http://acl.ldc.upenn.edu/N/N07/
NAACL 2007
http://acl.ldc.upenn.edu/N/N07/N07-4014.pdf	154	The Hidden Information State Dialogue Manager: A Real-World POMDP-Based System	academic achieved agenda approach aruba average based beach bootstrapping bridging cambridge completion computer correct decision demo dialog dialogue dialogues error evaluated forthcoming hawaii hidden honolulu human icassp ieee industrial information language management markov measured naacl number observable palm partially performance pomdp pomdps proc proceedings processes rate real recent recommendation recorded references research rochester schatzmann simulation speech spoken state study subjects system systems task technology thesis thomson total training turns university user using venue weilhammer were williams with word workshop world young
http://acl.ldc.upenn.edu/N/N07/N07-2035.pdf	108	Analysis and System Combination of Phrase- and N -gram-based Statistical Machine Translation Systems	aachen alignment analysis april avoiding banchs based best bleu boston ciency combination comparison computational computer computing conf consensus corresponds costa crego david decoder discussion each eacl english enhanced error experiments extensible fernando fonollosa from function genoa gispert giza gram hand haro here hermann html http hypotheses icslp improvements include informatik italy iwslt journal jussa khalilov lambert language lation linguistics lrec luis machine marino matusov mead merged method minimization modeling multiple mwer nelder ngram nist october outperforms output overpopulation pages pert phrase phrasebased presented problem proc processing produced ranked references rescored rescoring results richard rows rwth score search sentences september shown shows simplex simply software spanish spoken srilm statistical stolcke structural system systems table talp terms test that third time toolkit trans translation using versus vilar wwwi zens
http://acl.ldc.upenn.edu/N/N07/N07-3005.pdf	134	Analysis of Summarization Evaluation Experiments	abstracting acceptabilit advances association automated automatic automatique barzilay brandow cambridge chains computing condensation construction editors edmundson electronic elhadad farzindar ferret filtrage fluhr guid hovy information journal juridiques langues lapalme lexical machinery management mani massachusetts maybury methods mitze multidocuments pages pour press processing production publications qualit references selection sentence summarist summarization taln text textes traitement using valuation
http://acl.ldc.upenn.edu/N/N07/N07-1053.pdf	53	A Cascaded Machine Learning Approach to Interpreting Temporal Expressions	about adafre adverbials analysis ando annotating annotation approach artemov automatic available barco bethard boguraev cation chang charniak chen chinese cicling cjlin class comparative compliant computer corpus csie dale dates domain douglas editor editors emnlp empirical english entropy essays event events exploration expression expressions extracting ferro fissaha from gabbay gaizauskas gelbukh gerber hacioglu hara hitzeman honour http identi ijcai information inspired interchange interface interpretation interpreting irst kubota labeling language lecture library libsvm local lrec machines mani marseglia mart martin maximum mazur mckeever meaning mentions mitre munoz naacl natural negri news normalization notes nouns ohrstrom open oxford pages parser prepositions press proc proceedings processing pustejovsky reader reasoning recognition recogniz recognizing reference references report representation resolution rijke robust rochester sandgren sanitized saquete schilder science semantic semantics show software spanish standards support syntaxsemantics tagging talip technical temporal tern text texts their them thesis tides time timeml times timex trento university vector volume wiebe will wilson workshop zwol
http://acl.ldc.upenn.edu/N/N07/N07-1063.pdf	63	An Efficient Two-Pass Approach to Synchronous-CFG Driven Statistical MT	accurate algorithm appear assocation based best better campaign chiang chiori cient communications computational computing context daniel david evaluation franz free galley hierarchical hillel hopkins hori huang inferences international iwslt josef journal kevin klein knight koehn language liang linguistics machine machinery manning marcu matthias michael model models naacl overview pages parsing perles philipp phrase proc references rich scalable shamir spoken statistical syntax technologies training translation unlexicalized workshop
http://acl.ldc.upenn.edu/N/N07/N07-2045.pdf	118	Language Modeling for Determiner Selection	adaptation american annual antal association barbara based bond bosch chance charniak church closer coling computational conference consortium corpus countability daelemans data david empirical english estimates eugene francis gawronska graff great guide head immediate international jakub japanese kenneth kentaro language learner linguistic linguistics machine meeting memory models netherlands news noriegas north number ogura pages parsing problem proceedings reference references report sloot technical text than tilburg timbl translation university version walter zavrel
http://acl.ldc.upenn.edu/N/N07/N07-1038.pdf	38	Multiple Aspect Ranking using the Good Grief Algorithm	ability agreement algorithm analy analyzing answering applied approach aren aspects assigned based basilico between both cant categorization cation class classi classsi coling collaborative component conclusion considered content contrast counts crammer current dave dence depend dependencies dialogue different discourse down echihabi emnlp empirical employs exploiting extraction facts from future gallery generate goldberg graph guide hatzivassiloglou higashinaka hofmann icml identifying improvements individual input joint jointly lawrence learning learns lies ltering machine many marcu method mining model modeling models multiple naacl naturalistic nips opinion opinions orientation over pang parameter particular peanut pennock polarity pranking prasad prediction presented problem proceedings product questions rankers ranking ranks rating recognizing references related relation relations relationships relatively respect reviews rhetorical scales seeing semantic semi sentences sentiment separating signi singer single spoken stars state strength such supervised systems techniques textgraphs there thumbs towards tuning turney unifying unsupervised user using utterances vaithyanathan value walker well what when will with work workshop yields
http://acl.ldc.upenn.edu/N/N07/N07-4008.pdf	148	RavenCalendar: A Multimo dal Dialog System for Managing a Personal Calendar∗	access accommodate adaptable additional agenda already application applications approach assistant automatically based bohus bridging building calendar cation clari comp computers computing conference connect constantinides control conversational cscer current currently data decomposition deeplistener dialog dialogue eech ermit eurospeech event events expectation expected factors framework future generation grammar guide harnessing heidorn hierarchical horvitz huang human icassp icslp input integration interest interesting interface interfaces interspeech language larri last learning line list maintenance management managing many mipad mobile modify modifying multimodal music naacl natural next noti olympus onent onse open orts otential other paek parsing particularly proceedings prototype public raux ravencalendar ravenclaw real recognition references remote repair research resources resp rudnicky schema searches services several source speech spoken system systems taking talking task that there this tighter topic trainable urls user users uses using utility words work workshop world yankelovich
http://acl.ldc.upenn.edu/N/N07/N07-2039.pdf	112	Reversible Sound-to-letter/Letter-to-sound Modeling based on Syllable Structure	aalborg accuracy acquisition adamson allen analogy approach australia automatic based black blackwell blevins canada caves chung cient comparative computation computational conversion damper denmark denver dialogue directional edmonton english eurospeech evaluation font galescu generation goldsmith gram grapheme gustafson handbook icslp improves improving iwss jenolan joint knowledge kokkinakis language letter linguistics llitjos marchand meng mode model morphology multi multilingual naacl name names origin oxford pages parsing phoneme phonological proc pronunciation proper references rentzepopoulos reversible seneff sound speak speech spell spoken strategy syllable synthesis systems techniques text theory using wang with word
http://acl.ldc.upenn.edu/N/N07/N07-1016.pdf	16	Unsupervised Resolution of Objects and Relations on the Web	banko bhattacharya broadhead bureau cafarella census clustering coling current data domingos entity etzioni evidence extraction from getoor heterogenous icdm ijcai induction information jurafsky linkage logic markov mining multi open problems query record references relational report research resolution semantic sigkdd singla snow soderland state taxonomy technical time type washington winkler with workshop
http://acl.ldc.upenn.edu/N/N07/N07-4003.pdf	143	Adaptive Tutorial Dialogue Systems Using Deep NLP Techniques	aaai adapting adding aied albacete aleven anderson barr chiu cognitive conference control corbett dialog dialogue editors eliciting engine engineering explanation explanations flairs improves information jordan journal koedinger language larsson lavancher learned learning leeuw lessons makatchev management markov models move natural pages pappuswamy pelletier physics popescu press proc redundancy references science sciences scripted self state student support system taking toolkit towards traum trindi tutor tutorial tutors understanding using vanlehn
http://acl.ldc.upenn.edu/N/N07/N07-4016.pdf	156	Voice-Rate: A Dialog System for Consumer Ratings	adaptive amazon application applications architecture assistance assistant automated automatic automation behind bilenko chai closely cohen collaborative comparison computer computing conference confusable construction corston customer design dialog directory distance especially eurospeech fienberg filtering first florida free from gamon horvath however icassp ieee ijcai information integration intelligent interactive interface international internet issues item kambhatla kamm knowledge language large lecture levin linden magazine maison makhoul mangu matching melville metrics mining mooney name natarajan natural nicolov notes oliver online opinions orlando pages prasad previous proceedings product pulse rate ratings ravikumar recognition recommendation recommendations references related reviews ringger scalable scale schwartz science second sets shamieh sigkdd signatures singhal smith speech springer string stys summarizing system systems tasks technology telecommunications text unique user verlag voice work workshop yang york zadrozny zweig
http://acl.ldc.upenn.edu/N/N07/N07-1004.pdf	4	What Decisions Have You Made: Automatic Decision Detection in Conversational Speech	acoustics agreement algorithms analysis andrew annotations announcement annual ashby association automated banerjee bayesian bene bengio between boulis bourban browsing calo capture carletta cieri collection computational computer conf conference conversation conversational corpus data dependencies detecting differences disagreement extraction flynn from galley ganesan garofolo gatica genders gordon group guillemot hain hirschberg http human icassp identifying ieee interaction interest international joint kadlec karaiskos kavita knowledge kraaij kronenthal language laprun lathoud learning level lexical lincoln linguistics lisowska lrec machine mccowan mckeown meeting meetings methodologies michel miller model multimodal necessity networks nist observations ostendorf outcomes perez pilot playback post pragmatic press proceedings processing project quantitative recording references reidsma related research room rose rudnicky shriberg signal speech stanford story system tabassi technologies telephone tenth third topic walker wellner workshop zhang
http://acl.ldc.upenn.edu/N/N07/N07-4006.pdf	146	Cedit – Semantic Networks Manual Annotation Tool	allows annotation application applications aspects association automatically barbora been before berlin beyond boston brian carsten cation cole company computational corpus databases david deep dependency dodrecht eckstein edit edited editor elliott enabled engineering even example format frontiers functionality generic german germany gnrlich hagen hajic hajicova have heidelberg helbig hermann hladka information intended interface international into ircs james jarmila java knowledge language large linguistic linguistics london manual many marc massachusetts meaning meyers more multinet naacl natural netherlands network never novak pages pajas panevova pennsylvania petr philadelphia pragmatic prague praha primarily proceedings publishing references reidel reilly related report representation robert roles seen semantic semantics sentence sentences serves sgall speci springer stepanek structure structured structures support swing syntactic technical tectogrammatical that there this those tilburg tool toolkit tools transforming tred treebank ufal university users using vaclav various verlag vidova wood work workshop xmlbased
http://acl.ldc.upenn.edu/N/N07/N07-2023.pdf	96	A Geometric Interpretation of Non-Target-Normalized Maximum Cross-channel Correlation for Vo cal Activity Detection in Meetings	activity approximation asru ation audio based boakye brown campiglio cation channel classi computer conclusions cross crosscorrelation crosstalk derived detection ellis eurospeech evaluo feature features france fugen geneva geometric harper have huang icassp icsi ieee ikbal improved interspeech island italy jeju korea kraft kumatani laskowski learning lecture lfel madonna maximum mcdonough meeting meetings mlmi model multi multichannel multichannels multiparty multispeaker normalization notes ostendorf overlapped parameters particular pfau pittsburgh proc processing recognition recorder recordings references renals schultz science selection south speech springer stolcke stuker switzerland system text toulouse trans type unsupervised using wrigley
http://acl.ldc.upenn.edu/N/N07/N07-1009.pdf	9	Structured Local Training and Biased Potential Functions for Conditional Random Fields with Application to Coreference Resolution	about achieved adds algorithm algorithms already annotating annual application approach approaches baseline because begin bene berger between bias biased bpfs capture cardie carvalho case casespeci citation class clifford clustering cognitive cohen collins computational concept conclusion conditional conference constrained converge converges coreference cost crfs data dempster discriminative during elds emnlp emotions entropy evaluation exible experiments exploiting expressions extra extraction fast feature fields finite finkel finley fourteenth from functions gibbs global goldstone graphical graphs grenager hammersley hardly hidden high higher hinton howe icml identity ijcai improving incomplete incorporating incremental inference information initializing integer integrated interactions into issue iteration joachims journal justies kaufmann kluwer labeling lafferty laird language lattices learning likelihood linear linguistics local loyal machine machines manning manuscript markov massachusetts matching maximizes maximum mccallum message methods minority model models morgan mpqa natural neal nips nite noun novel observed opinion opinions other output over parameters parsing partially peng perceptron pereira piecewise pietra potential practice prediction present previous probabilistic proc procedure processing programming punyakanok quicker quite random reduce references report resolution resources roth rubin rule sampling science second segmenting sequence sequential sixth society sparse stacked state statistical step stoyanov structured substantially summarization supervised support sutton systems technical that theory this through time training twice uncertainty understanding undirected university unpublished using variables variants vector very view volume weighting weights wellner were while wiebe wilson with zimak
http://acl.ldc.upenn.edu/N/N07/N07-1051.pdf	51	Improved Inference for Unlexicalized Parsing	accurate adaptation adjoining algorithms analysis annotated annotation annotations applied approximate assume available barrett barto based bayesian best better brants brown building cambridge carlo cascading chain charniak chart chinese chiou coarse coling collins compact comparable complexity computational computatoinal conference context corazza corpora corpus cowan cross dataset directly disambiguation discriminative dreyer driven dubey edge eisner emnlp english entropy errors estimation even fails features fine finkel free general german gildea gold goldwater goodman grammar grammars grif harder head henderson http induction inference informed interpretable introduction johnson joshi jurafsky klein krenn language languages large latent learning level levy lexicalization linguistics lingusitic loss manning marcinkiewicz marcus markov massachusetts matsuzaki maxent mcclosky metrics minimization miyao models mohri monte morphology multi naacl natural nederhof order pages palmer parse parser parsing pcfg pcfgs penn pennsylvania performance petrov pipelines press probabilistic problem process processing properties references reinforcement report reranking roark santorini satta scale scheme self semantic shallow shanker sima since skut smoothing software solving some spanish statistical steedman structural sutton syntactic tags technique their thesis they thibaux this titov train training tree treebank tsujii unlexicalized updated uszkoreit variation very vijay what when with wkshop word work zeros
