http://acl.ldc.upenn.edu/acl2004/emnlp/
EMNLP 2004
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Toutanova.pdf	28	The Leaf Projection Path View of Parse Trees: Exploring String Kernels for HPSG Parse Selection	accompany active acyclic advances algorithm apllications application aravind baldbridge based beyond burges carl cation charniak chicago chris christina christopher classi coling collins colt conll constraint convolution cristianini cruz csli data david decision detlef directed disambiguation discrete discriminative distribution duffy editors eisaku ensemble entropy estimation eugene experience fast feature flach flickinger free gaertner generative grammar grammars graph haussler headdriven hierarchical hirao hpsg huma icml inexact inspired invited ivan iwpt jason joachims john johnson jonas joshi kernel kernels kristina kuang kuhn language large learning leslie lexicalised lexicalized libin linear lingo linguistic lloyd lodhi maeda making manning mark matching maximum measures methods michael miles modeling models motivation naacl natural nello nigel nips oepen osborne pages paper parameter parse parser parsing peter phrase pollard practical practice preliminary prescher press proceedings publications redwoods references rens report reranking rich riezler santa sasaki scale scholkopf selection shawe shen shieber smola statistical stefan stephan stochastic string structure structured structures stuart support suzuki svmbased talk taylor technical text theories theory thomas thorsten three toutanova training treebank treebanks trees tsutomu university using vapnik vector vladimir voting watkins wiley with written york yutaka
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Swier.pdf	19	Unsupervised Semantic Role Labelling	aaai abney academic acquisition alternations ambiguity analysis annotated annual answering applications approach argument arguments arti assoc association automatic baker bank based berkeley berwick bootstrapping british case cation chapter chen chicago choice chunks cial clark class classes classi coling computational computers conceptual conf conference consistent construction corpora corpus criteria dang deep disambiguation distinctions distributions domains ecml edition editors empirical english entropy estimation european evaluating experiments features fillmore fleischman foundations fourteenth frame framenet generative gildea guide hacioglu hierarchy hindle hirst hovy http humanities ijcai intelligence international investigation issue jones jurafsky kingsbury kipper kluwer kwon labeling language large learning levin levy lexical lexicon linguistic linguistics lowe machine manning martin maximum mccallum medical meeting merlo methods mining model models national natural necessity nigam palmer parsing pradhan predicate preliminary press principle probabilistic probability proc proceedings processing project proposition publishers question rambow recognition reference references relations restricted riloff rivaling role roles rooth schmelzenbach schulte second semantic sense senseval special statistical stevenson structural structure submitted supervised syntactic systems tagging tasks techniques tenny text thompson university unsupervised using verb verbargument very walde ward weir word workshop world yarowsky
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Ng.pdf	42	Chinese Part-of-Speech Tagging: One-at-a-Time or All-at-Once? Word-Based or Character-Based?	academic accuracy aids algorithm also analyser analyzer annotation approach bakeoff based between categorial chang character characters chinese chiou combination compression computational consistency developing differences direct eacl effective efficient emerson emnlp encode english ensuring entity entropy finite first florian fluidity from full fund gale gives good grant guidelines have hhmm howtogetachinesename huang ictclas implications improved individual ineffective information integrating international issues ittycheriah jing kovarik kroch kwong language lexical linguistics lrec made marcus maximum mcnab method model more much national okurowski palmer parser parsing part partial partially partof porting proc ratnaparkhi references relatively research revealed segmentation shen shih sighan singapore speech sproat state stochastic study supported tagging teahan testing text than that this time training tsou university using with witten word wordsegmentation workshop xiong zhang zhou
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Quirk.pdf	25	Monolingual Machine Translation for Paraphrase Generation	accomplished acoustics acquisition aligned alignment alignments alternations appear approach articles automatic available backing barzilay based best between binary broad brockett brown building cambridge canada capable china codes coling comparison computational conf conference construction continuous copenhagen corpora corpus correcting data database deletions della denmark detroit dirt discovery doklady dolan domain driven eacl electronic empirical encountered engineering evaluation exactly example exercise exploiting extracting fast fellbaum field finding from generate generating goodman gram hong http huang humans hypotheses ibrahim ieee improved inference insertions international ittycheriah japan jcluster joshuago kambhatla katz kneser knight knowledge koehn kong language large learning level levenshtein lexical linguistics louisiana machine marcu massively master matching mathematics mckeown melamed mercer methods microsoft mihalcea mining modeling models monolingual monotone multiple naacl news novel oriented orleans other pang pantel parallel paraphrase paraphrases paraphrasing pedersen phrase physice pietra preferred presented press problem proc proceedings processing produced quirk recognition references research reversals roukos rules sapporo search second sekine semantic sentence sentences sequence sequences shinyama showed sigkdd signal software soong sorts sources soviet speech statistical structural sudo sumita summit syntax system systematic techniques text texts that thesis this tillmann toronto toward translation translations tree trellis tribble truecasing unsupervised used using various venugopal vita vogel waibel which while word wordnet work workshop zhang zhao zubaiga
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Lita.pdf	57	Instance-Based Question Answering: A Data-Driven Approach	aaai abdessamad advances agichtein algorithm always annotation answer answering answers approach asked automatic badulescu banko based bayes before better bikel biomedicine boosting brill brown bunescu burger callan carbonell carroll categorization cation channel chua clark clarke classi classify clef clustering coden cogex coling collins comparison constraints context cormack czuba daniel derivation discovery disertation domain driven dumais durme dynamic eacl echihabi engine engines entropy event eventbased exact exploitation external facilitate factoid falcon ferro fleischman frederking gerber girju goodrum gravano greiff harabagiu head heads henderson hermjakob herrera hiyakumoto hovy huang huttenhower ittycheriah joachims judy junk kemkes knowledge kupsc language laszlo lawrence learning learns licuanan light lita logic lynam machine machines magnini maiorano marcu mardis maximum mccallum mihalcea mitamura mitre models moldovan morarescu more multi multiple naacl naive name natural nigam nitional noisy nyberg online open overview parsing part pasca patterns pedro peiado penas performance pinpointing planning prager predictive prover qanda query question questions radev ravichandran references reformulation relations resource rijke romagnoli roukos rule samn schwartz search selection semantic sigir some speci speech statistical strategies strategy structure structured support surface svoboda system tagging terra text than that they thompson tilker track transformations trec using vallin vand vector verdejo voorhees wang webclopedia weischedel what whole with workshop yang zhang zweigenbaum
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Daya.pdf	52	Learning Hebrew Roots: Machine Learning with Linguistic Constraints	aaai abraham acquisition advances ambiguities analysis analyzer annual approach approaches april arabic association august aviv based beesley benjamins building cambridge canada cards cation choueka class classi coling comparison computational computers computing conference conll csli daelemans darwish david dictionary disorders editor editors edits edmonton education eizenberg emnlp empirical entity erik even exact exhibition fien final florian foreword full ging grammatical haifa hamillon hapoal haxadash hebrew house iaai independent inference information inquiry international internet introduction israel jerusalem john joseph july kareem kenneth kiryat kiryath kruskal language languages learning line linguistic linguistics luxot macromolecules madison markov mccarthy meeting memory methods meulder michael mike miles mlim model modern montreal morphological morphology multi multilingual named natural nerbonne neural nips nite nonconcatenative number only operations ornan osborne overview page pages part philadelphia practice press proceedings processing prosodic publications punyakanok quebec radu recognition references reprint resolve root rosner roth sang sankoff schutze sefer semitic sepher sequence sequential shallow shared shimron shoshan shuly sigdat singer speech stacking stanford state string system systems tables taiwan task theory time tjong university using uzzi variable vasin verb walter warps wintner wisconsin with word workshop yaacov yehuda yizxaq zdaqa zohar
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Keller.pdf	47	The Entropy Rate Principle as a Predictor of Processing Effort: An Evaluation against Eye-tracking Data	able achieves also analyses analysis annual another artifact association assumption averaged aylett based baseline between british burnard cachebased cant carpenter cases charniak clarkson cmuin cognition cognitive collins comprehension computational computed computing conclusions conference congress connected consortium constancy constant context contribution controlled corpus correlate correlated correlation cult culty data difference different dmitriy duration editors effect effort empirical enabled entropy equally esca essentially eucambridge eugene evaluated even expected experiment experimental factor first found francisco from function future generalizes generate genzel greece guide holds however human ieee important individual inference informativeness intelligence international involves isolation jerome journal just kuhn language larger learning length level linguistics lorch machine made many marcel mark matthew mcdonald means measured measures meeting memory methods michael model modeling mori movement movements myers national natural newspaper normalized normalizing number obtaining over oxford pages paper parse partial patricia pattern philadelphia philip phonetic position positions possibility prediction predictions predictive predictor predicts principle probabilities probability proceedings process processing proposed prosodic psychological psychology raises range rate read reading record redundancy references regression relationships remains renato repeated replicated reported reproduction research restricted review rhodes richard robert roland ronald rosenfeld rospeech same sampled sapporo sciences scott sentence sentences service shillcock should show showed signi simple simply speakers speech statistical steedman stochastic structure subjects subset suprasegmentals syllabic test tested text that theory there this times toolkit tracking transactions transitional trees uence underlies understanding university used users using variation vision were what when which with work xations
