cluster clustering codes combining communications complexity computation computational conference cooper correcting crossentropy csisz data databases david davies deduction deletions dendrite discovery distance distributions divergence dmitry doklady donald dover editors eiter events fast fayyad february francisco frasconi frequency frequent function geometry global gregory hailperin hand hannu harabasz heikki hollm ieee imielinski imre inequalities inference informatica information insertions intelligence intelligent international internet items itemsets jaakko jajodia jerzy john jordan jouni judea kaufmann knowledge kullback large levenshtein local logical machine management mannila mathematical measure measures method methods mich michael minimization mining mixture modeling models monthly morgan networks neural padhraic pages paolo pattern patterns pavlov pearl peter physics piatetsky pierre plausible point possible prediction press principles probabilistic probability problems proceedings processing publications publishers rakesh ramakrishnan ramasamy reasoning recognition references reversals rules separation sepp sets shapiro siam sigmod signal smyth solomon sons soviet spectral srikant statistics sushil swami systems tadeusz tatti their theodore theory thesis thomas toivonen tomasz toon transactions university usama using uthurusamy verkamo vladimir washington weiss wiley with yair http://www.jmlr.org/papers/volume4/bounkong03a/bounkong03a.pdf 253 Supp ort Vector Clustering academic advances alex algorithmic algorithms analysis annals approaches based berkeley bernhard biocomputing biology blake blatt bottleneck burges cambridge chichester classification cliffs cluster clustered clustering clusters computation computational conference cooper current data databases description detection determining diego dietterich dimensional discovering distribution domain domany dubes duda duin editors elisseeff englewood estimating eugenics examination expression eytan fast fisher fletcher fukunaga gene granular guyon hall hart high horn information international interscience introduction irregular jain jiang john kernel klaus learning leen letters lipson machine machines macqueen magnet marcelo markovian mathematical measurments merz method methods milligan minimal model muller multiple multivariate nature networks neural neurons novelty number observations optimization order pacific parametric pattern platt practical prentice press probability problems proc procedures proceedings processing psychometrika recognition references relaxation repository ripley robert roberts sara scholkopf sequential shai shamir shapes sharan shawe siegelmann slonim smith smola solla some sons springer stability statistical statistics stork structure support symposium systems taxonomic taylor theory thomas tishby todd topics training tresp university unsupervised using vapnik vector volker wiley williamson wiseman york zhang http://www.jmlr.org/papers/volume6/bar-hillel05a/bar-hillel05a.pdf 62 Learning a Mahalanobis Metric from Equivalence Constraints aaai academic acta adaptive adini adjustment advances allerton american analysis analytica annual application applications approach association background banff basu becker belhumeur bell bernhard bialek bilenko bilinear biometrika blake blind boreczky bottleneck boundary cambridge canada canonical cardie changes chechik chimica citeseer class classification classifiers clustering communication comparison compensating component computation computer computing conf conference constrained constraints content control correlation criteria cvpr data databases david deconvolution detection diego direction discovers discriminant discriminative distance easier edition editor editors eigenfaces equivalence european evaluation exploiting extracting face fawcett first fisherfaces francisco freeman from fukunaga functions gaussian geladi generative hasselmo hastie haussler hertz hespanha heyden hillel hinton html http ieee illumination images information instance integrating international interpretation jaakkola johansen jordan journal kamvar kaufmann klein knowledge kowalski kriegman kurtosis lawrence learning least level linear linsker machine mahalanobis making manning mardia maximization maximum means measures merz method metric michael mishra mixture mlearn mlrepository models mooney morgan moses most mozer multivariate nature nearest neighbor network neural newbury nielsen nina obermayer objective organising pages pami park partial pattern pavel pentland pereira preservation press principle prior problem proc processing projection rand random recognition references regression relations relevant repository retrieval rogers rowe russell sage saul scholkopf schroedl sebastian sejnowski self semisupervised separating separation shental shot side skewness space sparr specific spie squares statistical stereograms still storage structures style surfaces systems techniques tenenbaum than that thing thompson thrun thrune tibshirani tishby touretzky turk tutorial ullman uses using video vision volume wagstaff wahington weinshall with xing http://www.jmlr.org/papers/volume7/heiler06a/heiler06a.pdf 154 Learning Sparse Representations by Non-Negative Matrix Factorization and Sequential Cone Programming