http://acl.ldc.upenn.edu/N/N07/N07-1068.pdf	68	Toward Multimedia: A String Pattern-based Passage Ranking Model for Video Question Answering	algorithms analysis annual answer answering application applied approach aref asian augmented automatic bakeoff bakshi based beaulieu building caption chaison chang chen chinese chua circuits clvq color comprehensive computational computer conference constructing content context cross dependency description detection development digital edge elmagarmid engineering entity entropy evaluation experimentation extraction federation fernandes fung generate good growing harabagiu high human huynh ianeva ieee image images imagina indexing information informedia integrating interact interaction interactive international journal karger katz knowledge kraaij lafferty language learning lecture levow lexico libraries lienhart life line linear linguistics localization localizing makes management marton matching maximum medium method methods models multifaceted multilingual multimedia named natural news ngai nunamaker okapi over overview pages parser pasca passage pattern performance proceedings processing quan quantitative question recognition references region relations research retrieval robertson role search seeded segmentation segmenting seke semantic shallow sighan sigir simple sinha siteq smeaton smoothing software song study suffix summaries summarization symposium system systems technology tellex text third time track trans transformation trec trecvid trees ukkonen using video videoocr videoqa videos voorhees wactlar walker wernicke what word workshop yang yuan zhai zhang zhao
http://acl.ldc.upenn.edu/N/N07/N07-3010.pdf	139	Unsupervised Natural Language Processing using Graph Models	able accurate achieved advantage again agirre aiming algorithm also amount amsterdam annotation appear application arxiv australia automatic automatically available babylonian based basis been biemann bordag budapest building canada categories chinese cicling city claim class cluster clustering clusters coincidence coling companion comparable computational conclusion conference confusion construct construction contributions corpusspecific csli davidov dependency described different discovering discovery discussed disentangling dissertation distributional domains dongen dorow dunning eacl efficient electronically empirically employing enable enables entries evaluating evaluation exemplified experiments extraction extracts framework francisco frequency frequent from generalization generation geometry giuliano gliozzo graph graphs have here high http hungary identification identify immediate independence induction inter invariants italy kaufmann lacalle language languages large leads lecture like linguistics main manifest manual manuscript mart meaning measures methods mexico might model montreal morgan most naacl namely natural networks nips notes npcompleteness occurring only optimizing output parameters parser part pattern patterns preliminary preprocessing previously problems proc proceedings processing produce promising prove proven random rankings rappoport references reflects relation relations rely replace replaced report resource retrieval revealed rochester roget rule scale schaeffer science section semantic sense senses separator significantly similar similarity some soroa space speech stanford statistical statistics steps steyvers structure summarized supervised support supporting surprise sydney symmetric system tagger tagging tasks technical tenenbaum teresniak text textgraphs that thesauri thesaurus these they this trento triplet understood unsupervised until used useful using very well which whispers widdows will with word words work works york
http://acl.ldc.upenn.edu/N/N07/N07-1037.pdf	37	Extracting Semantic Orientations of Phrases from Dictionary	aaai acquiring acquisition adjectives affect algorithms analysis anatomy andrea annual application applications applied arti association attitude automatic baron based brin cambridge cation causal chapter chenggang cial cikm classi cluster coling collocations companion computational computer conference connective contextual criticism cues data david determining dictionary disambiguation domain down eacl emnlp empirical engine esuli european evaluation exaggeration expansion exploring extracting extraction fabrizio faye frank from fully glasses gloss graduate graeme graph hatzivassiloglou hidetoshi hiroshi hiroya hirst hoffmann human hypertextual image inference information institute intelligence international inui isdn jaap janyce japanese joint kamps kanayama kathleen kazuyuki kentaro knowledge kobayashi labeling language large latent lawrence learning level lexical lexicon linguistics littman lrec maarten mackay mainichi manabu management markers marx mckeown measure measuring meeting method methods michael mihalcea model models modern mokken morita nara nasukawa natural networks nishimori nozomi okumura orientation orientations oriented oxford page pages path paul peter phrase phrases physical physics plenum polarity potts praise predicting press probability problem proceedings processing rada ranking recognizing recreant references resources restoration retrieving review reviews rijke robert rumor rumormongers scale school science search sebastiani semantic sense sentence sentiment sequence sergey shao shimbun simple slud smadja society spin spring statistical subjectivity summarization symposium systems takamura takashi tanaka technology terms tetsuya text theories theory theresa thesis through thumbs tohru transactions transmitted turney university unsupervised using variable variation vasileios version vocabulary volume wiebe wilson with word wordnet words xtract york yueh zhongzhu
http://acl.ldc.upenn.edu/N/N07/N07-2038.pdf	111	Agenda-Based User Simulation for Bootstrapping a POMDP Dialogue System	agenda approach architecture aruba asru audio average based beach built computer data decision dialog dialogue dialogues domain dutoit eckert effectiveness engineering errorrate evaluated evaluating evaluation experimental extensively figure framework georgila hawaii henderson hidden honolulu human icassp ieee including information interaction involved issue knowledge language learned learning lemon levin machine management markov mining model native nding observable optimal palm partially pieraccini pietquin policy pomdp pomdps portability probabilistic proc processes processing prototype real recorded references reinforcement results review rudnicky scenarios schatzmann seattle simulation speakers special speech spoken state statistical stochastic strategies strategy stuttle subjects survey system systems talk task techniques total tourist towninfo training trans user users using weilhammer williams with word words young
http://acl.ldc.upenn.edu/N/N07/N07-2026.pdf	99	Look Who is Talking: Soundbite Speaker Name Recognition in Broadcast News Speech	aaai applications applied asru automatic available barzilay behind blind broadcast broadcasts canseco chang chen cjlin collins combining comparative configuration csie data detection diarization domain extraction feature finally foundations gauvain grishman hirschberg http identification identifying improving initial interspeech lamel library libsvm machines manual maskey naacl name news obtained proc radio reference references relation resolution results role roles rules selection shown software soundbite speaker speech springer strategies study support svms tagging test transcriptions using various vector whittaker with
http://acl.ldc.upenn.edu/N/N07/N07-2030.pdf	103	On using Articulatory Features for Discriminative Speaker Adaptation	aachen aalborg absent acoustic adam adaptation additional also analysis applications approach approaches articulatory asela asynchrony audio automatic bahl based between beyerlein bielefeld brown budapest byrne cambridge cation chomsky chris clustering codebooks combination comparison computation conclusion conditional context conver cost cued december decoding denmark dependent discrimination discriminative diskriminative during each editor ellis england english estimation eurospeech evaluate evaluation experiments exponents extent fakultat feature features florian foundations frame future gales geneva gerasimos german germany given graf gunawardana halle hans harper have having heidelberg hidden hmms hochschule however hungary icassp identi ieee importance improves infeng informatik information interesting intuitive isca janin japan june karlsruhe katrin kirchhoff lalit leggetter likelihood linear mark markov maximum meaning mercer metze might model modellkombination models morgan morris multi mutual necessary nelson noam october only other outperformed pages paper parameters particular pattern peter phil phonetic phonological plus point potamianos povey present presented prime proc ready recognition references regression report requires rheinisch robert robust rwth sational scandinavia seattle september setup slightly some sound souza speaker speech spracherkennungssystemen springer starting stream streams such summary switzerland technical technis that thesis they this time tokyo training transformations translation universitat university used using verbmobil visual volume wahlster weights westfalischtechnische which william with wolfgang woodland word work wortschatz york
http://acl.ldc.upenn.edu/N/N07/N07-1043.pdf	43	An Integrated Approach to Measuring Semantic Similarity between Words Using Information available on the Web	aaai acquisition adaptive advanced advances ambiguity application approach approch ariti association automatic bagga baldwin bandar based between bichnell bigham bigrams blettner cambridge challenge charles checking chen chichester cial clarke classi clustering cognitive coling communications comparison computational computer conference conrath content contextual coreferencing corpora corpus correlates cross curran cybernetics data development document double ecml emnlp engine engineering ensemble entity entropy estimates evaluate extraction fact facts foundations frequencies frequency from function goodenough graph gurevich hamasaki hasida hearst heilman hyponyms icml ieee index information intelligence international ishida ishizuka jain jiang joint journal keller kernel knowledge language lapata large learning lexical lifchits likelihood linguistics machines manning margin massachusetts matsuo maximum mclean measure measures measuring menthods methods metric mili miller million minning model modelling models mori multiple naacl natural nets network nishimura nition novel obtain organizing outputs pages parallel pasca platt polyphonet press probabilistic problems proc proceedings processes processing rada random references regularized research resnik retreival rocling rosen rubenstein sahami sakaki sampling schutze search searching semantic short similar similarity smith snippets social sources space speech statistical statistics step support synonyms synonymy system systems takeda taxonomy terra text theoretic theory thesaurus transactions turney uchiyama unseen using vapnik vector versus wide wiley with word words world yossef zuhair
http://acl.ldc.upenn.edu/N/N07/N07-2016.pdf	89	Relationship between Non-Projective Edges, their Level Types, and Well-Nestedness	algorithms alternative bulletin cken coling constraints dependency detecting diploma drawings eacl edge edges electronically emnlp encyclopedia fernando hajic have havelka hence integer joakim kiril kuhlmann least levelchildi levelchildu leveli levelu line linguistics marco mathematical mathias mcdonald mildly models neil nested nition nivre njas nonnegative nonprojective optimal ordered other pages pair parsing pereira point prague proc proceedings programming projective projectivity projectivizing published references research ribarov rooted ryan saarbr saarlandes satisfy sequences since sloane spanning structure structures suppose syntactic systems that theorem theory thesis totally tree trees type typei universit using witnessed
http://acl.ldc.upenn.edu/N/N07/N07-1019.pdf	19	Worst-Case Synchronous Grammar Rules	albert algebraic alon amer andrzej annals applications approaches april arnborg arti based bayesian beginning bound bull canada causal chandran cial cient cliffs close combinatorial compiling completes complexity computational computations concentrators consequences constant constructing construction containing context contextfree converges corneil cult derek discrete each edge eigenvalue embeddings emnlp englewood enoch example expander expansion explicit explicitly fact family finn free from further generalized geometric giorgio goes gons grammars graph graphs grows hall have hile hoory implies increases inequalities information input intelligence into isoperimetric jeffery jensen journal known kristian lauritzen lemma length letters linial local lower main math mathematics methods milman nathan nding networks obtain october olesen outlined pages parsing permutation permutations peserico plug prentice prime probabilistic probability proceedings processing proof propagation proskurowski provide quarterly references require result results rules satta second section shafer shenoy shlomo shown siam some spectral statistics stefen steffen step strategies string subramanian superconcentrators synchronous tabular tanner term that their theorem theory this time translation tree treewidth turn ullman updating vancouver volume where which wigderson will with without yielding zero
http://acl.ldc.upenn.edu/N/N07/N07-3002.pdf	131	Learning Structured Classifiers for Statistical Dependency Parsing	aaai adaptation algorithms american analysis approximate automatic bayes bikel blitzer boosting charniak cherry clustering coling collins computational conll constraints context corpus correspondence corston crammer decision dependency derived discriminative distributional doman driven eacl eisner emnlp english entropy euralex exploration free generative grammar grefenstette head icml ijcai improved inspired intricacies iwpt klein koller language laplacian large largemargin learn learning lexical lexicalized linguistics lizotte local mach machines magerman manning margin matsumoto maximum mcdonald model models multilingual naacl natural nities north oliver online pages parse parser parsers parsing pennsylvania pereira point predictors probabilistic proc ratnaparkhi references regularization retrieval ringger schuurmans second similar simple southey statistical statistics strictly structural structured support taskar thesis thirdorder three tishby training tree university unsupervised using vector wang wilkinson with word words yamada
http://acl.ldc.upenn.edu/N/N07/N07-1062.pdf	62	Efficient Phrase-table Representation for Machine Translation with Applications to Online MT and Speech Translation	acoustics alignment american annual applications april arbitrarily arbor assoc automatic bannard based beam bertoldi boston brill budapest burch byrne callison canada chapter chiang cient computational computer conf confusion consensus continuous corpora december decoder decoding eamt edmonton error european federico finding france francisco haeb hawaii hierarchical honolulu http human hungary icassp ieee improvements japan july june koehn language large larger linguistics long longer machine mangu march marcu mathias meeting mexico minimization minimum model moses naacl networks north november october oerder other pages phrase phrasebased phrases proc processing rate recognition references sapporo scaling schroeder search signal speech spoken statistical statmt stolcke technology toulouse training tran translation umbach understanding vogel volume word workshop zens zhang
http://acl.ldc.upenn.edu/N/N07/N07-1012.pdf	12	Information Retrieval On Empty Fields	american application approach arti based blair bull callan cation cial classi clustering cohen conf cooperative croft daoud data databases defazio desai document extended form friedland friedman fuhr getoor goyal heckerman ieee ifip ijcai indexing information inquery integrating integration intelligence jensen journal koller lafferty language learning macleod manage meek microsoft minimization model modeling models neville normal pages pfeffer ponte probabilistic probability proceedings process query rdbms references relational relations report representation research retrieval risk rotorua sadri science segal semantics sigir smith society srinivasan support syst systems taskar technical text trees universal using vasanthakumar vries whirl wilschut with word working york zealand zhai