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Halpin.pdf	23	Automatic Analysis of Plot for Story Rewriting	abney about academic acquisition across adventures agents allen american analysis andrei approach association automatic baldwin bear benjamin beth bhembe biarritz breck bringing burstein cambridge cation charniak children christine chunks city claire cogniac coherence cole colin company computational conference construction continuing cross cummings daniel database decision dependencies discourse doubleday dumais east editors education electronic engine engineering entropy environment environments erik essays eugene evaluation evidence feedback fellbaum finding flexible foltz foundations frame france garden georgia goals graeme green grover hastings hickmann high hirst hybrid identi ieee induction inspired intelligent interactive interface international issue james jennifer jerry jill johanna journal judy kevin kintsch kluwer knight knowledge lagerloff landauer language latent learning level lexical life linguistic linguistics long machine mani marc marcu martha matheson maximum maya meaning measurement menlo mikheev model moens moore morgan mueller multi multiple murray naacl narrative natural need news nils nirenburg north page pages park parser perceptions person peter plato pollack precision press problem proceedings processes processing pronoun psychological publishing quinlan references representation resolution resources review roberston robertson robust roque rose ross second selection selma semantic sergei shanahan siler solution solving space special srivastava steven stories story storystation stroudsburg structure student stuff susan syntax systems teacher temporal text textual their theory thomas through time tokenisation tool trees tutoring understanding university vanlehn volume walter wiemar wilson with wonderful wordnet workshop write writing york
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Kalt.pdf	9	Induction of Greedy Controllers for Deterministic Treebank Parsers	abney accurate advances adwait ambiguity american annotated annual applications applied approach arti association automata based beatrice black brants brian brieman brown building cation cessing chapman chapter charles charniak chart christopher cial classi collins communication computational conference context corpus darpa david decisions dekai dels deterministic dissertation driven edge eech empirical english entropy entropyinspired erty eugene europ exact examples factored fast fernando fourteenth friedman from gimenez goldwater gram grammars hall head hermjakob history information intel jelinek jerome jesus johnson klein language large learning ligence lightweight linguistic linguistics luis magerman manning marcinkiewicz marcus mark marquez mars mary maximum mcallester meeting mercer methods michael mitchell motivations national natural neural north olshen pages parse parser parsing part pattern pcfg penn pennsylvania pereira predictive probabilistic proceedings processing ratnaparkhi recent recognition references regression relating representations resolution revisited rich richard richer roark robust roukos santorini sharon sixth speech stanford statistical steven stone syntactic systems tagger tagging technology texas thorsten towards translation tree treebank trees university unlexicalized using with wong workshop
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Gliozzo.pdf	55	Unsupervised Domain Relevance Estimation for Word Sense Disambiguation	albany athens automatic beall boosting british categorization cation cavaglia classi codes coling comaroni computational conference consortium corpus database decimal dewey disambiguation domain edition editors electronic engineering escudero evaluating evaluation fellbaum forest france gliozzo greece http index information integrating international into ject july june language lazy learning lexical linguistics lrec magnini many marquez matthews national natural pages pezzulo press proc proceedings references relative resources rigau role saarbrucken second sense senseval strapparava system text toulose unsupervised using word wordnet workshop york
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Riezler.pdf	29	Incremental Feature Selection and 1 Regularization for Relaxed Maximum-Entropy Modeling	acapulco accuracy advances alexander algorithm algorithms american analysis andrew annotated annual appendix arti association banff based beatrice between boosting boston boyd building cambridge canada canon carnegie cation chapter chen cial ciently college comparison computational conditional conference conjugate conll constraints convex corpus crouch deep della department descent discriminative dual elds emnlp english entropy estimation estimators evaluation exponential extension fast feature features following from fuliang function functional gaussian geman goodman gradient grafting grammar grammars hauke human ichi ieee incremetal inducing inequality information intelligence interface international invariance james japan john johnson joshua journal kaplan kazama kevin king lacker lafferty language large lasso learning lebanon lexical lide lieven likelihood linguistics logistic machine malouf manuscript marcinkiewicz marcus mark mary maxi maximum maximumlikelihood maxwell mccallum meeting mellon mexico microsoft minka mitchell mizes modeling models naacl natural neural nips norm north optimization otherwise parameter park parsing pattern penn perkins philadelphia phrasal pietra pittsburgh press prior priors proceedings processing proof proposition random redmont references regression regularization report research richard riezler robert ronald rosenfeld rotational royal santorini sapporo schmidt selection series shallow show shrinkage simon since smoothing society space speed stanley statistical statistics stefan stephen stochastic street stuart taipei taiwan technical techniques technology that theiler then thomas thus tibshirani tracy transactions treebank tsujii uncertainty university unpublished using value vancouver vandenberghe vasserman vincent wall weng with yaqian zhiyi zhou
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Baldridge.pdf	8	Active Learning and the Total Cost of Annotation	about acquisition active advances annual anoop applications arti associated automating baldridge based best boston brants building cannon cation change china christopher cial cient cohn committee computational conf conference consider considered continuum corpus corrected cost creating data david discussion editors either else emnlp engineering ensemble especially estimators eventually example experts exploiting factored flickinger form from geman genuine ghahramani going grammar grammars haim hinton hong hpsg icml improving induction information international into issue itself jason johnson joint jordan july kong kristina labeled labeling language learning leen lingo linguistics manfred manning mark meeting michael miles minimizing model models more motivation naacl natural naturally need networks neural oepen only opper osborne otherwise pages parse parser parsers parsing pennsylvania philadelphia preliminary press proc proceedings processing products prudent query random rather rebecca redwoods references riezler roukos salim sample sarkar selection semi seung shieber should sigdat some sompolinsky special statistical steedman stephan stephen stochastic strategy stuart substantially systems taipei taiwan tang task techniques tesauro than then theory there this thorsten total toulouse touretzky toutanova training treebank turn types uncertainty understood unlabeled unlikely used using volume well when will with workshop would xiaoqiang zhiyi zoubin
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Zhou.pdf	62	Multi-document Biography Summarization	aaai about above acknowledgement actors address algorithm american analysis anlp annual answer answers articles artificial artistry assembly association audiences august automated automatic background barry barzilay based berkeley berlin beth biographical biographies biography biographyrelated body brill brunswick cambridge carbonell case categorization chang chen chih chin chronological cikm cjlin class classifier coling combining communications composing compression computational concise conclusion conference construction consult consummate contain content context controversial cooccurrence core corpora corpus creating cross csie daniel database december described development developments died diego directly discussions document does dragomir ecml editor eduard electronic elhadad ended english enthralled environment errordriven evaluation events explain explanatory extracting extraction features fellbaum film fixed fluent follow form francine free from further fusion galliers generated generating generation george gielgud goldstein gram great greenaway hill home horacio hovy http human include inderjeet indirectly inferring information insightful intelligence interest international into introduction introductory investigate investigation items jade jaime jair jersey joachims john jones julia julian kantrowitz karen kathleen kaufmann kevin knight knowledge kristian kupiec language large last learning lecture length lexical liang library libsvm lies life like line lingual linguistic linguistics listing lovins machine machines main major management mani many march marcu maryland mateo mcgraw mckeown means mechanical meeting meta michael might miller mitchell mittal morgan multi multidocument naacl national natural neats news noemie normal notes november obtain occurring online open order ordering other outside pages paper part pedersen person peter plan point possibly press problem proceedings processing producing programs prospero provide purpose querying question questions quinlan radev recent references regina repositories research resources retrieval review rhapsody role rouge saggion scale schiffman sentence sentences shakespearean should sigir simone sources special specifically speech springer stage starting statistics statisticsb stemming step strategies structures study summaries summarization summarizer summary sunday support surreal system systems tagging tailored taipei taiwan techniques technology tenth teufel text texts than thank that them themselves this thorsten though through trainable trained transformation translation university uses using vector vibhu voice volumes what will with wordnet workshop would years zhou