acoustics additional aerosol akamatsu algebra algorithms amari ambient anal analysis appl applic application applications approach arch asian athena atmospheric audio auto based bayesian belmont benchmarking berlin bertsekas biological blind boston boyd brown buciu calculus categories cbcl center chemometrics cheng chichester chichocki clustering codes coding comp complements computation computational concave cone cones conf constraint constraints continuous contour convex correct cplex csiszar cvpr data database datasets dealing decomposition denmark department developm differences differentiability divergences document does donoho dordrecht dual duality econom econometrics editor editors ellis emmcvpr energy environment environmetrics error estimates examples expression expressions face facial factor factorization family features fergus fisher floudas formulation france from functionals functions gabor generalized generative gesture give global gong graham grundlehren gyoba handbook heiler hiriart horst horwood hoyer http hyvarinen icpr ieee ilog image images incremental independent info inform intelligent interior intl into isbn israel kamachi kluwer kronecker laboratory layer learning learns least lebret lect linear lncs lobo local localized locally london lyons mach manual march massachusetts math matlab matrix mech methods minim minimization mittelmann model models mosek multi music natural nature negative netherlands netw network neur neural nips nonconvex nonlinear object objects october oper optim optimal optimization order over paatero pages pardalos parts pattern perona pitas point polyphonic ponstein positive press primal principle problems proc procedure products programming programs rangarajan receptor recog recogn recognition reference references relaxed representation research retrieval reverse revision robust rockafellar schnorr scientific second sedumi separ series seung shen siam sigir sign singer smaragdis socp software solvers spaces sparse sparseness spatially specific springer squares statist stodden sturm subsets symmetric syst systems tapper technique tested theory third tilburg toland toolbox tools training transcription transformation turk university updated urruty user using utilization values vandenberghe variation variational version vision visual visweswaran volume wang wavelets wets when wissenschaften with works workshop wright yuille zdunek zhang http://www.jmlr.org/papers/volume4/bach03a/bach03a.pdf 242 Regularized Principal Manifolds academic accepted addison advances aizerman algorithm alspector american anal analysis annual anthony applic applications applied approximation architectures arsenin artificial association atomic automation available bartlett based basis bennett berlin between beyond bishop boser boston bradley braverman burges cambridge carl cbms chen classification classifiers clustering code coherence combinations combining compact compactness component compression computation computational computer computing concave conference connection control convex corporation cowan curves data december decomposition dempster density dependences design dimension discrimination distortion donoho dover edition editor editors eigen elisseeff empirical entropy equations equivalence equivalent estimation fast flannery foundations francisco frie from function functions gammerman gaussian generalization generative gersho ghahramani girosi golberg graphical gray group grzywacz guide guyon hamermesh harrison hastie haussler hill http huber ieee incomplete incorrectly induction inference information integral intel international isbn john jones jordan journal july kambhatla kaufmann kernel kernels kimeldorf kluwer krzyzak laird large learning least leen library ligence likelihood linder linear lkopf lloyd local london loqo lugosi machiens machine machines madison mangasarian manifolds manual mapping margin massive math mathematical mathematics maximum mcgraw mercer method methods mika minimax minimization minos mlps models morgan morozov motion muller murtagh negative networks neural neurocolt noise nonlinear numbers numerical observational operations operators optimal optimization pages pattern performance philadelphia philosophical physical pittsburgh plenum poggio posed positive prediction press princeton principal problem problems proceedings processes processing programming publication publishers pursuit quantization quantizer reading recipes recognition reduction redundancy reference references regional regression regulariza regularization regularized remote report reprint research results review revised robust roweis royal rozono rubin rvdb saunders scholkopf schuurmans scientific series sheffield siam signal smola smoothing society solution solutions solving some sons sparse spline springer squares stanford statistical statistics stephani stitson street stuetzle subroutine support suppport svens systems tchebycheffian technical technische tential tesauro teukolsky their theoretical theory thesis tikhonov tion topographic training trans transactions type unifying universit university user value vanderbei vapnik vector verlag version vetterling vision volume vovk wahba washington watkins wesley weston wiley williams williamson winston wisconsin with workshop york yuille zeger