http://acl.ldc.upenn.edu/N/N07/N07-1032.pdf	32	Cross-Instance Tuning of Unsupervised Document Clustering Algorithms∗	acoustics algorithm algorithms allerton alpha anal analysis ando annual application april arbor aspects association august available based beaulieu bialek blum boot bootstrapping bottleneck cambridge caruana categorization cation church classi clickthrough cluster clustering clusters code coifman colt combining communication communications comparative compstat computation computational conf conference corpus cover cross csiszar cspl cutoff data david decision denoising development dhillon disambiguation discovery divergence divergences divisive document editorial eisner elements emnlp empirical engines evgeniou exploiting extraction feature february fraley framework friedman from gatford generalized gorman guest guglielmi hamza hancock hastie hayes healy hero http human icassp icml ieee indexing inform information integrated intel international intl issue iterative jacobs january japan jelinek jensen jmlr joachims john jones journal july june karakos khudanpur knowledge krim kumar labeled laboratory lang language learning lewis linguistics lter mach machine mallela march marchette matlab maximization mclust measure measures meeting method methods michel michigan mining mitchell model multi multiple multitask mutual natural netnews nslonim october okapi optimizing pages park pedersen pereira perspective perspectives philip pontil predictive press priebe princeton proc proceedings processing raftery rates recognition references regularized relatedness release renyi report research retrieval rivaling robertson schuller search selection sense sensing sequential sigir signal sixteenth slonim socolinsky software solka sons special speech springer statistical statistics structures study submitted supervised symp symposium systems task tasks technical technology text theoretic theoretical theory third thomas tibshirani tishby torkkola training trans transactions trec trees tsao university unlabeled unsupervised using variable verlag walker wegman wiley with without word workshop yang yarowsky yokohama zhang
http://acl.ldc.upenn.edu/N/N07/N07-2029.pdf	102	Hybrid Document Indexing with Spectral Embedding	algorithms american analysis anne assignment ayman based bientinesi bottleneck chapman christian clustering clusters cross deerwester dense development dhilon dictionary document douglas dumais eigensolver farahat furnas geijn generalized george gerard gina hall harshman hill inderjit indexing information introduction irina issue journal landauer language latent levow management matrices matveeva mcgill mcgraw method michael micheal modern multidimensional multiple munkres naftali noam oard pages paolo parallel philip problems proc processing ranlp references relatively report representation representations research resnik retrieval richard robert robust royer salton scaling science scott semantic siam slonim society special susan symmetric technical techniques term thomas tishby transportation trevor using word
http://acl.ldc.upenn.edu/N/N07/N07-1024.pdf	24	Hybrid Models for Semantic Classification of Chinese Unknown Words	about accuracy achieved acquisition annotation applications association automatic based basic categories character charactercategory characters charniak chen chinese ciaramita cilin classification classifying coling combining commercial component compound compounding compounds computational conclusions construction contemporary contextual cooccurrence corpus corpusbased curran disambiguation distributional duan emnlp fine flexible framework grained hong important improve information into introduction issue johnson journal knowledge knowledgebased kong language lexical lexicon linguistics model models noun nouns occurrence pages peking performance phrase presented press proceedings processing references require research retrieval roark ronis rule semantic semi sense sighan similarity special state statistics student supersense tagging task template that thesaurus three tongyici tseng university unknown uses using weeds weir while word wordnet words workshop
http://acl.ldc.upenn.edu/N/N07/N07-1040.pdf	40	Whose idea was this, and why do es it matter? Attributing scientific work to citations	aberdeen ambiguity analysis anaphor anaphora anaphoric annotating annotation application articles artstein astronomy athens attribution australia automatic beijing beverly bhuptira birmingham bollacker burger cation china citation citeseer citing classi codata cohen computational computer computing conference connolly connor content coreference corpus czech daarc data deciding digital discourse emnlp exible experiments francisco frank function genomics giles greece grosz grover hersh hills hirschman hobbs humanities idea ijcai implementations improve indexing information international introduction iwpt java johnson karlovy kaufman kaufmann kraemer krippendor labels lagos language lawrence learning libraries linguistics lisbon literature loquium lrec machine management matheson message methodology mikheev mining model moens morgan multipaper nanba natural okumura overview passonneau poesio portugal practical prague proc processing pronoun publications punctuation readings recognition reference references relevance reliability republic resolution resolving retrieval rhetorical ross sage scheme science scienti scoring siddharthan sparckjones speech statements status summarising summarization sydney system techniques technology teufel theoretic theory third this tidhar tokenisation tool tools towards track trec understanding using vary vilain webber whose wiley with witten york
http://acl.ldc.upenn.edu/N/N07/N07-1039.pdf	39	Extracting Appraisal Expressions	accurate agrawal analysis andrea anthony applied appraisal association based bases bremen brill cardie choi christopher cikm claire collier commun computational computer conditional conf conference corston customer data database dayal determining diverse editors elds ellen emnlp empirical english eric esuli evaluation extraction fabrizio free from gamon generalized george gloss grammatics gray http human identifying information intelligent international kaufmann klein knowledge language large lecture lexical linguistics london machines madrid management manning martin meeting methods michael miller mining morgan mullen natural nigel nishio notes october oliver opinions orientation pages palgrave parsing part patterns patwardhan peter proc proceedings processing pulse rakesh ramakrishnan random references riloff ringger rule rules science sebastiani semantic sentiment september shojiro siddharth sigir simon simple sources speech springer srikant support symposium tagger technology terms text through tony umeshwar unlexicalized using vector verlag very vldb white with wordnet yejin
http://acl.ldc.upenn.edu/N/N07/N07-2040.pdf	113	Are Some Speech Recognition Errors Easier to Detect than Others?	addition analysis andrew approach arrangement automatic based boley both brill case centered challenges chang changes communication complements computational conclusion consecutive contrast corpus corresponding cues cult culty daniel data david dekang dependency detect detected detection determine diane different direction discovering dissection divisive domain dongsong driven early ease easier easy ecml edlund eric error errors evaluation example failed failures from further future gabriel greenberg have high highlights hirschberg impact importance increased information interested isca jens journal julia knowledge lacking language learning level lina linguistic linguistics litman make management manual marc mckoskey millennium minipar modi natural nominal nouns other output pages parsing part participants particularly partitioning pcomp performance phrases poses preposition principal proceedings processing prosodic ratio recognition recogntion reference references relations requires revealed robustness scrutinizing sears sentence showed shown shuangyu signal skantze some speech steven study subj subject successfully such summary swerts switchboard syntactic systems table tagging that they this transaction transformation travel typical user using verb were when which will word words work workshop yongmei zhang zhou
http://acl.ldc.upenn.edu/N/N07/N07-1069.pdf	69	Can Semantic Roles Generalize Across Genres?	alternations argument arguments beth carreras categorial categories center chen chicago classes colorado combinatory computational conference conll cslr daniel deep dowty edward emnlp empirical english evidence features gildea grammar hacioglu hockenmaier identifying international introduction investigation james japan john julia jurafsky kadri labeling language levin linguistic linguistics loper machines marquez martha martin methods natural owen pages palmer parsing pradhan preliminary press proceedings processing proto rambow recognition references report research role roles sameer sapporo selection semantic shallow shared spoken support task technical thematic ting university useful using vector verb ward wayne workshop xavier
http://acl.ldc.upenn.edu/N/N07/N07-3009.pdf	138	Combining Evidence for Improved Speech Retrieval	alicante alzghool approaches archives audio based building cantly cation classi clef collection computer contribute contributions conversational cross dissertation document domain evaluation experiments fast forum history human icassp improving information inkpen knowledge language languages lecture modal multi multilingual notes numerous oard olsson oral papers proceedings published references research retrieval science search selected series seventh several sigir signi spain speech spoken spontaneous subword task temporal test text this unconstrained ways will with workshop
http://acl.ldc.upenn.edu/N/N07/N07-4015.pdf	155	Text Comparison Using Machine-Generated Nuggets	across allow along annotator annotators answers appeared appendix automatic automatically collins comparing content definition demner dissertation driven each emnlp equivalence equivalent evaluating evaluation extracted frequency from fushman gram group grouped groups head hovy indicate individual into language method models modify multiple naacl natural nenkova nugget nuggets number occurrence passonneau pennsylvania present proceedings processing pyramid questions references score select selection semantic sentences statistical statistics summaries summarization text texts their there university using when with
http://acl.ldc.upenn.edu/N/N07/N07-2046.pdf	119	Entity Extraction is a Boring Solved Problem – or is it?	absence also among appear arbor around atypically based bergen classifier cohen columbia combination composition conducting conll crfs current deep dependent detection eacl easy editorial editorially edmonton email emnlp entity extracting extraction features find finkel florian from gazetteers gibbs grenager grishman grover improving incorporating informal information into involvement ittycheriah jing keyword kosseim learning like local manning mccallum mikheev minkov moens mohit most much name named names news noms obvious occurrence partir personal poibeau pristine proc programs propres recognition reference references relation relaxed remain research resolution sampling seeming semi similar skin sources standards struggle study suggests sundheim supervised surprised sutton syntax systems tagging taln techniques text textes texts that this through toulouse transfer trivia used vancouver vari very wang were what while with without word zhang
http://acl.ldc.upenn.edu/N/N07/N07-2031.pdf	104	RH: A Retro Hybrid Parser	about absolute accuracy acquired actually adaptation additional adjust advantages alex alon alone also analysis analyzer andy anette aoife approach approximations arbor automatically available average back bank barcelona based baseline becker been bernd berthold best beyond boston boundaries budapest build burke cahill cambridge chanod charniak chunk chunker chunks classifier claude coling collins columbia combine communications compete complexity consistent constrain constraint construction corpora corpus corpustrained coverage crouch crysmann current dalrymple data daum david deep deepunification degrading dependency development disambiguation discussion distance donovan driven eacl efficiency efforts elegant ends engineering english estimate eugene evalb evidence facilitate fast field finally foth frank from full gala genabith generate generation geneva german given good grammar grammars graphs have head hpsg http hybrid hybrids ideal identifier idiom improved including incremental information integrated integration interpreted itself jean john johnson josef kaplan kenji kiefer kilian king language large lavie leads length limited linc linear linguistically long maintenance mark markus mary maxwell mcclosky meets menzel method methods michael models mokhtar more much naacl natural network never newman notations numbers nuria obtained only open operating organization other parc parse parser parsers parsing paula pcfg pennsylvania pierre practical prefer press prevents problem proc provide questions rapid rather reduce reducing references related relative relatively remain require reranking research resolution results reversible richard riezler robust robustness ronald roux ruth sagae salah same sapporo satoshi schaefer sekine self sentence several shallow shallowness show shown significant single slow soft software speed speedups statistical stefan stephan stochastic sydney tagged tagger techniques technologies test texttree than that their them there these thesis they time times topological topp tracy training transition ulrich unification university using vancouver vasserman very wcdg while whose wide william with wolfgang woods words work workshop would
http://acl.ldc.upenn.edu/N/N07/N07-2034.pdf	107	An integrated architecture for speech-input multi-target machine translation	ability acoustic algorithm amparo approach approaches architecture attempt automata barrachina based basque been both brown cambridge cant carry casacuberta classical communication compared composition computational computer concluding contrast decoding decoupled della editors embedded emmanuel enable engine english enrique errors estimation extracted feature fernando finite francisco franz from fully further garc giza gonzalez ines inferred input integrated integrates into introducing itself january language languages linguistics llorens machine mart massachusetts mathematics measure mehryar mercer michael model models mohri molau mono most multi multiple nature nevado nite pages parameter pastor pereira pico pietra press proceedings process processing proposed recognition recognized reference references remarks results riley roche sanchis schabes series several sfst sfsts shown signal signi simultaneously some source spanish speech star state statistical stochastic systems table target teresa terms text that these this tillmann torres training transcription transducers translation unique using varea varona vidal vilar viterbilike weighted with work workshop yves
http://acl.ldc.upenn.edu/N/N07/N07-1033.pdf	33	Using “Annotator Rationales” to Improve Machine Learning for Text Categorization∗	aaai algorithm analysis august away based beineke black blum bounds cannot categorization cation cause classi complete computation computer conf could culotta cuts depends dumais education effort emnlp european factor feature features gale global hastie highlighted hints humanprovided ieee ijcai important improving including information intelligent interactive joachims jones july kuusela labeling learning lewis machine machines madani magazine many mccallum minimum mostafa network networks neural node ocone pages pang pixels possibly prediction proc raghavan reducing references relevant review rivest sciences selection sentiment sentimental sequential side sigir simply since structured subjectivity summarization support suppress svms system systems take tasks techniques text that this those thumbs training using vaithyanathan vector white with would
http://acl.ldc.upenn.edu/N/N07/N07-4005.pdf	145	Automatic Segmentation and Summarization of Meeting Speech	accelerated acoustic also amount annual approach association australia automatic being better boundaries broadcast browser carletta cation chain chapter characterizing classi coarser cohesion combining communication comparing compresses computational conditional conference constituent conversation conversational data degree dialogue discourse domain done each emnlp erent european explored extractive features following fosler from galley gradually granularity greater have hirschberg hsueh importance important informative jing leaving less level levels lexial lexical linguistics lisbon lussier maskey material mckeown meeting moore most multiparty murray nding news only performs portugal predicting previous problem proceedings prosodic random ranking recent recordings references related removes renals scored segmentation september showed skip speech structural studied summarization supervised sydney technology that topic user using utterances words work years yield
http://acl.ldc.upenn.edu/N/N07/N07-2050.pdf	123	Chinese Named Entity Recognition with Cascaded Hybrid Model	adaboost application arti boosting carlo carreras chain chapman cial coling computational computer conll decision extraction foundations freund generalization genesereth gilks hall intelligence kaufmann language learning line logical london markov marquez mateo michael monte morgan named natural nils nislsson padro pages practice publishers references richardson robert schapire sciences spiegelhalter system taipei theoretic tity using xavier yoav
http://acl.ldc.upenn.edu/N/N07/N07-1014.pdf	14	A Random Text Model for the Generation of Statistical Language Invariants	accurate advances albert algorithms allauzen annual association barab cancho coincidence complex computational conference constructing debrecen dunning emergence extension ferrer garcia generalized graphs human international language linguistics london mathematicae meeting methods ming models mohri networks phrases proceedings publicationes random references roark royal scaling science sicilia smallworld smith society statistics surprise systems texts words zipf