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Daume.pdf	22	A Phrase-Based HMM Approach to Document/Abstract Alignment	agreement alignment annual aposteriori approach arti assessing association august automatic available banko based berkeley beyond brown carletta cation chains channel christoph christopher cial classi coling comparison compression computational conf conference construction corpora daniel daume december decompose decomposition della development docs document estimation extraction foundations franz gaussian gauvain generation hdaume headline hermann hidden hinrich hong hongyan http human ieee information intelligence jean jing josef kappa kathleen kevin knight kong language large linguistics machine manning marcu markov mathematics maximum mckeown meeting mercer michael michele mittal mixture model modeling models multivariate natural noisy observations october pages parameter pbhmm peter phrase pietra press probabilistic proceedings processing references research retrieval robert scale schutze sentence sentences sigir statistic statistical stephan stephen summaries summarization summary systematic tasks tillmann transactions translation unpublished using various vibhu vincent vogel witbrock word written
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Cucerzan.pdf	44	Spelling correction as an iterative process that exploits the collective knowledge of web users	above absolute accuracy acquisition additional agreement analysis applying approach arguably attempt automatic automatically based basil bayesian binary blackwell brill brodt byrd channel cherkassky chodorow church coling collective column comm communications components computational computed computer computers computing conclusion confirm consider context contextbased contextsensitive contextual corpora corpus correct correcting correction correctness could critiquing damerau design detection development different difficulty each english epistle error errors evaluation experiment experiments first fisher gale garside give gold golding google grams hall hanson heidorn hybrid icml ieee importance improved information investigations jensen jieee journal jurafsky kernighan knowledge kukich language large leech list logs longman lower made management mangu martin mays mcilroy measure measures measuring mercer method miller model modeling moore needed noisy number observed oxford pages paper peterson philosophical postprocessing precision prentice problem proceedings processing program programs pronunciation queries query reasonable recall references refers relative result rieseman roth rule sampson search second sensitive sent show shows slightly something speech speller spelling springer standard stored string successful such suggested suggestion suggestions surveys synthesis system systems table technique techniques test text than that them they this times total toutanova trans transactions useful usefulness using vassilas verification verlag very wagner were when while winnow with wittgenstein words workshop
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Hana.pdf	35	A Resource-light Approach to Russian Morphology: Tagging Russian using Czech resources	alena algorithm alignment american analy analysis analyzer andrei andrey anlp anna annotated annotation appear approach april article articles association atala automatic barbora based beatrice bemova blackwell bloomington brants building case chapter cognates colloquium combination company comprehensive computational conference copernicus corpora corpus czech data david dependency design dictionaries dictionary dublin dzeroski eacl east ected elworthy engine engineering english erjavec evaluating evaluation fast feldman france from generation goldsmith grammar greg guessing hajic hana hladka html http identifying ilya indiana induced international iseg issues jakub jarmila jean jiri john june karel kondrak kovalev krbec kveton language languages large lasvegas learning linguist linguistics liubov liubushkina marcinkiewicz marcus mary maslov meeting michail midwest mikheev minimally mitchell models morphological morphology morphosyntactic multext multexteast multilingual multimodal mystem naacl narod natural north ofspeech oliva pages panevova paris part pavel penn petkevic phonetic portable prague probabilistic proceedings projects references resources richard rules russian santorini saso search seattle second segalovich semantic serge serial sigdat similarity slovene statistical statistics stemka study supervised syntactic synthesis tagger taggers tagging tags tagset tagsets technology terence texts thorsten titov tomaz toulouse treebank ukranian univaix unknown unsupervised veronis vextal vitaly vladim wade washington wicentowski with word words workshop yablonsky yandex yarowsky zavrel
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Agirre.pdf	10	Unsupervised WSD based on automatically retrieved examples: The importance of bias	accent acquisition agirre ambiguity amoros annotation annual anual appear application applied argamon arpa arti association atserias automatic available barcelona based bench bootstrapping brants bunker cambridge canada carroll castillo cation chodorow cial classi clustering coling colingacl collocation committee computational concordance conference conjunction content corpora corpus cotton cruces dagan daude decision disambiguation eacl edmonds empirical engelson evaluating evaluation examples exploring exretriever fernandez finding france french genre gonzalo hong human identi information intelligence intelligent international joint journal kilgarriff koeling kong language large leacock lexical lexicographic linguistics lisbon lists lopez lrec luxembourg mapping martinez mccarthy meeting methods mihalcea miller montreal natural nominal overview padro pages palmas part portugal predominant princeton probabilistic procedings proceedings processing publicly references relations research resolution resources restoration retrieval rigau rivaling sample seattle second selection semantic sense senses senseval sigdat signatures similar sixth spain spanish speech statistical statistics structural supervised supporting systems tagged tagger technology tengi text tool topic toulouse tugwell turmo uned unsupervised untagged using variations verdejo very volume wasp weeds with word wordnet wordnets words workshop yarowsky
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Sporleder.pdf	16	Automatic Paragraph Identification: A Study across Languages and Domains	aaai adam adwait alexander alistair annual applied approach arti association attribution authorship bala based beeferman behavioral berger boostexter boosting boundaries boundary broadcast budapest cambridge castellan categorization chapter character charniak christensen chunshen cial clarkson coherence computational conference corpus critique dale data department detection discourse dmitriy domain doug draft driven edinburgh empirical entropy esca eugene europarl european eurospeech evaluation experiments fall fifth from fuchun function functional gaizauskas genzel gotoh grammatical hauptmann head hearst heather heidi hill hitoshi http identifying immediate improvement independent information informedia integrating intelligence into isahara jeffrey john keselj knott koehn kolluru lafferty langauge language learning level linguistics longacre machine mark markings marti masao maximum mcgraw meeting meta methodology methods metric michael model modeling models more motivating multi multilingual natural news number once pages paragraph parametric parse parsing passages peng pevzner philip philipp proceedings processes processing project publications ratnaparkhi references relations renals retrieval reynar rhodes robert ronald rosenfeld sapporo schapire schuurmans sciences seattle segmen segmentation segmenting semantics sentence sentences shaojun sidney siegel singer smith speci speech stark statistical statistics steve stevenson style subtopic summarisation sunderland symposium syntax system tation text texttiling thesis toulouse translation trees unit university unpublished using utiyama variation video vision vlado wang washington what yoram york yoshi
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Chklovski.pdf	11	VERBOCEAN: Mining the Web for Fine-Grained Semantic Verb Relations	aaai abelson acapulco accomplish accuracy acknowledgments acquiring acquisition adding addition algorithm alternations among analysis annotation answering antohe antonymy appear appendix applications approach aspects assemble assertions assess assigning associated associates austin authorize authors automated automatic automatically available badulescu baker barker barwise barzilay before behavioral benefit berkeley between bigrams bolohan boston brisbane cafarella california canada castellan certain chicago chinese classbased classes classification clause click coling comments complete computational concepts condemn conference construction contrasts corpora coverage create cruse dang database demonstrated derive detain diego discovering discovery dismantle distributional distributionally document dolan double downey ecml electronic eleventh elhadad enablement engineering english enhance enhanced enroll entailment erlbaum etzioni examples experimental extractable extracted extraction fellbaum fields fillmore fine forms framenet frames france freiburg frequencies from further gardent germany girju glosses goals gomez google graduate grained grant happens harabagiu harris have hearst helpful high hill hillsdale hope hovy human hyponyms identification identified identifying ijcai implied index inference inferring information innocence inquiry interactive international interpretation into investigation jair journal katz keller kingsbury kipper kittay klavans knowitall knowledge labeling lacatusu language lapata large lawrence learning lehrer level levin lexical lexicography lexicon linguistics logic lowe lrec machine marcus martinich maximize mcgraw mckeown methodist methods mexico miller mindnet mined mining models moldovan montreal morarescu multidocument naacl nantes natural network nonparametric notebook noun novischi obtain online open opposition ordering other oxford pacling pairs palmas palmer pantel partly patternbased patterns penn permit perry philadelphia philosophy phrases pittsburgh plans popescu position preliminary present press proceedings process produce project prosecute querying question ravichandran reduce references refine regard relation relations relationship relationships reschedule research resource restrict results retrieval review reviewers revisited richardson roast robust role rules sample scale schank schedule sciences scripts semantic semantics sentence shaked shock show siegel similar similarity simple situations soderland some southern spain startle stated statement statistics strategies strength strong structure structures structuring such summarization supported supporting surface surprise synonyms system szpakowicz tatu text thank that their thesis this thoughts toefl tools topic translation trec treebank turney uncompromising understanding university unseen using vanderwende verb verbs versus volume webber weld which wide wish with wordnet words work workshop yates york zhao zhou
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Kudo.pdf	36	Applying Conditional Random Fields to Japanese Morphological Analysis	accurate aided algorithm analysis andrew appear arti asahara atsushi bfgs bound bruce byrd carnegie chen chikashi chinese cial ciently coling computing conditional conference conll constrained corpus croft data david dayne della descent detection dictionary dong early elds emnlp entity entropy exponential extended extraction fangfang fast feature features feng fernando freitag from fuchun function gaussian goodman gradient grafting high hitoshi icml ieee incremental inducing induction information intelligence isahara james japanese jmlr john jorge joshua journal kevin kiyotaka labeling lacker lafferty large lexicons limited machine markov masayuki math matsumoto maximum mccallum mellon memory method models morphological naacl named nineteenth nobata nocedal optimization pages papers parsing part pattern peihuang peng pereira performance perkins pietra pinto prior priors probabilistic problem proc programming random recognition references report research results richard ronald rosenfeld satoshi scale scienti segmentation segmenting sekine selection sequence shallow siam sigir simon smoothing space speech spontaneous stanley stephen table tagger technical thiler tools transactions uchimoto uncertainty university unknown using vincent webenhanced with word xing yamada yuji
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Chelba.pdf	43	Adaptation of Maximum Entropy Capitalizer: Little Data Can Help a Lot	accumulating adaptation advances adwait after algebraic algorithm algorithms allowed amount andrew appendix approach argument arti association audio automatic baker based basic because becomes berger boston bounding brill capitalization carnegie caused change chen church cial cients closely coef collins computational computer conditional conf conference const context convenient convexity corpus covariance daniel darpa data decrease degree della derivation derivative design diagonal difference discriminative does doug dumais each editors elds empirical encountered entropy equality equation equivalent eric experiments exponential features february fernando fields following follows francisco gaussian generation goodman guarantees hand hidden highest hltnaacl holds hwan ieee increase independent inducing inequality input intelligence interna iteration ittycheriah janet january japan jensen jersey john joshua journal july kambhatla kaufmann kenneth labeling lafferty language largest last learning likelihood linguistics lita loglikelihood lower machine main manipulations march marcu markov massachusetts matrix maxent maximized maximum mccallum mean mellon method methods michael model models modi morgan national natural nding newton nonnegative obtain pages part partial paul pennsylvania perceptron pereira philadelphia philip pietra pittsburg polynomial possible predicted prior priors probabilistic proc proccedings proceedings processing random ratnaparkhi reduces references regularized report respect respectively results right ronald root rosenfeld roukos salim same sapporo school science second seek segmenting sequence side since smoothing solution solve solving some somerset speech stanley street substi summation survey susan tagging taking technical technique techniques that then theory this thus tional training transactions transformation truecasing tute unique university update updates using value values very wall where which whose with woodland workshop
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Nepveu.pdf	31	Adaptive Language and Translation Models for Interactive Machine Translation	accurate acoustics adam adaptation adaptative alberta alignment among analysis annual approach april association august based berger beyond brown budapest building cache cachebased canada chapter cherry cient clarkson cocke colin coling compositionality computational conference constraints data decaying dekang della dependent description divergence driven eacl editors edmonton emnlp empirical entropy estimation european eurospeech evaluation exercise exponentially foster france franz fredrick friendly george hermann hong hungary ieee improved intelligence interactive international jelinek john josef july june kong kuhn langlais language lapalme learning liermann linguistics machine martin mathematics maximum meeting mercer methods michel mihalcea minimum mixtures model modelling models moore mori munich naacl natural october pages pami parallel parameter pattern paul pedersen peter philadelphia philippe pietra prediction proalign proceedings processing rada recognition references relationships renato robert robertson roland roossin search shared signal simard simple speech statistical stephen system task text texts topic toulouse towards transactions translation translators transtype user using varigrams vincent with word words workshop zens
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Nielsen.pdf	17	Mixing Weak Learners in Semantic Parsing	aarseth adding advances algorithms amit annotating annotation application approximate argument arti automatic bagging bartlett berkeley bies boosting boston breiman britta cambridge cation cial classi colorado comparing computation computational computer conference cslr daniel data decision decisiontheoretic dietterich donald estimation extraction ferguson forests freund geman generalization gildea grace hacioglu harabagiu hill html http improvements induction info information intelligence james john journal jurafsky kadri karen katz kingsbury krugler labeling large learning line linguistics machine machines macintyre marcinkiewicz marcus margin mark martha martin mary mcgraw mihai mining mitch mitchell necessity neural nielsen other palmer parsing paul penn platt pradhan predicate predicateargument predictors press probabilities probability proceedings quantization quinlan random randomforests randomized recognition references report robert rodney roles rulequest sameer sanda schapire schasberger schuurmans sciences scolkopf semantic shallow shape smola statistical statwww structure structures supervised support surdeanu systems technical tests thomas tools treebank trees uncertainty university users using valerie vector ward wayne williams with yali
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Fung.pdf	14	Mining Very-Non-Parallel Corpora: Parallel Sentence and Lexicon Extraction via Bootstrapping and EM	academic acquisition adaptive alexander aligned aligning alignment ambiguity american anglo annual approach april asian association barzilay better bilingual bing bootstrapping brown calculated cambridge canada cheung christopher church classifier coling collections collocations comparability comparable computatinal computational conference conjunction corpora corpus cross dale daniel data dekai dekker della determine disambiguation distributed dordrecht dragos each editor editors edmonton elhadad emnlp engineering estimation example extraction finally finding foundations framework frank franz fraser from fung future gale genichiro grefenstette gregory handbook harold hermann hinrich hiroyuki hong human identifying ieee improved incorporated increases indian information into isbn japan jean josef june kaji kathleen kenneth kikui kluwer kong language languages large learning lesal lexical lexicon like linguistics machine manning marcel march marcu maryland matching mathematics mckeown meeting mercer methods mining mixed models moisl monolingual mumbai munteanu naacl natural ness news noah noemie nonparallel north other pages pair pairs parallel parameter pascale performance peter philip pietra press proc proceedings processing program proposed provide publishers quantified query rapp references regina reinhard resnik resolving results retrieval retrieving robert sapporo score scores sense sentence sentences september show shun smadja smith somers south statistical stefan stephan such suggest technology terminology text texts that this translating translation translations type unsupervised used using veronis very vogel with word words workshop would xiaohu xtract york yuen zhao
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Civera.pdf	51	From Machine Translation to Computer Assisted Translation using Finite-State Models	aachen acoustic acoustics aided alberto algorithm alignment amengual american americas amta andreas andres andrew annedore annex annual antonio appliquee approach approaches april assisted association asuncion asymtotically august automatic baltimore bangalore barrachina based bened bleu bounds brown carlos casacuberta castano castellanos celer center centre chapter china cisco codes comparison completion computational computer conf conference continuous convolutional ctor curin data david decoding della devices dieter driven editor editors engineering enrique error estimation europe eutrans evaluation experimental farwell federico final finite foster fran francisco franz gamma garc george gerber hermann hochschule hong hopkins hovy icassp ieee improved informatica informatik information informatique instituto international ismael jahr jimenez jing john johns jose juan july kevin kishore knight kong laboratory lafferty langhorne langlais language lapalme learning lecture lehrstul linguistics linguistique llorens london machine mart marzal mathematics meeting melamed mercer mergel method methods michael models moises molau montreal munich nevado nite noah noll north notes october onaizan optimal organization paeseler pages papineni parameter pastor paths peter philadelphia philippe pico pietra prat press proc proceedings processing project purdy puting recherche recognition references report research rheinisch ricardi robert roukos salim sanch sanchis schlumbergersema science second sergio shortest signal sirko smith societe soluciones some soup speech springer state statistical stephen stochastic system technical technische tecnologico theory third todd transactions transducers translation transtype typing unit university varea verlag vidal vilar vincent viterbi vitter volume ward westfalische with workshop xerox yarowsky yaser zaroliagis