http://acl.ldc.upenn.edu/N/N07/N07-1018.pdf	18	Bayesian Inference for PCFGs via Markov chain Monte Carlo	accuracy achieving acquisition algorithm algorithms analyses analysis andrew annealing annotation annual approximate associates association australia available aware based basic bayesian bias boltzmann brian carlo cascading cases catherine centroids chain chan charles child chris christopher cient component computational computer conclusion conference constituency context corpus cross data demuth dependency described ding dirichlet distribution distributions donald dynamic edition editor eisner empirical ensemble erlbaum errors estimation estimator exact exchange finkel free from geman generation gentle gibbs given goodman grammar grammars harvard hastings hillsdale http ieee images induction infer inference inside intelligence international james jason jenny joshua joshuago journal katherine klein language lari lawrence likelihood linguistic linguistically linguistics machine macwhinney manning markov match maximum meeting methods microsoft models monte morpheme morphological multilingual natural noah number outside over pages paper parse parsing pattern pcfg pcfgs performing pipelines posterior prediction presented prior probabilities problem procedure proceedings processing product production programming random references relaxation research restoration returns right rose sampling score secondary sesotho slobin smith snow solution solving speech springer standard stochastic strings structural structure stuart study sydney syntactic system task terminal their these thesis this thus transactions trees trivial uninteresting university unlabeled used useful using verbs volume weighted which word yields york young
http://acl.ldc.upenn.edu/N/N07/N07-1070.pdf	70	Towards Robust Semantic Role Labeling	adaptation adrian american analysis arbor automatic bacchiani bejan benfeng best bilingual boas boston brian brown carreras cation charniak chen chinese chunk classes coarse computational computer conll construction corpus cosmin daniel dictionaries disambiguation discriminative edinburgh emnlp english eugene extraction features framenet francis fung gildea girju grammars hans harabagiu henry identi ijcai interdisciplinary introduction johnson jurafsky kucera kudo labeling language learning linguistics lrec machine mark marquez matsumoto maxent melinger michael michiel morarescu naacl nelson pages parser parsing pascale paul performance present press proceedings providence references relation representation reranking richard riley roark role roles roth sammons sanda schulte scotland semantic semantics shallow shared speech sproat stochastic support taku task tokenlevel translation university variation vector verb verbnet walde workshop xavier yuji
http://acl.ldc.upenn.edu/N/N07/N07-1015.pdf	15	A Systematic Exploration of the Feature Space for Relation Extraction	achieve acyclic adding also among anthony aone arguments aron automatically based basic been before best between both bunescu chad chinatsu combination combined combining complex complexity composite conduct considered contains context convolution corpora corpus coupled culotta cumby data dependency different directed discovering dmitry each effective effectiveness eisaku emnlp entities entropy even existing experiment explored exploring extract extracting extraction feature features found framework from future general gives global good graph graphic grishman guodong hasegawa have heuristics hierarchical hirao hurt icml improve includes information instead integrated jeffrey jian journal kambhatla kenrel kernel kernels knowledge language large lead learning levels lexical local machine maeda maximum methods models mooney more most much naacl named nanda natural nips only oriented overlap parse path performance posed previous proceedings pruning ralph raymond razvan reasonably references relation relational relations representation representations research results richardella roth sasaki satoshi search sekine semantic sentence sentences sequence sequences serves shortest showed shubin slightly some sorensen space starting statistical statistics structured studied studies study subsequence subspace subspaces suggests suzuki syntactic systematically takaaki task that these this three tree trees tsutomu unit using various weighting when which will with work yutaka zelenko zhang zhao zhou
http://acl.ldc.upenn.edu/N/N07/N07-2021.pdf	94	Semi-Supervised Learning for Semantic Parsing using Support Vector Machines	aaai accuracy achieves acknowledgments across advances algorithm also amount annotated approach arbor august australia average based bennett best bled borland both bottou built call cambridge cantly categorial cation chapelle chen classi clause clearly cluster coling collins collobert comparison conclusions conll constructors corpus cristianini curves data database demiriz demonstrated dempster dong during ecml edinburgh editors effort examples experimental experiments exploitation extracting figure form formal freiburg from germany given google grammars grant guide gure hand humanannotation icml improved improvement incomplete increased inductive inference information integrates international introduction joachims john journal july june kate kernel kernels laird languages large learning letters level likelihood limited lodhi logic logical machine machines mastodon maximum measures method methods mooney multiple natural neural number obtained obtains only other over pages paper parse parser parsers parsing particularly pattern perform performance pittsburgh portland precision presented press probabilistic proc processing programming progressive prolog provided queries reaches recall recognition reference references research resulting results retrained royal rubin same saunders savings scale scholkopf scotland scotts seen semantic semantics semi sentences shawe show shows signi sinz slovenia small society sons statistical string structured supervised support supported svms sydney syntax systems tang taylor text that then theory these this trade training transductive transform turbo unannotated university unlabeled uses using utilizes utilizing valley vapnik vector wang watkins were weston when while wiley with wong zelle zettlemoyer zien
http://acl.ldc.upenn.edu/N/N07/N07-2001.pdf	74	Comparing User Simulation Models For Dialog Strategy Learning	aaai approaches arti asru assessment automatic based cambridge cial communication communicator comparing corpora cozar data design dialog dialogue diss effects empirical evaluation experiments features from georgila henderson hybrid ijcai intelligence journal kearns knowledge learning lemon litman lopez managment means model naacl njfun optimizing policies practical proc quantitative real reasoning references reinforcement research rubio schatzmann schef segura sigdial simulated simulation singh speech spoken state statistical strategies stuttle supervised system systems technique techniques tetreault torre university user using utility walker weilhammer with workshop young
http://acl.ldc.upenn.edu/N/N07/N07-1023.pdf	23	Lexicalized Markov Grammars for Sentence Compression∗	aaai accurate adjunct against algorithms also analysis andtrim annotated annotation approach assembler association automatic based best better between bigrams bilexical building careful charniak cient coarse collins combination compression comput computational computer computing constituent constraints contrast corpus dependencies dependency describes directed directly discriminative discriminatively distance does donald door dorr driven each eacl editing elementary emphasis english estimates evaluates evidence extracted factored fast feature features framework from furthermore generation grammar hardly head headline hedge include incorporating jing johnson joshi journal klein knight language large leads learning leaves levy lewis lexical lexico likelihood linguistic linguistics literature machinery makes manning marcinkiewicz marcu marcus maxent maximum mcdonald model models modi more much naacl natural noisy open optimization optimize order pages paper parse parsing partially pcfg penalized penn pennsylvania power problems proc proportionally pushdown puts quality reduction redundant references related relies representations reranking robustness santorini schwartz science selection sentence shasha siam simple since sister soft statistical statistics stearns stems step strength strives structure such summarization supervised syntactic syntax system takahashi text than that thesis think this thus transduction translations tree treebank trees turner type ullman understanding uninformative unlexicalized unsupervised uses vertical volume weighted well what which with word work zajic zhang
http://acl.ldc.upenn.edu/N/N07/N07-1072.pdf	72	Computing Semantic Similarity between Skill Statements for Approximate Matching	aaai accurate agreement alternative annotated annotators approach assessing attachment automatic automatically baker bank based berkeley budanitsky california cambridge carletta categorization charniak chklovski class classification cogsci coling comparison computational concepts conrath construction contextually corpus dang database definition demonstration dependency diab disagreement disambiguation ellsworth entropy evaluating evaluation exploiting extracted features fillmore framenet gildea giuglea gordon grammars hacioglu hirst hulth human icml information inspired intelligent international jiang johnson journal jurafsky kappa keywords kingsbury kipper klein labeling lexical lexicography lexicon line linguistics lrec machines manning martin maximum mccord measures measuring megyesi memory michelizzi mihalcea miller minipar moschitti naacl pages palmer pantel parse parser parsing paths patwardhan pedersen petruck phrase practice pradhan prepositional proceedings processing propbank proposition psychological quillian ranlp references relatedness resnik review rocling role roles ruppenhofer semantic sense shallow similar similarity slot statistic statistics study support swanson systems tasks taxonomy text theoretic theory tree tversky unlexicalized unsupervised using vector verb verbnet ward word wordnet wordnetbased words workshop
http://acl.ldc.upenn.edu/N/N07/N07-1067.pdf	67	Question Answering using Integrated Information Retrieval and Information Extraction	adam allan andrew answering blair borchardt bowden brownell christine clark conference croft dang david dependencies description diane document employing english evaluation external felshin focused forthcoming fourteenth giridhar goldensohn grishman harabagiu hickl http information integrating james jimmy kathleen katz kelly knowledge kumaran line loreto louis markov marton mckeown metzler meyers mitchell model models moldovan mora nist overview pages patrick proceedings publication queries query question ralph random references relation retrieval rhetorical riao rosenberg sanda sasha semantic sigir sources stiller summarization system systems techniques templated term text track trang trec understanding uzuner wang westbrook wilcox workshop
http://acl.ldc.upenn.edu/N/N07/N07-1013.pdf	13	Improving Diversity in Ranking using Absorbing Random Walks	america anlp arti assoc automatic based brandow carbonell centrality cial cluster condensation david diversity document documents doyle dragomir edition electric electronic erkan extraction gather geoffrey goldstein graphbased grimmett gunes hearst hypothesis intelligence jade jaime journal kantrowitz lexrank lisa manage mark marti mathematical mittal mitze multi naacl networks oxford pages pedersen probability process processes producing publications radev random reexamining references reordering reranking research results retrieval salience scatter science selection sentence sigir snell stirzaker summaries summarization text third vibhu walks workshop
http://acl.ldc.upenn.edu/N/N07/N07-2025.pdf	98	Exploiting Rich Syntactic Information for Relation Extraction from Biomedical Articles∗	aaai abstracts accuracy andrade aone automatic based best between binary biological biomedical blaschke boost both bunescu cantly cation claudio combination combining comparison composite comprehensive conclusion contrast convolution cross culotta current dependency different discovery discriminative drug eacl emnlp entire entities entity entropy experiments exploiting explored extract extraction features feldman finkelstein flat fold frequently from functional gene genes genome gives gures higher hurvitz identi identifying informatics information interaction interactions ismb journal kambhatla kernel kernels kogan labeling landau large lavelli learning lexical linguistic literature lower machine manually many maximum medline methods miller mining model models mooney more moschitti naacl named novel number obtains originally outperforming outperforms ouzounis over overall paper park parse parsing path popowich poster precision previously proc products proposed protein provide ramshaw recall references regev relation relations report research results rich richardella role romano sarkar scienti seen sekimizu semantic session shallow short shortest shows signi simultaneous sorensen speci statistical structured study syntactic system table task techniques ternary text that they this tree trees tsujii used using usingstatistical valencia validation verbs weischedel which with word zelenko zhang zhou
http://acl.ldc.upenn.edu/N/N07/N07-1044.pdf	44	An Information Retrieval Approach to Sense Ranking	accurate acquired across alberta alexander algorithm allwords analysis annotation annual antal application association australia automatically barcelona bart based benjamin berlin biewald bosch briscoe brody budanitsky cambridge canaria carlo carroll chapter christiane christopher church clark claudio cognition computational computers conference corpus cotton daelemans dang daphne database david decadt dekang delfs determining diana differentiating disambiguating disambiguation discourse distance diverse dixon domains dominance durkin editor editors edmonds electronic empirical engineering english ensemble entire european evaluating evaluation experimental fellbaum finding firth florian francisco freiburg frequency gale gambl general genetic germany giuliano glimpse gliozzo graeme gran gruyter harris hirst homographs homonymy hoste human humanities information international intraword italy jocelyn john journal julie kenneth kevin kittredge koeling koller language lapata large lauren learning lehrberger lexical lexicon linguistic linguistics luke machine madison manber manning marc martha mccarthy meaning measures meeting memory method methods mihalcea mining mirella mohammad naacl natural navigli nition norms optimization oriented other oxford pages palmas palmer parameter peter phil philological pittsburgh polysemy predominant press proceedings processing psycholinguistic rada radu references relatedness relative research resources restricted retrieval roberto robust saif salience sample samuel scott search semantic sense senses senseval similarity snyder society spaces statistical stokoe strapparava subjective sublanguage sydney synonyms synopsis systems task tasks taylor technical technology text teyssier theoretic theory thesaurus through toefl tool toulouse trang translation trento turney twilley university unsupervised untagged usenix using vancouver verb veronique versus vickrey walter weeds william winter word wordnet words workshop yarowsky york zellig
http://acl.ldc.upenn.edu/N/N07/N07-4011.pdf	151	Learning to find transliteration on the Web	accepts added ague alen algorithm alla along amta anchor andi answer answers anti appears application approach associated bart base bata belo beth between bilingual bottom buch chang char chien chinese chinet choose click clicking clustering coling computational conference containing corpora data delete demonstration deng dictionary dinstl document each edited engine expanded extracted extraction feature finder finds form from graehl improved information intelligence international journal knight koehn kwok likely linguistics list local look machine microsoft mining monolingual morphe most nagata name names ngra noun number obtained opens parallel peng performance phonetic phrase phrases postfixes prefixes previously prob proc proceedings program queries query references relationships results retrieve rich right runs saito save script search select selects sending showed shows slide snippet snippets statistical string suzuki system systems text them things this tois transactions transitive translation translations transliteration transliterations triage under user using view where windows with workshop
http://acl.ldc.upenn.edu/N/N07/N07-3007.pdf	136	Dynamic Use of Ontologies in Dialogue Systems	advisor aist allen amanda annie annika annual antonio applications arbor architecture baader batista benjamin built byron calvanese cambridge campana chambers colea components conference construction cooking cslu cummings daniele deborah demonstration description design dialogue diego diverse domain donna dzikovska edition editor editors ellen engineering eriksson ferguson fernando filipe flycht franz galescu generic george giovanni gregory handbook hans human iberamia implementation inesc information interactive international james joana joint joseph knowledge language linkoping lisboa logic lucian main mamede mariani mary mcguinness meeting michigan myroslava nardi nathanael natural nuno ontologies ontology pages pardal patel paula paulo pena peter pittsburgh porf portugal poster press proc progress providing recent references report ribeiro ricardo sbia sbrn schneider school scott sessions shell spoken state stent stoness survey swift systems technical technology theory thesis trips understanding university using uszkoreit varile victor workshop zaenen zampolli