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Gildea.pdf	34	Dependencies vs. Constituents for Tree-Based Alignment	above aligned alignment american annotated annotation annual approach arpa association automatic automatically based beatrice bikel bilingual bracketing brown building chapter chinese christopher cocke cohesion collins companion comparison computational conference consistently constit constits constituent corpora corpus correspondence daniel dekai della dependency design ding driven eisner emnlp empirical engine english estimation evaluating first france franz frederick galley generated gildea grammars guidelines hand harder head heads heidi hermann hong hopkins human ijcnlp improved international inversion ircs isomorphic japan jason jelinek john joint josef june kenji kevin knight kolak kong lafferty language large learning levy lingual linguistics loosely machine manning mappings marcinkiewicz marcu marcus mark martha mary mathematics meeting mercer methods michael michel mitchell model models multi naacl natural nianwen north october okan pages pairs palmer parallel parameter parse parsing paul penn pennsylvania peter philadelphia philip phrasal pietra proceedings processing projection proportion rebecca references report resnik robert roger roossin rule santorini sapporo statisti statistical stephen stochastic syntaxbased table technical technology thesis toulouse transduction translation translational tree treebank treelet trees university using vincent volume weinberg what with words workshop yamada yuan
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Andrew.pdf	26	Verb Sense and Subcategorization: Using Joint Inference to Improve Performance on Complementary Tasks	access accuracy accurate acquisition across adam address algorithm also although amsterdam analysis analyzing anna annual appear applicable approach arguments association attempt basic basis benjamins besides bikel case cations changing chapter chose christopher clauses cognitive collins colorado communications computational computers conditioning conference constructions corpora could daniel data database david decision dempster determination disambiguation discussed distribution diverse done doublegeneration douglas driven editors empirical engineering english especially estimate evaluating example extracting florian framework from future head headwords hierarchical hope humanities imagine importantly improve improvement improving incomplete information issue japan john joint joseph journal judita jurafsky kilgarriff klein korhonen laird language large lexical lexicon likelihood linguistics lists made manning maximum meeting merlo methodological methods michael might miller model modeling models modi more most movement natural nding only other pages pairs paola paper parameter parser parsing particles parts passive perceived performance phrases possible preiss prepositions probabilistic probabilities probability proceedings processing radu references relative results roland rosenzweig royal rubin sapporo science second section sense senseval sentence separately sequence series several sigdat similar society some spaces speci speech statistical stevenson structure subcategorization such suzanne syntactic system tags tasks test that then thesis third this types under underlying university unlexicalized useful using verb verbs very with word wordnet words work would yarowsky
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Moore.pdf	49	On Log-Likelihood-Ratios and the Significance of Rare Events	accurate acquiring acquisition agresti alan alberta alignment analysis applications arti attneave between beyond brian bruce building cambridge cant categorical choice cial coincidence collocations computational computing conference cover data diana driven dunning edition edmonton elements england evaluation exercise flannery fred graeme group hirst holt information inkpen intelligence interest john kayaalp lexical lexicon linguistics machine mehmet methods mihalcea naacl national nearsynonyms numerical oregon parallel pedersen pennsylvania philadelphia portland press proceedings psychology rada rebecca recipies references relationships rinehart saul scienti second siglex signi sons special statistics surprise teukolsky texts theory thomas translation university unsupervised using vetterling wiley william winston word workshop york
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Bikel.pdf	30	A Distributional Analysis of a Lexicalized Statistical Parsing Model	abney academic advances algorithms andrew annotated anton appear april august based beatrice bikel bilexical black broad building bunt california cation center charniak christopher classi coling collins comparing computational computer conditional conference copenhagen corpus cover coverage cubictime daniel darpa data david dependency design development diego discriminative distributional divergence document driven editors eisner elds elements empirical engine engineering english entropy erty estimation eugene evaluation exploration ezra fernando flickenger frederick friedman gdaniec gildea grammar grammars grishman grove harriman harrison harry head hebrew hindle history icml ieee information ingria inspired international intricacies israel jason jelinek jerusalem jianhua john kaufmann klavens klein kluwer labeling language large learning leibniz liberman lillian lingual linguistics machine magerman manning manuals marcinkiewicz marcus mary maximization maximum mccallum measures mercer methods michael mitchell model models morgan multi naacl naftali natural nijholt noam october other paci pages parallel parser parsing pattern penn pennsylvania pereira performance philadelphia pittsburgh probabilistic procedure proceedings processing publishers quantitatively random recognition references report reranking richer robert roukos salim santorini school science seattle segmenting sequence sequential shannon similarity slonim sons speech statistical structure strzalkowski syntactic technical technologies their theory thesis thomas three tishby towards transactions treebank university unsupervised using variation versus washington wiley workshop york
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Clark.pdf	21	Object-Extraction and Question-Parsing using CCG	accurate algorithm almost amit annotation answering antecedents approach aravind argument bangalore barcelona briscoe building burning cambridge canaria carroll categorial charniak clark coling collins combinatory combining computational conference coverage curran data deep dependency development dienes doran driven dubey edinburgh emnlp empty entropy envgram eugene forest from general generative geneva grammar gran harabagiu head high hockenmaier hockey hopely importance information inspired interpretation james japan johan john johnson joshi julia language linear linguistics lrec madrid maintaining marius mark matching maximum meeting methods michael models naacl nasr natural nodes orleans pages palmas parser parsing pasca pattern pennsylvania performance peter philadelphia predicate press proceedings process processing question rambow recovering references representations research retrieval robust rosenzweig sanda sapporo sarkar seattle semantic shallow sigir simple spain srinivas statistical steedman stephen structure structures supertagging surface switzerland syntactic text their thesis underbrush university using wide widecoverage with workshop xtag
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Mihalcea.pdf	58	TextRank: Bringing Order into Texts	algorithms anatomy annual application applied arti association august authoritative automatic barcelona based brin canada cial cohesion coling companion computational computer conference council digraphs disambiguation document domain edmonton empirical engine english environment evaluation extract extraction figa frank from geneva given gram graph gutwin halliday hasan herings hobbs hovy http hulth human hyperlinked hypertextual improved information institute intelligence international isdn japan joint journal keyphrase keyphrases keyword kleinberg knowledge laan language large learning linguistic linguistics lingusitics longman manning measuring meeting methods mihalcea model more naacl national natural networks nevill nist nlpir nodes occurrence page pagerank part paynter power proceedings processing projects ranking references report research scale search semantic semantics sense sentence sources spain speci statistics summaries summarization switzerland systems talman tarau technical technology text tinbergen turney understanding university using volume with witten word yale
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Goweder.pdf	38	Identifying Broken Plurals in Unvowelised Arabic Text	abuleil addison analysis approaches arabic baeza beesley berthier cairo coling comparing computational computers discovering egypt evens finite generation hani index individual information jasist kharashi khat language languages lexical mahmoud martha methodologies modern morphological neto newspaper omari over press proceedings publishing query reep references retrieval ribeiro ricardo roots saleem salem semitic state stemming stems system tagging terms text wesley words workshop yates
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Tsukada.pdf	61	Efficient Decoding for Statistical Machine Translation with a Fully Expanded WFST Model	according accuracy achieve adam added alex algorithm alignment ambiguity american amta andrew annual apparatus applied apply approach approximately association automata automatic average bangalore based beam beforehand beginning berger bleu blue brown byrne cambridge cause chapter christoph church cient clarkson coding communication comparison compile complexity composed composition computational computer conclusions conference consisted context conventional cooccurrence correspondences corresponds course daniel darpa data dddh decoding degrading della described devices difference different doddington driven dynamic each editors effectiveness eiichiro emmanuel english estimation european eurospeech evaluation example expanded expansion experiment experimental express fast faster fernando fertility figure finite fourth franz full fullexpansion fully gale general generally germann giueseppe gram hermann hmms however human hypotheses identifying implementation including independent itself jahr japanese jing josef july june kehler kenji kenneth kevin kishore knight kumar language large length like linguistics machine march marcu massachusetts mathematics measure meeting mehryar mercer merged merging method michael model modeling models mohri more naacl natural nicola nist nite north null onaizan optinal original pages paper papineni parallel parameter patent paths pereira peter pietra pitra press previous proc processing produced programming proposed provided pushing quality reach recognition reduce reduced reduction references reordering replacement represented result results riccardi riley robert roche rosenfeld rough roukos salim schabes score scores search section sentence sentences september shankar showed shows size slight slightly some special speech srinivas state states static statistical statistics stephen submodel submodels sumita summit symbols system systematic table taro technology template test texts than that these this tillmann time times todd toolkit transducer transducers transitions translation translations trimmed ulrich united used using various vincent vocabulary waibel wang ward watanabe weight weighted were wfst wfsts which while william with without word words workshop yamada yaser yves
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Feng.pdf	54	A New Approach for English-Chinese Named Entity Alignment	acquisition aligned alignment also among ankara annals applications approach approaches assistant author automatic based berger bilingual brew brown budapest cambridge canada cannot cascaded chang changning checking chen cherry chinese choi chunking class coling come comments comparison computational conference continuous corpora cost darroch data dekker della dictionary discriminative driven eacl ease edmonton eduard encouragement english entities entity entropy equivalence estimation extracting extraction fastus finitestate first france from generalized graehl handbook hang heidorn hmmbased hobbs hongkong hovy huang hungary improve improved information intelligent international iterative japan john kevin knight koehn korean language learning lexicography linear linguistic linguistics machine marcel marcu mathematical mathematics maxent maximum mckelvie melamed memory mercer methods minimization mixed model models monolingual moore msra multifeature multilingual naacl named natural novel number onaizan pair pairs paper parallel parameter philadelphia phrase phrases pietra press probability probst processing propose ratcliff references resources results sapporo satisfactory scaling scoring similarity special statistical statistics structural support systematic techniques text texts thank thanks their this tillmann toulouse towards traditional training transducer translating translation translations translingual transliterated transliteration unified using valuable various visiting vogel volume waibel wang while with word words work workshop writing yunbo zhou
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Xu.pdf	48	Random Forests in Language Modeling	acoustics advances algorithms amit analysis andreas annual applications backing bahl based bengio berkeley bigram breiman brian brown california cambridge cation chapman chapter charniak chelba chen chou ciprian classi clustering communication computation computer conf conference connectionist constructing continuous decision denver department dependencies ducharme empirical eugene european extensible florida forests france frederick friedman gauvain geman gerasimos goodman gram group hall harvard head hermann holger icassp ieee immediate improved information intelligence intl jean jelinek joshua july kneser language large letter liermann machine martin massachusetts meeting mercer method methods model modeling models motivations natural neural october olshen optimal orlando pages parsing partitioning pascal pattern peng pennsylvania philadelphia potamianos predictive press probabilistic proc proceedings processing providence quantization random randomized recognition references regression reinhard rejean report richer roark robust schwenk science shape signal smoothing souza speech spoken srilm stanley statistical statistics stolcke stone structured study subspace syntactic systems technical techniques thesis toolkit toulouse trans transactions tree trees trigram university vincent vocabulary volume with word york yoshua
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Kudo2.pdf	45	A Boosting Algorithm for Classification of Semi-Structured Text	aaai adaboost adjectives algorithm algorithms algoritms application applied arching arikawa arimura asai automated available bact bartlett based bernhard boostexter boosting boser breiman categorization cation chasen ciently classi collins colt composition computation computer computing convolution corpora cruz data david decision decomposition discovery discrete down duffy effectiveness explanation fabrizio forest frequent freund from games generalization gunnar guyon haussler hiroki hisashi http hypotheses iaai icml implementation isabelle janyce japanese journal kashima kawasoe keiji kenji kernels klaus koyanagi language learning line machine margin margins methods michael mining mohammed morhishita muller naoyoshi natural neural nigel onoda optimal optimized orientation over pages parsing perceptron peter pkdd prediction proc processing progress ranking ratsch references report reviews robert santa schapire science sebastiani segments semantic semi setsuo shinichi shinji sicences sigkdd singer soft software springer structured structures structuring subjective substructure surveys system tagging takashi taku tamura tatsuya technical teruo text theoretic thumbs training trees turney unsupervised vapnik vladimir voted voting wada wiebe yoav yoram zaki
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Zhao.pdf	33	Phrase Pair Rescoring with Term Weightings for Statistical Machine Translation	adaptative alex algorithm aligning alignment based beijing better bilingual bing bleu brown canada china church city collection collins computational conf conference confusion corpora daniel data della denver discriminative edmonton ellen emnlp estimation extensible fifth franz from gale have hermann hongkong ieee improved improvement integrated international interpreting intl japan joint josef kantor kenneth koehn language learning linguistics lisbon lrec machine maebashi marcu mathematics mercer michael mining model modeling models much naacl natural need news nist parallel parameter parsing paul peter philadelphia philipp phrase phrasebased pietra portugal probability proc proceedings processing program references report reranking retrieval robert scores segmentation sentences spoken srilm statistical stephan stephen stolcke system text toolkit track translation trec vincent vogel voorhees waibel william wong ying zhang zhao
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Szpektor.pdf	12	Scaling Web-based Acquisition of Entailment Relations	aaai abdessamad acquisition alignment answering applicability applied approach arti articles automatic automatically barzilay based bootstrapping buckland canada case ceedings christian cial coling collin conference corpus dagan daniel deepak dekang dependency dictionaries diego discovery duclaye echihabi editors edmonton eduard ellen engineering entailment evaluation exploitation extracting extraction florence form france francois from gaithersburg generating generic glickman granada grenoble grishman hermjakob hovy human huttunen identifying improved inference information intelligence jacquemin jones kathleen kevin kiyoshi knight language learning level lexical lillian linguistic logic lori lrec marcu mckeown methods mining minipar model modeling moldovan multilevel multiple naacl national natural news nist olivier oren pages palmas pang pantel paradigmatic parallel paraphrase paraphrases parsing pascal pasi patrick pattern patterns philadelphia probabilistic proceedings question ralph ranlp ravichandran references reformulation reformulations regina representation representations resource retrieval riloff roman rosie rules satoshi scenario sekine sentences sequence shinyama silja single sixteenth spain study sudo surface syntagmatic syntax system systems tapanainen technology term text textual toulose transformation translations trec understanding unsupervised using variability variation vasile verbs voorhees wordnet workshop yangarber yusuke yvon
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Cohen.pdf	46	Learning to Classify Email into "Speech Acts"	aaai aalborg access action activities acts adaptive advances agreement algorithm algorithms alto amherst annual anti appear application applying approach approaches argamon assessing assistant association attempts authorship automatically avoids bayesian behavior bell bellotti bled blei boosting boston bruce bulletin burden california camino canada capture carletta carvalho categorization centered central challenge challenges chel chinese clarity classification classifies classify cohen combine communication computational computer computing concluding conf conference confidence construction cooperative coordinating coordination corpus creech data decision deliberation denmark dept design designed desktop development dialog dialogue discourse discrimination doing ducheneaut dumais eighth electronic email emnlp empirical engineering entropy espinosa evaluation evaluative evidence experimental extract extraction factors feature fick filtering finke first flexibility flores florida freitag freund from fussell game goldkuhl griffiths group gunderson hainsworth have heckerman heuristics hierarchical horvitz howard http human humancomputer identifying ifile illocutionary important improved inducing inferring information integration intelligent intention interaction interactions interfaces international introduction isaacs issue issues joachims joint jones jordan journal junk kappa kephart knowledge kopenhagen kraut language languageaction lapata large lauderdale lavie learning lerch levin lewis lines linguistics ludlow machine machines madison mail management margin markov martin mass matwin maximum mayfieldtomokiyo mccallum messages messaging methods milewski millen millewski mind mining minneapolis minnesota minorthird model modelling models moor mountain multiple murakoshi names nardi nashville natural negotiating nested netherlands networks neural numbers observes ochimizu office ontological ontology organizational organize organizing pacific palo pang parse passively perceptron pereira pergamon personal perspective polzin predictions presented press previous proc proceedings process processing properties propose rated ratnaparkhi reduces references regularities relations remarks rennie reply replying representation research responses restaurant results retrieval review ries rules sacrificing sahami schapire schoop science scott searle segal segmentation sentiment shimazu sigdial siggroup sigir sigkdd signature simulated singer sium slovenia smith social some sourceforge spam speaking speculative speech spraque spring stanford statistic statistical stein structure structured studies study style support supporting swiftfile sympo symposium system systems taking task tasks taxonomy teams techniques tenenbaum terveen text that theory thesis this thumbs tilburg tool topic transactional transactions under univ university unlike user users using vaithyanathan vector view waibel washington waterloo wermter which whittaker wiebe wiegend wilson winograd with work working workshop yang york zechner