http://acl.ldc.upenn.edu/N/N07/N07-2002.pdf	75	Automatic acquisition of grammatical types for nouns	acquisition against allow annual applied automatic baldwin bond bootstrapping briscoe cambridge carroll chesley china city conclusions conference congress copestake corpora corpus count countability courses csli data deep definitions dictionary distributional domain domains duden economy edition eibe english euralex evaluating expected extraction fang feature features fifth frames francisco frank french from genoa german global grammar grammars grammatical hpsg ijcnlp impact implementing information international joint kaufmann korhonen language learned learner learning learnt lexical linguistic little lrec machine manual mass medicine meeting mining morgan natural nouns obtained occurrence ones parser pcomp performance practical prec proc proceedings processing publications references relevant report resources results salmon sanya should show shows shulte siglex signatures single statistical structure subcategorisations subcategorization successfully table technical techniques test testing that their there tools training trans typed ucam university unseen using valid variation verb walde washington when with witten workshop
http://acl.ldc.upenn.edu/N/N07/N07-1008.pdf	8	Direct Translation Mo del 2	adria alignments annual arbor australia banchs based bilingual brown building carnegie chiang computational computer conference crego darpa data david della department distortion divergence elds entropy erty estimation euparl features foster george gispert given grams hierarchical hong improvement inducing international john josep june kishore kong lambert linguistics machine marino mathematics maximum meeting mercer michigan minimum model models niyu onaizan pages papineni paral parameter peter phrase pietra presentation proc proceedings rafael random references report robert science statistical stephen sydney technical texts tides translation trik university using vincent word workshop yaser
http://acl.ldc.upenn.edu/N/N07/N07-1034.pdf	34	Combining Reinforcement Learning with Information-State Update Rules∗	able achieves action actions acts adding after airline allen allowing allows also analyzing another application applied approach arti audio automatic average barto based bene bohlin both cambridge canada capture characteristics cial cognitive cohen combined complex computational conclusion converge conversants cooper cost crafted cult database demonstrated depends designers determine dialog dialogue diego discourse does domains easily eckert effects elements email emnlp enable enables enabling engdahl engine engineering engines english epochs eurospeech example exibility experiments figure formalism formalizing formulated from georgila goebel good grounding hand happen heeman henderson human ieee ights ijcai improved including information initiative intelligence intention interaction journal kearns knowing knowledge language large larsson learning lemon less levin lisbon litman logical machine mackworth management managment matheson means might mixed model modelling more move much naacl natural need nitions njfun none number numbers obligations obvious occur october optimal optimizing oxford pages paper parameters partial perrault pieraccini plan poesio policies policy poole portugal practical preconditions press problems proceedings processing quickly reasoning references reinforcement relative requery require research rich rules same schef science search seattle second seen selection setting should simulation simulations singh situations smaller solution space speech spoken standpoint state states stochastic stops strategies strategy sutton system systems takes than that theory these this thus toolkit transactions traum trindi under university until update used user users using utterances vancouver versus walker whether which with workshop young
http://acl.ldc.upenn.edu/N/N07/N07-2004.pdf	77	JOINT VERSUS INDEPENDENT PHONOLOGICAL FEATURE MODELS WITHIN CRF PHONE RECOGNITION	acero acknowledgments acoustic agency also anonymous approach articulatory attributes authors automatic based bayesian berkeley bielefeld capturing cation chang classi combining computer conclusions conditional conference conjunction continuous corpus crfs dahlgren darpa data designing detectors disc discriminative dynamic effect elds elitist evaluating evaluation event expressed feature features february fiscus fisher fosler four framework frankel funding gaithersburg garofolo grant greenberg gunawardana hasegawa have hidden hierarchical hopkins html http icassp icsi icslp ieee independently information institute international interspeech intrinsic itrw johns johnson joint king kirchhoff knowledge labeling lafferty lamel landmark large launay learning livescu lussier machine mahajan mccallum modeled modeling models moore morris national network networks neural nist nistir october often opinions over package pallett pereira performance phone phonetic phonological platt posteriors probabilistic proceedings pronunciation provide providing published quicknet rajamanohar random recognition recogntion redundant references report reviewers robust science segmenting sequence sharenborg shown siohan software speech standards summary summer superior supported surendran technical technology thank that thesis this those through timit towards trained understanding university using variation vocabulary well wester with within work workshop
http://acl.ldc.upenn.edu/N/N07/N07-2036.pdf	109	Stating with Certainty or Stating with Doubt: Intercoder Reliability Results for Manual Annotation of Epistemically Modalized Statements	aaai academic adverbs affect agreement amsterdam annotating applications applied approach attitude attribution beads benjamin benjamins between biolink biological biometrics bioscience braclets cambridge cardie categorical categorization certain certainty chicago citation coates coefficient cognitive cohen comment company computing concepts conceptualization conference construction contemporary contexts could criteria cscsi cues data databases dimarco directions disciplinary discourse doctoral doubt educational emotions english epistemic epistemological european evidentiality exploring expressing expressions facts fifth frequency future fuzzy hedges hedging holmes hoye hyland identification identifying information international iral john journal kando kluwer knowledge koch kroon lackoff landis language less liddy light linguistic linguistics linking literature logic london longman manual meaning measurement mercer modality model mood more mushin narrative netherlands newswriting nominal nuyts observer ontologies opinions palmer paper papers part philadelphia philological possibly pragmatic pragmatics presented prespective press proceedings psychological publishers publishing references relc research results retelling retrieval review rubin scalarity scales sceio scientific sented series society speculations spoken spring springer srinivasan stance stanford statements studies study symposium syracuse tagging teaching text texts that theories theory thesis think transactions university unpublished verlag westney wiebe wilson writing york zuck
http://acl.ldc.upenn.edu/N/N07/N07-1003.pdf	3	Avoiding and Resolving Initiative Conflicts in Dialogue∗	aaai acoustical action adapted allaby allen america analysis annotation arthur articulation attention barbara brown cambridge candace carroll clark coding cognitive cohen collaborative communicative computational connected core crystal cumbaa curry damsl deanna dialogue dialogues discourse duncan duration during evidential fall fitzgerald gibbs grosz groups guinn herbert house humans initiative intentions interaction interactions james jennifer journal language latulipe linguistics machines mark michael model modeling moussa notes oriented pages pooley process qian rate references referring robin scheme science selection siddiqi sidner society speakerauditor speaking speech starkey stress structure syllables symposium task thomas tracking turns user what wilkes with working
http://acl.ldc.upenn.edu/N/N07/N07-2005.pdf	78	K-Best Suffix Arrays	algorithms arrays associative below bentley binary case charles church clifford communications considerably constructed cormen demonstrating easier edition figure gene introduction kenneth leiserson line louis manber mcgrawhill method more multidimensional myers note popularity press queries random references represented rivest ronald samples scenario search searches searching second simulate soda squares stein string suffix than that these thiesson thing thomas trees typical used well wild worst
http://acl.ldc.upenn.edu/N/N07/N07-1035.pdf	35	Estimating the Reliability of MDP Policies: A Confidence Interval Approach	acknowledgments active algorithms anonymous application approach assumption available bagnell barto based baseline becomes boosting bounds case categorization cation certain chickering classi comments communicator comparing complicated computer concept conference construction converge converging cumulative currently data decision dence detail dialog dialogs dialogue dialogues discourse does drew dumais email enough estimation european eurospeech expected extension features finally four frampton from future georgila grants group groups heckerman help henderson however human hybrid ijcai indicate inductive information initial interaction interval intervals investigating itspoke jair jaulmes jeff jordan kearns keep knowledge last learning least lemon levin like litman machine makes management markov matrix model more most msri naacl ndings networks neural nips nonlinear observable overview paek paper parameters partially personnaz pieraccini pineau pitt platt policies practical precup press proc processes provided references reinforcement reliable remain repetition representations research reviewers reward rivals sahami schapire schneider selection sigdial singh size some spoken squares state stochastic strategies strategy such supervised support sutton system systems tetreault text thank that their this tighten transition unreliable user using utility version walker well what where wide wkshp work workshop would
http://acl.ldc.upenn.edu/N/N07/N07-2014.pdf	87	Arabic Diacritization through Full Morphological Tagging	acoustic analysis analyzer ananthakrishnan andreas annotated annotation approaches approximations arabic automatic bangalore barcelona based bies british buckwalter building cairo center challenge coling columbia computational computer conclusion conference context corpus data david detailed diacritics diacritization diacritizer dictionaries dimitra disambiguation discussion draws egypt elds emnlp entropy extensible fast feature features fell full functional further general generator geneva gimenez habash hajic have highly hubert icon icslp imed india international jeffrey kanpur katrin kernel kirchhoff kudo kulick language languages large learning lexical linguistic lrec maamouri machine machines marquez matsumoto maximum methods modeling mohamed morphological morphology naacl narayanan nemlar noah number optimized otakar outperform parsing partof penn proceedings processing rambow random recognition references report resource resources restoration ruhi sarikaya scale script seth shown smith smrz society sorensen spain speculate speech spoken srilm stolcke support svmtool switzerland swoop system systems tagger tagging taku techncial text that theories through tokenization toolkit tools transcripts translation treebank treebanks tromble university uses vector vergyri version with work workshop yuji zitouni
http://acl.ldc.upenn.edu/N/N07/N07-2010.pdf	83	Situated Models of Meaning for Sports Video Retrieval	action activity alignment allen analyses analysis annotation announcers approach artificial automated automatic automatically average barnard baseball based berger blei both bouthemy bremond brown camera change characterization circuits classification combines combining common comp computational computer conference content context croft cucchiara currently data decamp della detection direct discuss duygulu each effect elements engineering entropy estimation event events expected feature features figure five fleischman forsyth four freitas ganansia gelgon general generated given gong grana highlight hongen ieee image improving inbaseball indexing information intelligence intentional interestingly international jordan journal lafferty language learning length limited linguistic linguistics machine marchi matching mathematics maximum mean measure mercer methods metric mining model modeling moore motion movement mpeg multimedia natural nevatia occurring only parameter patterns performance performs pictures pietra ponte poor precision prep probabilistic proc proceedings processing produces queries query ranking recall recognition references relevance represent representation research result results retrieval returned robert semantic shot show shows sigir sigmm significantly similarly simple situated snoek sports statistical system systems tardini technology temporal tendency text that this though time together topics training trans translation tutorial understanding unified used using video videos vision word words workshop worring
http://acl.ldc.upenn.edu/N/N07/N07-2012.pdf	85	A Filter-Based Approach to Detect End-of-Utterances from Prosody in Dialog Systems	accurate across algorithm almost annual another applications approach arabic area audio automatic average back bagging based basic bayyari behavior best better bleeding boosted boosting calculated cascade case cation certain channel channeling channels classi combinations computer conclusions conference conversation cross crow cues cvpr data decision detecting detection dialogues different direct does done each endof endpoint ensembles entropy error exception experiments faster feature features ferrer fold frank frequencies from fuentes future generated generation good graphics hakkinen hariharan have hoiem hollingsed however icassp icslp icsp identi ieee implementations improved interactive interesting international interspeech into iscslp jones journal kaufman laurila learning lter ltering lters machine mapping master method methods mining mislead mistakes modeling more morgan music object obtained ofutterance only other overview paso perceiving performed perhaps possible practical precision prediction proceedings processing produce programs prosodic prosody pruning quinlan random rapid real recall recognition reduced references require responsive results robust runs seed shown shows shriberg similar simple simplifying solorio speaker speech spoken still stolcke streams studies study stumps such suitable sukthankar summed system table tables taking task techniques tested texas texture that these thesis this time tool tools trees turn tutorial university used using utterance utterances validation viola vision ward weka well were which with witten work yield
http://acl.ldc.upenn.edu/N/N07/N07-2011.pdf	84	Exploring Affect-Context Dependencies for Adaptive System Development	about across acts analysis analyzed answers applied architecture arti assessment associated atlas audio australia automatically barry based between both bratt cant cantly cause challenges cial clark coach cognition cognitive collaborative communication computer computers contexts continuous cornelius correlations cowie critical current dependencies describing design detecting develop dialog dialogs dialogue dialogues directions discourse during education embedded emnlp emotional emotions empathy engineering essay events examine exploiting expressed forbes frustration graesser gravity health human hypotheses ieee instruction intelligence interacting interactive international jordan journal klein language learning litman magliano march monitorable moon more murray narayanan natural naturalistic next only pages patterns performance person peters physics picard plan predicting proc proceedings processing psychology qualitative questions references repetition responding responds responses results riley rose rotaru schultz signi siler similar some speech spoken states step structure student such sydney system systems table that theory these this topic towards tracks transactions transition tutorial tutoring uncertain uncertainty user vanlehn variable will with workshop writing
http://acl.ldc.upenn.edu/N/N07/N07-2008.pdf	81	A Fast Method for Parallel Document Identification	acquisition alexandre aligned automatic available based bilingual bottleneck building canada canadian cation chen comput corpora corpus craig cross csail daniel data diab dictionaries disambiguation documents download ecir europarl evaluation franz from germann hansard hansards http iadh identi jian jiang josef koehn langlais language light linguist linguistic machine marcu mcewan method mining mona multilingual naacl natural noah ounis pages parallel parliament patry people philip philipp philippe phrase proc references release relieving resnik resources riao ruthven sense smith statistical tagging text translation ulrich unsupervised using with without word