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Teufel.pdf	60	Evaluating information content by factoid analysis: human annotation and stability	aaai abstracts advances american analysis automatic barzilay benjamins between bleu cambridge conference consensus content directions document documentation eacl ectiveness electronic elhadad evaluating evaluation examining experiments factoid factors firmin formation halteren hirschman house hovy html http human information initial intel international jing john jones judgements klein know ledge ligent machine management mani manual maybury mckeown measurement method methods naacl nenkova nist nlpir notes papineni passonneau press proceedings processing projects pubs pyramid radev rath references relative relevance resnick retrieval roukos savage selection sentences sparck spring summac summaries summarising summarization sundheim symposium teufel text tipster translation twelfth understanding using utility variations voorhees ward with working workshop
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Freitag.pdf	40	Trained Named Entity Recognition Using Distributional Clusters	aaai adaboost algorithms anlp annual aone applied april arti august austin based bennett bikel boosted boosting bootstrapping brown carpuat carreras categories cation cial clark class classi clustering clusters collins colt combining computational computer conditional conference conll context contexts contextual corpora cucerzan della dence dept desouza dhillon discriminative distribution distributional eacl early elds emnlp empirical enhanced entity evidence experiments extraction feature freitag gram guinness hidden highperformance improved independent inducing induction information intelligence introduction joint july kushmerick language languageindependent large larsen learning lexicons linguistics mallela march markov marquez mccallum mercer method methods miller models modha morphological multilingual naacl name named national natural nder ngai nymble observation padro pages part pattern perceptron pietra predictions proc proceedings processing random rated recognition references report results riloff sang schapire schutze schwartz science semantic september shared sigdat singer speech syntactic tagging taipei taiwan task technical texas texts thelen theoretic theory through tjong training unsupervised using very weischedel with word wrapper yang yarowsky zamanian
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Smith.pdf	13	Bilingual Parsing with Factored Estimation: Using English to Parse Korean	across aligned alignment alshawi annotation approach asian automaton baker bangalore based bilexical bilingual bootstrapping bracketers brown cambridge canon catalog catalogentry catalogid cationbased charniak cherry chiang chinese cient clark cocke cohesion collections comp companion comparing comparison computational conditional conf context corpora correspondence crim crouch data declarative della dependency description development discriminative divergences dorr douglas dyna dynamic eisner elds emnlp empirical english entropy estimating estimation estimators evaluating evaluation exact example extracted factored fast finite formal free from functional geman gildea goldlust graehl grammar grammars head hockenmaier http icml implementing improve inducing inversion ircs isomorphic jelinek johnson joshi kaplan king klein klex knight kolak korean labeling lafferty language learning lexical lexicalized linguistics loosely ltag machine manning mappings mathematics maximum maxwell mccallum melamed mercer methods mildly model models morphological multilingual multitext naacl natural ngai nips nite osborne paci pages palmer parallel parameter parsers parsing penn pereira phrasal pietra press probabilistic probability proc processing programs projection proposed random recognition references report resnik riezler robust roossin ruhlen sarkar satta schabes segmenting selection sensitive sequence shieber smith solution speech state statistical steedman stochastic summit synchronous syntax syntaxbased systematic taggers tagging technical techniques transducer transducers transduction translation translational transliteration tree treeadjoining treebank upenn using various wave weinberg with word workshop yamada yarowsky
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Taskar.pdf	7	Max-Margin Parsing	accuracy accurate altun analytic annual association based bunt cambridge canon carroll cation clark collins computational conditional conference cristianini crouch curran cyclic data deep developments discriminative distribution driven dynamic editors eech elds eling emnlp endency entropy erty estimation estimators feature features forests free geman giorgio grammars guestrin harry head hidden hofmann human icml introduction john johnson joint joshi kaplan kernel king klein kluwer koller language learning linear linguistics ltag machines manning margin markov maximum maxwell mccallum meeting methods miyao models naacl natural network networks nips other pages parameter parse parsing part pennsylvania pereira platt practice press probabilistic proc proceedings programming random references reranking rich riezler sarkar satta segmenting sequence shallow shawe shen singer sparseness statistical stochastic supp support tagging taskar taylor technology theory thesis toutanova training tsochantaridis tsujii university unlexicalized using vasserman vector with
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Mullen.pdf	59	Sentiment analysis using support vector machines with diverse information sources	aaai academic adjective adjectives analysis annotating annual applied arti articles association attitude austin based brunswick burges cambridge capture capturing cation chapter charles chichester church cial classi classify collier computational conference corpora cristianini criticism data deep dependency discovery down effects empirical engineering european favorability florida framework from george global gradability hanks hatzivassiloglou illinois india inference information instructions integrating intelligence intelligent international introduction jaap janyce jarvinen jeonghee joachims july kamps kawazoe kernel kluwer knowledge language large learning lexicography lillian linguistics littman maarten machine machines marten marx mckeown meaning measurement measuring meeting methods mining mokken mullen mutual mysore nasukawa national natural newspaper norms opinions orientation osgood other pang parser pattern percy philadelphia pittsburgh praise predicting press proc proceedings processing projective publishers recognition references report reviews rijke robert second semantic sentence sentiment seventh shallow shawe shivakumar springer statistical structures subjective subjectivity succi support systems takeuchi tannenbaum tapanainen taylor technical techniques ternational tetsuya texas text theory thorsten thumbs tois transactions turney tutorial university unsupervised using vaithyanathan vapnik vector verlag very vladimir washington wattarujeekrit wiebe wiley with word wordnet words
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Erkan.pdf	53	LexPageRank: Prestige in Multi-Document Text Summarization	anlp april automatic based baseline bigram blair brandow brin bringing budzikowska canada carbonell centroid chin cial citation code codes common condensation conference cross dence development dialogue discourse diversity document documents dragomir edmonton electronic evaluation experiments extraction first from fusion goldensohn goldstein gram hong hongyan hovy information interval jade jaime jing june karl kong language lisa malgorzata management manual mead mitze motwani multidocument multiple naacl occurrence october order orleans page pagerank pages peer proceedings processing producing publications radev ranking references reordering report reranking research retrieval rouge sasha scores seattle selection sentence september sigdial single sources stanford step structure studies submissions summaries summarization summary system table task technical technology text theory understanding unigram university user using utilitybased winograd workshop zhang
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Liu.pdf	15	Comparing and Combining Generative and Posterior Probability Models: Some Advances in Sentence Boundary Detection in Speech	adaptive affordable agenda annotation appear applied approach assp automatic based berger boundaries boundary brants broadcast brown carnegie cation challenges chen christensen class classes combined communication computational conditional consortium conversational cope darpa data della desouza detection disambiguation ears effective emnlp entropy estimation eurospeech evidence fall fast fifth florian fourth franco from gadde gaussian generation gotoh gram harper hearst hidden http icml icsi icslp identifying ieee imbalanced index information institute internal into introduction ipto isca january juang june klein labeling lafferty lane language learning linguistic linguistics machine magazine manning markov maximum mccallum mellon mercer metadata millennium models naacl national natural ngai nist november pages palmer part pereira period pietra presentations prior probabilistic proc processing prof programs prosody punctuation rabiner random ratnaparkhi recognition references renals report research reusable reynar rosenfeld schmid segmentation segmenting sentence sentences sequence shriberg simple sixth smoothing speci speech spring standards statistical stolcke strassel structure stuttgart system tagger technical technology tests text tokenisation topics transcripts transformationbased uency understanding university unsupervised using versus with woodland workshop
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Mochihashi.pdf	50	Learning Nonstructural Distance Metric by Minimum Cluster Distortions	acquisition acyclic addison adjugate advances algorithms alternatively american analysis andrew annual answers appendix application association automatic bach baeza based bases berry berthier blake calculated categorization cation cell choi christopher christos classi clustering collins come comparisons computational concept condition conf conference constraint convolution corpus data databases david decompose decompositions deerwester denotes derivation deriving dessert dharmendra dhillon diag differentiating dimensionality dimensions directed discriminative distance documentation domain duda duffy dumais ecml edcb edition eisaku eric euclidean even example examples expanding exploiting expository faloutsos features figure filtering formed foundations francis freddy from fruit function furnas generative george graph hart haussler hearst here hierarchical high hinrich hirao html http hvus ieee ihug independent inderjit indexing information international introducing introduction inverse irregular ishikawa jaakkola jiang joachims john jordan journal kernel kernels kikui lagrange lang language large latent learning least length like linear linguistics lrec lter machine machines maeda manning many marti mathworld matlab matrix matthew mean meeting merz method methods metric michael mika mindreader minimize mlearn mlrepository models modern modha moore moorepenrosematrixinverse much muller multi multiple multiplier naacl namely natural netnews neto networks neural newsweeder nigel nips normal number obtain orthonormal pages paragraph pattern penrose peter pinv press proc proceedings processing property proposal prove pseudoinverse query querying rank rate ratsch ravishankar reduction references registration relative relevant repository retrieval ribeiro ricardo richard riemannian right rqug rsvd russell salton sasaki satis schultz schutze science second segmentation semantic sentence setting shortest sideinformation simply singular society solution some sons sparse speci spectral squares statistical stork structured stuart subramanya sugaya supercomputing support susan suzuki synonymous systems takezawa term text that then theorem therefore thorsten through tommi total tsuda tsutomu twelfth under unique using values vector very volume weisstein wesley when where wiley with wolfram wppy xing yamamoto yang yates yoshiharu yutaka zero
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Almuhareb.pdf	27	Attribute-Based and Value-Based Clustering: An Evaluation	abattoir abdulaziz abdulrahman academic according acne acquiring acquisition adjectival adjectives aircraft airplane alan almuhareb american anaphora anger animal ankle anthrax appendix apple approach arabia arbitrary architect arthritis artist associative asthma atlas attributes automated automatic automobile available banana base based bear behavior beijing berland berry bicycle bigrams blouse boat bobrow body book bookcase booklet books brachman bradford brochure budapest builder building bull burgess cabinet california cambridge camel cancer caraballo catalog center century chair charniak cherry child cholera cirrhosis city clarendon class classes cloth clubhouse clustering cluto coat cognitive coling collins columbus community comparing computational computers concepts conceptual conference construction constructor cooccurrence cook cookbook core corpora corpus costume couch cousin cradle craftsman creator criteria cruiser curran daniel data database daughter decade deer descriptions designer desire desk details developer diabetes dictionary discovery disease distributional dixon documentation dormitory dresser eacl eczema editor editors effectiveness electronic elephant empirical encyclopedia engineering english evening experiment extracting extraction face fall family farmer father fear feature feeling fellbaum finding finger fisher foot foundations frequencies from fruit furniture generative george girl glaucoma gloves google grammar grandchild grandfather grandmother grape graph greenhouse grefenstette guarino hall hand handbook happiness hatzivassiloglou head hearst helicopter hepatitis heuristics hierarchy highdimensional holder horse hospital hotel hour house http husband hypernym identification ieee implementation improvements incremental information instrumentation intelligent international introduction inventor ishikawa issue jacket jeans journal june kacst karypis kaufmann keller kilgarriff king kitten kiwi knowledge labeled lamp language lapata large laurence learning lemon leukemia levesque lexical lexicon library linguistic linguistics link lion lounge love lrec lund machine magazine maker making malnutrition mango manning manual manufacture margolis markert maryland mckeown meaning melon meningitis methods mining minnesota modjeska moens monkey month morgan morning mother motorcycle mouse musician natural neckpiece networks night nissim nlpke nominal nose noun nursery obtain offspring ohio olive ontological orange originator overtime oxford oyster pages pain painter pajamas palmas pants part parts passion peach pear pereira perspective phonebook photographer pickup pineapple plague pleasure poesio press principles proc processing producer producing property publication puppy pustejovsky quarter reading readings reference references relation relations report representation research resolution restaurant retrieval rheumatism riyadh robe rocket sadness salience saudi scales scarf school schuetze science season seat semantic semantics semester sensitivity sextant shame sheep ship shirt shoulder sibling similar skyscraper smallpox sofa some sources spaces special spring statistical strawberry structuring studies submitted suit summer supported swets systems table tailor tavern technical technology text textbook thank theater their thesaurus thesauruses tiger time tishby tongue toolkit tooth towards treatments trousers truck turtle understanding uniform university unseen unsupervised used using vehicle very vieira villa walde want wardrobe watermelon week weekend what whorehouse wife winter wonder woods wordnet words workbook workshop wrist wwwusers year york zebra
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Daume2.pdf	39	NP Bracketing by Maximum Entropy Tagging and SVM Reranking	adam algorithm alsmost american annual approach april aravind association bangalore based bergen berger brandts brill cascaded case chapter charniak chunking clickthrough collins compositional computational conditional conference conll corpora dagan data december della detection discovery discriminative driven eacl elds engines entropy entropyinspired eric erik errordriven eugene evidence fernando first franz head hermann hltnaacl hong incorporating joachims josef joshi journal july knowledge kong krymolowski kudo lance language large learning linguistics machine machines march marcus markov matsumoto maximum meeting memory michael michell mining models naacl natural north norway noun optimizing pages parser parsing part partial pereira philadelphia phrase pietra porter proceedings processing program ramshaw random references repeated research sang search seattle shallow speech srinivas statistical stephen stripping study supertagging support tagging taku text third thorsten tjong training transformation translation using vector very vincent washington with workshop yuji yuval
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Ney.pdf	41	Error Measures and Bayes Decision Rules Revisited with Applications to POS Tagging	about abstract academic acoustics agreements algorithm allocate already analysis approach assignment automatic bahl based basically bayesrisk beale beijing berger bisani bloothooft boca bootstrap byrne canada carstensen category cation chapter china chou church classi cocke codes coding colloquium communications computational computer conf corpora corpus countries decision decoding della dence denke derose direct disambiguation dordrecht duda editors emotionen empirical engineering english entropy error errors estimates european evaluation examples expansion florida gefallt geneva german geschichte glasgow glauben goel gram grammatical handout hart have idee ieee information internal intervals investors jelinek john juang kanalisieren kinscher klute kluwer knowledge language large lead leaving leute linear linguistic linguistics made martin maximum mercer merialdo methods minimizing minimum model modelling montreal much munster nach named natural negotiations northern noun oneout optimal optimization pages parser part parts pattern performance phrase pietra plant police prasentiert press probabilistic proc processing program project publishers rate ratnaparkhi raviv recent recognition reference references report ronneby rota rule sandmannchen scaled scene scotland sharply signal sommerset sons speech started statistical steiner stochastic stolz string stronger sundermann sweden switzerland symbol symposium synther table tagger tagging tannenbaum text that theory think this trans university unrestricted very warum wessel wiley will with written year years york young
http://acl.ldc.upenn.edu/acl2004/emnlp/pdf/Pado.pdf	20	The Influence of Argument Structure on Semantic Role Assignment	adding algorithms alternations analysis analytic andrea anette anlp annotated annotation antal assignment automatic available baker baldewein barcelona based berkeley beth bigrams book bosch brants brno building cambridge canada cation cavtat charles chicago christopher city classes classi cluster clustering coling college collin collins comparison computational conll corpora corpus corpusinduced croatia cynthia czech daelemans daniel data decision dependency diego distributional downloads ecml ellsworth emnlp english entropy estimation extensive fillmore finding fleischman frame framenet frames frank frequencies from generalisation generalisations generative german gildea grammar groups gruber guide hajicov helmut hong hovy html http icsi interpretation introduction investigation jackendoff jakub japan jeffrey john johnson jurafsky katrin kaufman keller kingsbury kong kowalski kwon labeling labelling lapata large learner levin levy lexical lexicalised lillian linguistics lissabon lowe lrec madrid malouf manchester manfred manning marcus martha maximum measures memory michael mirella mitch model models montreal namhee nemlap obtain over pado pages palmer papers parameter park parsing part paul penn petruck pinkal portugal practice prague preliminary prescher press probabilistic proc proceedings project quaderni reference references relations report republic resource resources roger role roles rousseeuw rules ruppenhofer sapporo schmid seattle sebastian semantic semantica semantically semantics senseval similarity sloot sons spain speech statistical studies tagger tagging taipei taiwan technical tectogrammatical theory thompson thorsten three tilburg timbl towards treebank trees understanding university unseen using verb version walter wiley with wood working workshop york zavrel