http://acl.ldc.upenn.edu/N/N07/N07-1071.pdf	71	ISP: Learning Inferential Selectional Preferences	accept accepted accuracy accurate achieved acquisition across algorithms alternations analysis answering antonymous antonymy applications approach arguments artificial assumption asymmetric australia average baker barcelona barzilay based because behavioral berkeley better between both call canada cases castellan categorization chicago clark classes cogex cognition cognitive coling columbus common communications computational computing conclusion constant constraints coppola corpus cover cutoff dagan database dirt discovering discovery distributional domain eacl edition edmonton electronic elements emnlp encyclopedia engineering english entailment entire entities erlbaum errors expense explored extracting extraction fairly false family fass fellbaum figure fillmore filter filters finally first fodor form framenet france from generally generated generic greiff harabagiu hence here hickl hill hillsdale history human hypothesis identical impact include incorrect increase induction inference inferences inferential information infrastructure inspected intelligence interestingly investigating investigation investment italy john katz knowledge kouylekov language large lavelli learn learning lenat levin lexical light lloyd logic lowe made maiorano manually many match mathematics maximum mcgraw mckeown methods minimum missed model models moldovan montreal more most naacl natural nearly nonparametric occurs open overall overgeneration pantel parallel paraphrase paraphrases parsing pazienza pennacchiotti performance permissive place positives preference preferences preliminary presented press problem proceedings produced project prover quandary question range ranking rather realization recognized references rejected relation relations remained resnik results romano rosch rule rules same scale scaling science sciences second section selectional semantic semantics senses sensitivity sets several shapiro shorter shown siegel similar since sons source spain specificity statistical statistics staying step stepped strategies structure such suffer sydney system szpektor take tanev test text textual than that theoretic theory these they this thomas those threshold toulose trento types typically university used using value values verb verbs version were what wiley wilks will with within without word wordnet zanzotto
http://acl.ldc.upenn.edu/N/N07/N07-2032.pdf	105	Subtree Mining for Relation Extraction from Wikipedia	accumulates accumulation adequate advances agichtein also although analysis anne annotate answering applications based baseline baselines basis best between betz brin bunescu burges calculates capture carry category cation chasen chooses chua ciently classi collected collections compared conference coreference culotta data databases degrade dependency detection develop difference different digital discover during each editors effectiveness entities entity error evaluate evaluation experiments external extracting extraction extractor fact feature features finally focuses forest four freqt frequent from generate gives graphs gravano highest hovy http improvement instances integrating interesting international irrelevant joachims kernel keyword label language large learning levels libraries light like long making manually many mccallum method methods mined miner mining minipar model models mooney more morton naacl natural necessary obtained only order over pages pair parameter parser parsing passage path paths patterns performance pipeline plain polynomial poor poteet practical precision preliminary press principal probabilistic proceedings processing proposed question quite randomly ravichandran reason recall references relation relations relationships report resources respect result retrieval reveals scale scholkopf score scores second secondary select selection sentence sequences shows sigir sigkdd smola snowball software step steve substantial subtree subtrees suffers support surface system systems table testing text that their they those trace trained training tree trees types used using values vector when which wide wikipedia with word words works workshop world zaki
http://acl.ldc.upenn.edu/N/N07/N07-2027.pdf	100	Tagging Icelandic text using a linguistic and a statistical tagger	able about accuracy achieve additionally advances algorithms along analyzer angelova anttila appeared applied approach automatic auwera average based been berlin bonev borovets brants briem bulgaria calculations call chapter cius classes combination combinations compared comparing complex components computational computer conclusion conference consists constraint corpus corresponding data department development dictionary does driven dublin each eacl editor editors effort errors european evaluation experiment facilitate fewer finland fintal first following forlag found frequency further future generate generated german germanic germany ginter given global grammar gruyter guessing hahn hand have heikkil helgad hence heuristics high holmboe iceland icelandic icemorphy icetagger improve improving increase increased indeed individual induction infrequent insson institute integrations international ireland karlsson label language languageindependent languages learning less lexical lexicography lexicon like linguistic linguistics lled lling local loftsson logical london made madrid magn make makes manner marked mechanism methods mikheev missing months morphological morphologically morphy mouton museum nakov natural nordisk nouns obtain obtained often only other overall pahikkala paper parsing part pind probability probably proceedings processing pyysalo recent record references report resources resulting reykjavik routledge rule rules salakoski samuelsson science seattle secondly seems shef since spain speech sprogteknologi sson state statistical stochastic such syntax system table tagger taggers tagging tags technical testing text than that this through took training ttir turku tusculanums university unknown unrestricted used uses using various voutilainen were when with word work worth would
http://acl.ldc.upenn.edu/N/N07/N07-2028.pdf	101	Efficient Computation of Entropy Gradient for Semi-Supervised Conditional Random Fields	accuracy active algorithm allows based been bengio calculating calculation cation cient classi coling computation computer conditional crespi crfs cybenko data duplicate dynamic elds entity entropy experimental forward given grandvalet greiner hernando hidden icml ieee improved information jiao label labeling lafferty learning like literature loose machine madison many markov mccallum method minimization model models naacl named nips notably observation opposed output pages paper pass pereira predominantly present probabilistic proceedings program proposed random recognition references related report reporting results schuurmans sciences segmentation segmenting semi sequence similarities since some song structured subsequence supervised survey technical theory this training trans university unnecessary used wang which wisconsin with work
http://acl.ldc.upenn.edu/N/N07/N07-3003.pdf	132	Creating a Knowledge Base From a Collaboratively Generated Encyclopedia	aaai acquisition application approach automatic baldwin banerjee based bicknell blettner budanitsky cambridge chains charles checking classes cognitive coling common compounds computational computing concept conference constrained context contextual coreference corley corpora corpus correction correlates cybernetics database detection development disambiguation distance domingos electronic entailment evaluating exploiting fellbaum finkelstein framework from gabrilovich generalized greenwood hearst hirst hyponyms icos ieee ijcai ijcnlp induction inference information interpretation jacquemin jarmasz knowledge kohomban labeling language large learning lexical light linguistics links logical machine malapropisms mass matias mccarthy measures mechanization metric mihalcea mili miller monz morin naacl nets noun onge output over pattern patwardhan pedersen phrases placing ponzetto poon press proc proceedings processes programs projecting punyakanok rada ranlp references relatedness representations resolution revisited richardson rijke rivlin roget role roth ruppin search semantic semantics sense senserelate similarity singla solan soon statistical stevenson strapparava strube systems szpakowicz targetword teddington text thesaurus thought transactions unifying using weight wikipedia wikirelate with wolfman word wordnet zimak
http://acl.ldc.upenn.edu/N/N07/N07-1005.pdf	5	Automatic Evaluation of Machine Translation Based on Rate of Accomplishment of Sub-goals	adequacy akiba alberto alex alon amigo analysis annotation anthony applications applied arabic assessment automatic automatically babych banerjee blatz bleu bogdan branches cance cation center chin chine chinese chiori coling conference correlation criteria cyril data dence developer distance distances draper edit edition eiichiro emnlp english enrique erin error estimation evaluating evaluation extending extrinsic fast final fitzgerald fluency foster franz frequency from gandrabur george gimenez goutte gram gregor harry hartley hermann hitoshi hopkins hori http human ikehara imamura improved inside intrinsic isahara iwslt japanese jeida jesus jing john johns josef joseph jsai judgments kenji kentaro kishore koehn kulesza language lavie leusch linguistic lrec luke machine measure measures melamed meteor method metric metrics minimum multiple nicola niessen nist norman novel occurrence ogura orange output package pages papineni philipp point proceedings processing projects qarla quality rank references regression report research rouge roukos salim sanchis satanjeev satoru satoshi sets shen shirai signi simona smith sonja speci speech statistical statistics string sumita summaries summarization summer summit systems technical test tests text tides todd tool training transactions transassess translation translations turian university upenn using view viii ward weightings wiley with workshop yasuhiro
http://acl.ldc.upenn.edu/N/N07/N07-1047.pdf	47	Applying Many-to-Many Alignments and Hidden Markov Models to Letter-to-Phoneme Conversion	acknowledgements alan albert alberta alexander algorithm algorithms aligning altun analogy analysis annual antal applications approach arthur association baayen bartlett based bazin black bosch building canisius cation celex chapter charles cherry colin collins communication complex computational conference constraint conversion correction csli daelemans damper data database david dempster dennis discriminative distance donald edit edits eighth emnlp empirical engineering english eric esca european experiments from general grapheme group guide gulikers harald hidden hofmann icml ieee improve improved improving incomplete independent inference instance intelligence interest international ioannis issues jakub john joseph journal june kevin kibler kristina kruskal laird langauge language learn learner learning lenzo leon letter lexical like likelihood linguistics machine machines macromolecules marc marchand markov marsters maximum meeting members memory methods michael modeling models moore morpho morristown multistrategy natural networks oriented other pagel pages parallel pattern paul perceptron peter phoneme phonemes phonological phonology piepenbrock proceedings processing progress pronounce pronunciation publications reference references report research richard ristad robert rosenberg royal rubin rules sander sankoff satisfaction sejnowski sequence series sigphon sloot society sound special speech spelling springer statistical string support susan sven syllabi synthesis systems taylor technical technology terrence text thank that theory third thomas tilburg timbl time tophoneme toutanova training transactions tsochantaridis university using vector version vincent walter warps with workshop would yannick yasemin yianilos york zavrel
http://acl.ldc.upenn.edu/N/N07/N07-2055.pdf	128	A Semi-Automatic Evaluation Scheme: Automated Nuggetization for Manual Annotation	accomplished agreements annotator answers applications automatic automatically automation based close collins conclusion concrete conference content created currently data definition demner deploy description design dissertation document driven emnlp environment evaluating evaluation exercises following from fullscale fushman generated given goal gold grade gram hamly have head high hovy html implementation index information ings insure inter into judgments language look machine main marton measured method meticulous models naacl natural nenkova nist nlpir nugget nuggeteer nuggets occurrence paper passonneau peer pennsylvania preparing procedure proceed proceedings processing projects provided pyramid quality questions radul rambow ranlp reference references selection semiautomatic standard statistical statistics summaries summarization summary systematic them these this took understanding university using various very
http://acl.ldc.upenn.edu/N/N07/N07-1001.pdf	1	Exploiting acoustic and syntactic features for proso dy lab eling in a maximum entropy framework	aalborg abstract accent acoustic adell aguero almost altun analysis ananthakrishnan annual approach april arti association audio automatic bangalore based batliner beckman beijing berger black bonafonte borys boston both brenier budapest bulyko caley cambridge chavara chen china choi cial classi cohen cole college communication components computational computer concatenative conditional conference conkie context conversational corpus coupled detection direction discourse elds emnlp emphatic english ensemble entropy eurospeech events features festival festvox france from generation giving gmmbased gregory hasegawa hirschberg http hufnagel hungary icassp icslp ieee information informativeness intel inter intonation intonational johnson joint joshi june jurafsky kompe labeling labels language large learning lehiste lexical ligence ling linguistic linguistics machine march margin maryland maximum mcgory mckeown meeting model modeling models monologues multistream nakatani narayanan natural naturalvoices news niemann ostendorf pages park parsing patterns philadelphia pierrehumbert pietra pitch pitrelli predict predicting prediction press price proc proceedings processing prominence prosodic prosody radio rambow random recognition recognizer references reliability report representation riccardi rose ross scaling segments selection shattuck shen shimei silverman simultaneous speech spoken standard supertagging suprasegmentals syntactic synthesis synthesizer syrdal system taylor technical text tilt tion tobi toulouse transactions transcriber transla tree understanding unit university using utterances verbmobil voices volume wightman word words yoon zhang
http://acl.ldc.upenn.edu/N/N07/N07-1048.pdf	48	Analysis of Morph-Based Speech Recognition and the Modeling of Out-of-Vocabulary Words Across Languages	aalborg acquisition adaptive agglutinative algorithm alumae application applied arabic australia automatic backoff based bazzi beijing berton bilmes bisani bntu book brent brietzmann broadcast byrne cambridge challenge chen china cient coling communication complexity compound computational computer contextual continuous corrective creutz czech darpa decoder decomposition denmark dependencies discovery diss dutch ected ectional emnlp empirical entropy eurospeech extensible fetter finke finnish galescu geneva german geutner glass goldsmith goldwater goodman gram grif growing hajic hall helsinki hessen highly hirsimaki http hybrid icassp icslp improved induction inquiry interspeech into ircing isbn jelinek johnson jong khudanpur kirchhoff koehler korean kurimo kwon labs lagus language languages lansdowne large larson learning lexical library limited linguistics lisbon machine marcken model modeling modelling models mohri morph morpheme morphemes morpho morphology multilingual naacl natural news ollason open ordelman park parliamentary particle pellom performance philadelphia pittsburgh portugal probabilistically proc projects pruning psutka puurula pylkkonen recogni recognition recombination references regel research rigoll riley rissanen robust saraclar scheytt science scienti seattle segmentation series shafran sigphon siivola smoothing sound speech speeches splitting srilm statistical stochastic stolcke study switzerland sydney system systems techniques technology thesis tion toolkit tools transcribing unit units university unlimited unsupervised valtchev varjokallio vergyri version virpioja vocabularies vocabulary whittaker willett with woodland word words workshop world york young
http://acl.ldc.upenn.edu/N/N07/N07-1061.pdf	61	A Comparison of Pivot Methods for Phrase-based Statistical Machine Translation	about additional address adria again alex alexandra aligning alignment also although amittai amount amta analogy andreas anoop appeared approach approaches arise arti articles association automatic available axelrod based beam below better between bilingual birch bleu blue boitet bridge bridging brings brown burch callison cance cantly catalanenglish charles chris christian christof cial cient city classes coling compare compared comparison computational conclusion conditions conducted conference conll cons consistently constructs controlled conviction corpora corpus cross daniel david decisive decoder della described description determining developing direct directions directly discriminative diverse divided document dragomir dragos eacl edinburgh editors emphasis emphatically english enormous entropy error estimation europarl europe european evaluating evaluation even experimental experiments exploiting extensible fail features foris framework franz fraser french from fugen further future german gildea gispert gollins great guiding haifeng halves harold have heart hermann highest hitoshi hltnaacl hope human icslp implicitness importance improved improving including incorporated indicates inducing information insisted institutions intelligence intend intended intergovernmental interlingua international into investigate isahara issue iwslt jain japanese jose josef june katherine kauers kenji khudanpur kishore klaus koehn kumar language languages large lexicons libin limited linguistics lrec machine main makoto manual manuel marcu mari mark masao materialise mathematics maximum maxwell mayne measures mechanical mercer method miles minimum minority modeling models monz morphosyntactic much multilingual munteanu must naacl nagao namely nato never news nirenburg noah ones opinion opportunity osborne other otherwise outperformed outputs pages pairs paper papineni parallel parameter paraphrases parliament performance peter pharaoh philip philipp phrase phrasetrans pietra pivot point possibility poster press principle proc proceedings promising pros proves public published quite radev rate readings real reference references reform rejection rejects relative relatively reliable remaining research resnik resources results retrieval robert role roukos salim same sanderson sanjeev sarkar scarce sceptical scepticism schafer schubert scores scoring search section selects sentence sentences separately sergei serious sessions shankar shen should showed shows sigir signi similarity situation situations smith smorgasbord snttrans some somers sonja source spanish speech srilm statistical statistically stefan step stephan stephen stolcke strategies strategy subject summit surprised symposium system systematic systems table talbot target template than that then these third this those through todd toolkit toon towards train trained training transfer translates translation translations treaties triangulated turn types used uses using utiyama various very vincent viren vogel waibel wang ward weijing well were when where wilks will with without witkam word work workshop yamada yarowsky yorick york zhanyi zhen
http://acl.ldc.upenn.edu/N/N07/N07-3001.pdf	130	Query Expansion Using Domain Information in Compounds	able ahlgren alan baeza based been berthier bethesda biomedicine book boras brake broms cambridge cloth cognitive collection college combination compound compounds croft cross cruse database decide demo department developing dictionary dimitrios domain effectiveness effects evaluation expansion finland from full future germany goteborg hand have headings hedlund http indexing information institutet international january jena karolinska kokkinakis language library lingustics medical medicine medlex mesh mining modern national needed neto parts presented press proceedings publications query references report research resources resurser retrieval retrieved ribeiro ricardo semantic should sity smbm spraakdata sprakdata stockholm strategy strategyquery studies subject svedk svensk swedish symposium tampere technical term test text this towel true turid unfortunately univer universitetsbiblioteket university used valfrid volume when william work yates york
http://acl.ldc.upenn.edu/N/N07/N07-2020.pdf	93	ILR-Based MT Comprehension Test with Multi-Level Questions	analysis arabic arpa association automatic based calandra clare conf conference connell descriptions dlpt doddington engines english european eval evalplan evaluation gale govtilr gram http intel interagency jones language level machine measuring methodology nist occurrence plan proceedings program quality references roundtable shen skill speakers speech statistics task tate testing tests translation using voss website white with workshop
http://acl.ldc.upenn.edu/N/N07/N07-2037.pdf	110	Joint Morphological-Lexical Language Modeling for Machine Translation	about additional afify alignment allows amta analysis analyzer approach arabic automatic available based beam berger besacier bleu broadcast brown buckwalter building called chen chollet choueiter coling comp comparison computational conclusions conf constraints continuous decoder decomposition della demonstrate dialectal directed discriminative echelotte encouraging entropy eurospeech evaluation france future gauvain ghaoui gram hawaii honolulu huang icassp improve improved improvement information integration interspeech isbn items jmllm joint kirchhoff knight koehn language lexical linguistics lisbon lvcsr machine makhoul march marcu mathematics maximum method model modeling models mokbel morpheme morphemebased morphological naacl natural news nguyen over pages papineni parallel pennsylvania pharaoh phrase pietra pittsburgh portugal povey presented proc processing provides quality recognition references relabeling results roukos sarikaya schwartz schwenk search segments sentence sources space speech statistical syntax systematic technique text that tight toulouse towards training transcription translation trees university using various version ward will work workshop xiang yang yvon zweig
http://acl.ldc.upenn.edu/N/N07/N07-1050.pdf	50	Incremental Non-Projective Dependency Parsing	abeille afonso algorithms analyzer annotation asahara available based bick black bohmova brants buchholz building chang chapter cheng chinese cjlin conll cormen corpora csie darpa dependency deterministic dipper dordrecht floresta grammars haber hajic hajicova hansen history hladka http ijcnlp introduction jelinek kluwer lafferty language leiserson level lezius library libsvm lrec machines magerman marsi matsumoto mercer models multilingual natural parsed parsing portuguese press probabilistic proc references richer rivest roukos santos scenario shared sinta smith software speech structure support task tica tiger towards treebank treebanks using vector workshop
http://acl.ldc.upenn.edu/N/N07/N07-2041.pdf	114	Simultaneous Identification of Biomedical Named-Entity and Functional Relations Using Statistical Parsing Techniques ∗	acab acids also analysis annotation april articles automated bacteria based best bies bikel bioinformatics biomedical blum boston brady brinkman cantly charniak coarse current data database defays discriminative discussion distributional documents donnes emnlp entities examples experiments exploiting extract extracted extraction from fully function further future gardy gold graphical have hoglund hunter identi implemented improved increasing infer inferring information input integrated integrating introduced johnson kohlbacher kulick laird lambert large lexical lexicalized libeman localization localizations made mandel maxent mcdonald medline method miguel miller model molecular multiple naacl nair named novel nucleic outperforms output pages palmer paper parser parsing parsingbased plan poster prediction predictive proc protein psortdb ramshaw references relation relations reranking research results rich rocheford rochester rost sarkar schein score semi sentences sequence shatkay show signi statistical subcellular supervised syntactic take takes terms text that them this through track train ungar unlabeled volume weischedel which with work
http://acl.ldc.upenn.edu/N/N07/N07-2049.pdf	122	Combined Use of Speaker- and Tone-Normalized Pitch Reset with Pause Duration for Automatic Story Segmentation in Mandarin Broadcast News	acoustic across america analysis automatic based because boundaries boundary broadcast cannot chen cheveigne clearly combined comm computational consortium cues data different discriminated duration estimator fluent framework frequency from fundamental future greatly hakkani http integrating into investigate investigated journal kaufmann kawahara learning levow lexical linguistic linguistics machine mandarin modeling morgan motivates music naacl news normalized pairs paper pause pitch proc programs projects prosodic prosody quinlan references reset segmentation sentences shows shriberg society speaker speakerand speakernormalized speech stolcke story summary syllables that this tone topic topics tseng upenn utterance varies wang work
http://acl.ldc.upenn.edu/N/N07/N07-1066.pdf	66	A Probabilistic Framework for Answer Selection in Question Answering	allan amsterdam answer answering bolivar boosting bunescu buscaldi clarke clef conference cormack database detection electronic engines evaluation exploiting falcon fellbaum from grju harabagiu international jijkoun knowledge language level lexical lynam mihalcea mining moldovan morarescu notes novelty pasca press proceedings question rantwijk redundancy references resources retrieval rijke rosso sang sentence sigir surdeanu task tjong trec university wade wikipedia wordnet working
http://acl.ldc.upenn.edu/N/N07/N07-1058.pdf	58	Combining Lexical and Grammatical Features to Improve Readability Measures for First and Second Language Texts	adelson annual assessment association baltimore bates biology bizzi bonner brain calissano categorization choice clarke computational conference della development dobson dulbecco edition enciclopedia examination focus fondata frontiers fuchs giovanni goldstein grammar homo howard information istituto italiana jacob language level levi linguistics machines meeting methods michigan models montalcini nature nurture ostendorf oxford pearson press proceedings read reader reading references reflect research retrieval rome sapiens schwarm sigir silberstein statistical support text trecanni university using vector volterra westheimer yang
http://acl.ldc.upenn.edu/N/N07/N07-1046.pdf	46	A Log-linear Block Transliteration Model based on Bi-Stream HMMs	abduljaleel adam alex algorithms alignment almut americans amta andreas annals annual approach approaches arabic arbabi arbor asif association australia automatic bandyopadhyay barcelona bart based beam berger bilingual bing bleu bonnie brown building canada channel cheng chinese chiori choi christoph cluster clustering cohesion coling comparison computational conf conference considering context contextual copenhagen cross darroch decoder della denmark denver development edmonton ehara ekbal elizabeth emnlp empirical english entity entropy eric estimation evaluation extensible features figure fischthal fourth franz generalized google goto graehl haizhou heidi hermann hewavitharana hildebrand hoon hori http huang incorporation information international intl isao island iterative iwslt jeju jian jing joint jonathan jong journal july june kato kevin khudanpur kishore knight knowledge koehn korean language languages larkey leah learning linear lingual linguistics lingustics louisiana machine madrid management mansur march mathematical mathematics matthias maximum mercer method methods michigan model modeling models modi montreal multi name named names naoto naskar nasreen natural noriyoshi november october onaizan orleans pages pair paola papineni parallel parameter peter pharaoh philadelphia philipp phrasal phrase pietra proc proceedings processing pronunciation proper quebec ratcliff recognition references research retrieval robert roukos rules salim sanjeev sanjika scaling scott search semitic sighan silja source sourcechannel spain speci spectral spoken srilm stalls statistical statistics stephan stephen stolcke summit system systematic taipei taiwan technical template terms terumasa text texts tillmann todd toolkit translate translating translation transliteration uratani using vancouver various vincent virga vogel volume waibel ward washington with word workshop xing yaser zhang zhao
http://acl.ldc.upenn.edu/N/N07/N07-2054.pdf	127	Speech Summarization Without Lexical Features for Mandarin Broadcast News	aaai accuracy achieve acoustic alone also amsterdam analysis another approach aspects automatic aver average barzilay baseline boersma broadcast chang chen chih chinese chung communication compared comparing comparison competitive computer conclusion conference consistent demand different discourse distribution document doing draw elhadad empirical english eurospeech evaluation even evidence experiments expo features figure fmeasure found from functional furui good have hidden higher hirschberg holds hori hsieh huang icassp icme ieee important improvement information inoue institute intelligent international interspeech jing koumpis language latent lexical library libsvm machines mandarin markov maskey mckeown measure messages methods mikami models multimedia naacl nding news pages paper performance phonetic phonetics placing play praat presented probabilistic proc proceedings processing produced produces program prosodic prosody recognition references relatively renals report result results role same sciences semantic sentences shows similar sources speech spoken stringent structural study summarization summarizer summarizing summary superior support symposium system table text than that therefore these this topical transactions tslp university using vector version voicemail weenink with without yamashita yields
http://acl.ldc.upenn.edu/N/N07/N07-1055.pdf	55	A Unified Local and Global Model for Discourse Coherence	alexander althaus analysis annual applications approach aravind assoc association barbara barcelona barzilay based beal best carl catching centering charniak coarse coherence coherent computational computing conference content discourse discourses discriminative drift edward entity ernst eugene foltz framework generation ghahramani grosz hidden johnson joshi karamanis kintsch koller landauer lapata latent lillian linguistics local locally main mark markov matthew maxent measurement meeting mirella model modeling models naacl nikiforos nips nite pages parsing peter probabilistic proc proceedings processes rasmussen references regina reranking scott semantic summarization textual thomas walter weinstein with zoubin
http://acl.ldc.upenn.edu/N/N07/N07-1022.pdf	22	Generation by Inverting a Semantic Parser That Uses Statistical Machine Translation	april arbor automatic available bangalore based belz brown carroll cation chart chen chiang ciency cient cliffs comparing compiling computational copestake della eacl englewood estimation evaluation ewnlg flickinger france generation generator grammar grammars hall hierarchical high http human ijcnlp inlg island israel italy jeju july june korea later lexicalist linguistics machine manual mathematics mercer metrics mitzpe model october oepen pages parameter parsing phrase pietra poznanski prentice proc projects rambow ramon realization references reiter robocup semi server soccer sourceforge sserver statistical systems theory toulouse translation trento ullman users version whittaker widecoverage
http://acl.ldc.upenn.edu/N/N07/N07-1049.pdf	49	Tree Revision Learning for Dependency Parsing	algorithms approach attardi automatic available based between biology brill buchholz chang charniak ciaramita cjlin coarse collins computational concepts conll connl corston crammer csie dependency discriminative driven emnlp error experiments fine free gangemi grammar hidden http icml ijcai induction introduction johnson journal language learning library libsvm linguistics machine machines markov marsi maxent methods models molecular multiclass multilanguage multilingual natural nbest oliver online ontology parser parsing perceptron problems proceedings processing projective ratsch reference references relations reranking research rojas saric semantic shared singer slovene software spanish support swedish task text theory training transformation ultraconservative unsupervised vector with
http://acl.ldc.upenn.edu/N/N07/N07-2048.pdf	121	Modifying SO-PMI for Japanese Weblog Opinion Mining by Using a Balancing Factor and Detecting Neutral Expressions	academic accuracy achieved added adding after analysis annual anthony applied applying approach approaches araki atad balance balanced balancing based basic between cascaded cation cations chaovalit characteristics chunking classi collecting comparatively comparison conclusions conference contrast corston customer data deal decision dependency described detection does done down eifidom emnlp eric etzioni evaluative evitag evitagen evitisop expanding experiment expressions extracting extreme factor features finally free from fukushima gamon guangwei hicss high hsilg ieice improved information intelligent inui ipsj japan japanese journals kenji kentaro kobayashi kudoh leaning lina maria matsumoto meeting michael mining modi morf movie must napaj near negative neutral nigiro nomrah note notes nozomi obtained oliver operator opinion opinions oren orientation original parts peter phenomenon pimwadee poor popescu positive preliminary presented proceedings processing product proposed pulse quite reference references result results review reviews ringger section semantic sides simon society some structure study such supervised support system taku tateishi technique technology test text this three thumbs thus toshikazu turney unsupervised using very wang weblog well whether while with word work yuji zhou
http://acl.ldc.upenn.edu/N/N07/N07-1057.pdf	57	Multilingual Structural Pro jection across Interlinear Text	aaai across allows analysis bailyn based bonnie brown charniak cohesion computational conference context coordination csli current daniels description dislocation divergences dorr editor emnlp engine estimation eugene feature formal found frederick free gerhild grammar heidi hpsg http international inversion issues john kind linguistics linguists machine mathematics mercer michael neutrality optionality osed parameter parsing pennsylvania peter philadelphia phrasal pietra proc proceedings prop publications references reminiscent resnik russian search slavic solution statistical statistics stephen structural text this translation umiacs unlikes vincent which with word zybatow
http://acl.ldc.upenn.edu/N/N07/N07-1026.pdf	26	Data-Driven Graph Construction for Semi-Supervised Graph-Based Learning in NLP	abstraction active algorithms alpaydin analysis approach aren balcan based bengio cald carnegie cascading categorization cation chen chew chia chklovski classi cluster clustering complementarity cuts data disambiguation dong ducharme education elisseeff emnlp empirical english entropy erkan evaluation extraction features feedback finding focused from ghahramani giuliano gliozzo goldberg graph graphbased graphs grozea haghighi high hong icml identi image images inference irst kaynak kernels killgariff knowledge kybernetika label labeled labeling language large learning leslie lexical machines manning many matching maximum mellon mihalcea minimum model mohammad multiple naacl natural neural nips noble ofspeech optimal otterbacher pages pang parameter part partially pattern pedersen performance person probabilistic proc proceedings processing propagation protein question radev random ratnaparkhi references relation report retrieval robust sample seeing semi semisupervised sense senseval sentence sentiment sentimental sequence settings siglex similarity simple sources stars strapparava subjectivity subspace summarization supervised support syntactic syntalex tagging task technical term textual there thesis training university unlabeled unsupervised using vector vincent vocabulary walks webcam weston when with word workshop xiaojin zhang zheng zhou
http://acl.ldc.upenn.edu/N/N07/N07-2047.pdf	120	Kernel Regression Based Machine Translation	achieved algorithm although amta applied approach approximating based beam capability cation class classi comparable complexity conceptual contribution corpus cortes could cristianini decoder demonstrated discussions error europarl expensive extensible face from general give gram hernandez hltnaacl icml icslp inversion kernels koehn language larger learn learning linear lodhi mach machine main margin matrix maximum method minimum model modeling models mohri much multiclass multiview operation operators orientation parado parallel pascal performance pharaoh phrase practice previous problem proc proposed rank rate reduced references regression report saunders scale search selection sentences several shawe solution southampton space speed srilm stage statistical still stocke string subset summit szedmak tasks taylor technical technique techniques text than their this thousand tillmann time toolkit training transductions translation unigram using watkins weston will with work
http://acl.ldc.upenn.edu/N/N07/N07-1031.pdf	31	Automating Creation of Hierarchical Faceted Metadata Structures	aaai abstraction acknowledgements acquisition additions adjectives ageing aging aldo algorithm allocation allowed although andrew anick another application applied approach april architects architecture assessing assigned assignment assistant associates assumed athens automated automatic based before bene berkeley bernardo biomedical blei brain browsing bruce called captured caraballo castanet categories categorization category caution christiane cluster coarse codes coling collection collections compared computing concept conclusions conference construction cooking corpora could coverage craig creates creation croft currently cutting data database dataset dave david degree deriving descriptions designing digital dirichlet disambiguated discussed displayed document documents domains done douglas east editor electronic emilia ence engineering even example exhaustive extending fabrizio facet faceted fagan fails feedback fellbaum figure findex flairs flamenco fort from funded fussing future gangemi gather generated generation grained greece have headings hearst help here hierarchical hierarchies hierarchy history hofmann http hypernym hyponyms ijcai image important improve improvements included including incorrect information integrating intelligent inter interface into iterative john jordan journal july kaki karger kevin kirsten knowledge labeled large larry latent lauderdale learning leaves level lexical lexically libraries likely louis lrec machine magnini manning mark marti matched mechanism medical mesh metadata michael mihalcea mika model moldovan more morphological morville move naacl nally navigli nding nearly needed nevill noun obtained occur often ontology organizing orlando other paola paraphrase part partial participants paynter pedersen peter piek piles ping pratt presented press principles proc processing procs program promote questionnaire quite rada ranking rashmi recipes recognize references reilly release remove required research resources result retrieved revealed roberto rosenfeld same sanderson scale scatter search sebastiani second seeking seen sense several sharon should sigir sinha sites some spelling stir stoica stolckholm stroudsburg subject subsumption successfully such summer superb surveys susesh swearingen systematicity systems table tags taxonomies technical term terminological terminology terms text than thanks that their themselves this thomas thought tipirneni titles topic translation trees trimming tukey undergraduate unknown unsupervised useful users using variations velardi verbs view vossen wanda weinberger when wide with within witten wordnet work workshop world
http://acl.ldc.upenn.edu/N/N07/N07-2007.pdf	80	Combination of Statistical Word Alignments Based on Multiple Preprocessing Schemes	aaai abraham ackerman across aligned aligner alignment alignments american amta analysis approach arabic arti association ayan based cial cohen college combination combining communication comparison computational condon conference corpora cross disambiguation discriminative dpbased emnlp english entropy features fell fourteenth fraser freeman habash hltnaacl improved improvement inducing intelligence ittycheriah julien klein koehn lacoste learning linear linguistic linguistics link lopez machine marcu maryland matching maximum models monotone morphological multilingual naacl name ngai park partof personal phrase phrasebased preprocessing projection rambow references resnik robust roukos rules sadat schemes search semi speech statistical supervised swoop systematic tagging taskar techniques text thesis tillmann tokenization tools training translation trees university using valued various vogel what wicentowski with word yarowsky zubiaga
http://acl.ldc.upenn.edu/N/N07/N07-4004.pdf	144	POSSLT: A Korean to English Spoken Language Translation System	accuracy adding advance airport alignment amta annual approach associative automatically based beam been benefits besides built cascading categories changing chen china coling consists conversion corpus coupling coverage decide decoder decoding deleting depends developed divided domain donghyeon dynamically english eojeol extensible extracted fast features gary geunbae geveva grapheme have higher hongkong hotel icslp improved improving increase information input integrating interspeech into jeju jinsik jonghoon kikui koehn korea korean language lattice loaded loads machine many meeting model modeling models more morpheme morphemes needed obtained october order page pharaoh phoneme phrase preprocessor proc quality recognition references rules saleem schultz search selection seungwon several shopping situation soong spacing speech srilm statistical stolcke such system systems task techniques than there tighter toolkit tospeech translation travel typically unified unit units unloads used useless user uses using utterances vogel washington watanabe which word words yamamoto zhang
http://acl.ldc.upenn.edu/N/N07/N07-1060.pdf	60	Automatic and human scoring of word definition responses	aarts acad analysis balass bolger breadth british burnage burstein cesses classify computerized conf contextual corpora corpus depth design discourse duction dunlop encoding engl english exploitation foltz free haan integrating intl into intro kintsch kluwer landauer landen lang language latent learning lexical lexicons meanings models national nijmegen nitions oostdijk perfetti press references relationships responses semantic sigir techniques using variation wolff word words
http://acl.ldc.upenn.edu/N/N07/N07-2053.pdf	126	Selective Phrase Pair Extraction for Improved Statistical Machine Translation	achieved against alexandra alignment alignments american americas amoung andreas annual approach association attempts based beam best beyond bilingual birch bleu bode both building burch callison cases chapter chris ciency closely comparison competitive computational conference consistently constraining constraints create daniel data decoder denero development difference directly discriminative discussion driven empirical equivalence error evaluation exercise experiments figure foster franz from full generative george gillick growing growth heldout heuristics higher howard improved included incorporates increase james john johnson join joint josef klein koehn kuhn language learning length limited line linear linguistics linking local machine marcu matched maximum meeting melamed methods mihalcea miles minimum model models moore most naacl natural north opment original osborne outperforms pair pairs parallel pedersen pharaoh philipp phrase phrases phrasetable points possible probability proceedings processing quality rada rate recent references related results robert roland scaling score scored search seems selected selection sentence shows sixth size slower smoothing stastical statistical surface system table tables test tested texts that these this trained training translation under underperform unseen using varied verify were while william with wong word words work workshop zhang
http://acl.ldc.upenn.edu/N/N07/N07-2013.pdf	86	Document Similarity Measures to Distinguish Native vs. Non-Native Essay Writers	academic addison analysis answer automatic available average background behavior bleu campbell center characterizes chodorow compositions computer computers cosine described discourse discussion easier education evaluation foltz from future harder humanities information integrated introduction laham landauer language latent leacock lecture less machine maximum measure method more native others paper papineni populations processes processing prompt questions rater reading references relying report research retrieval roukos salton scoring semantic short show significant skillful speaker speakers students successfully task technical text than this transformation translation trend ucla university ward weley with words work writing
http://acl.ldc.upenn.edu/N/N07/N07-1030.pdf	30	Joint Determination of Anaphoricity and Coreference Resolution using Integer Programming	aggregation algorithm andrew appelt approach argument barzilay based bell berger computational coreference della emnlp entropy extraction frequencies generation hongyan information interpretation ittycheriah jing kambhatla kehler language lapata linguistics maximum mentionsynchronous mirella naacl nanda natural pages partitioning pietra predicate probabilistic proceedings processing pronoun references regina resolution roukos salim simma taylor tree utility xiaoqiang york
http://acl.ldc.upenn.edu/N/N07/N07-1045.pdf	45	Near-Synonym Choice in an Intelligent Thesaurus	agirre american annotation annual answering association automatic canada chapter charles choosing church clarke coling computational conference content context david decision development disambiguation document edmonds edmonton egidio eneko estimates exploring factoid frequency hanks hltnaacl human information intelligent international kenneth language lexical lexicography linguistics lists luxembourg madrid martinez measures meeting most mutual nancy network norms north occurrence pages passage patrick philip proceedings question references research retrieval saarbrucken semantic sense sigir similarity spain statistical technology terra toronto typical using with word workshop
http://acl.ldc.upenn.edu/N/N07/N07-1007.pdf	7	Generating Case Markers in Machine Translation	aaai according achieved additional adequacy agreed agreement alignment also among annotators applicable articles automatic based baseline being believe best better between bleu both cambridge candidates case caused causing center chander cherry chiang choosing clarkson clause coling collins compared comparing conclusion conditional context contributes cover degradation degradations deneefe dependenc described differ differing ding discriminative documents dorr efficiently eisner english enough entropy erov error esca eurospeech evaluation extension extremely features fields final fluency fraser galley generally generated generation gildea graehl haji have hierarchical higher hopkins huang human improve improved improvements improves improving including indicating inducing inference informed integrating jain japanese johns judges judgments khudanpur knight koehn kumar language large learning list machine making manner many marcu marker markers maximum mccallum means mejrek menezes method minimally minimum model modeling models more naacl natural nbest needed number only outperformed output over papineni particular parton penn performed phenomena phrasal phrase postediting precise predict proceedings processing proposed purpose quality question quirk radev rambow random ranking rate references reliable remarkable report restructuring result results rich rosenfeld roukos sarkar scalable sentences shen should shown sign significant significantly smith smorgasbord special specific speech standard statistical statistically such summer summit suzuki syntactic syntactically system table target technical test that thayer there therefore this toolkit toutanova training translation translations tree tribble university using variants venugopal vogel waibel wang ward were when where which without workshop would yamada zhang zhao
http://acl.ldc.upenn.edu/N/N07/N07-1028.pdf	28	A Case for Shorter Queries, and Helping Users Create Them	algorithm allan anick assistant ballesteros bikel broglio bruce buckley callan cardie chris claire clustering croft daniel feedback hongming information inquery iterative james janet jinxi john learning learns lisa machine management mandar mitra name pages paraphrase peter proceedings processing ralph references richard schwartz search seeking sigir smart superconcepts suresh terminological that tipirneni trec using walz weischedel what within
http://acl.ldc.upenn.edu/N/N07/N07-1056.pdf	56	Randomized Decoding for Selection-and-Ordering Problems	acknowledge acknowledgements acyclic algorithm algorithms alon amalgam applications applied approximate approximation authors automatic awerbuch azar balas banko based beam been blowing blum broadcast cambridge career closely coding collecting color completeness computationally computer conclusions conll corston decoder decoding development difference directed does downey each eacl eisner employs even exact fast fellows figure fixed formalized formulation found foundation gamon generated generation germann global grant graph guarantees hard hauptmann have headline highest hltnaacl improved inference inlg iteration iterations jacm jahr jelinek joint journal knight language large larger learned less linear local lower machine marcu match method minimumweight mittal module moore motwani national natural nding neighborhoods networks news oliver optimal ordering other overview paper parameter path performance permutations point presented press prize problem problems proceedings processing produced programming proportion quality raghavan randomized references research ringger roth rouge running salesman salesmen scale science scores search second selection sequential size solutions speci speech spoken stack statistical stoc substantially such summarization support task tasks tested than theoretical this those time title titles tractability translation traveling trees tromble unable university used using vempala very weighted witbrock with without workshop yamada yielding york yuster zwick
http://acl.ldc.upenn.edu/N/N07/N07-1052.pdf	52	Approximate Factoring for A∗ Search	acaseo acknowledgments adaptive additive admissible algorithm alignment allard alternative andersen anonymous ariel arti ation basis bell bilingual bounds boyd cambridge cannot charniak christopher cial cient ciently collins comments compare complex compression comput conclusion conference context convex corpora cost creating cybernetics daniel database dekai demic dennis determination driven editor effective emnlp enumerated error estimates estimating eugene even events exact exhaustively factored fellowship felner formal free frequency functions future galley generalized generate ghallab gildea giorgio graehl grammar grammars hanan hart have head heuristic heuristics high hopkins ieee ijcai improved inference intelligence inversion investigate jair jonathan kevin klein kluwer knight korf language learning lieven like linear linguist malik manning marcu mark melamed method methods michael michel minimum model models mosek multitext naacl national natural near nilsson novel ntptnmizlength objective optimization optimizer over parallel parse parsing paths pattern performance plan point presented press probabilities problem problems provably publishers ramming raphael references reviewers richard rule sarit satta science search sequences since statistical statistics stephen stochastic supported systems technique terior text thank their this timothy training transactions transducers transduction transitions translation tree trees university used vandenberghe wellington what when where with witten word work would zero zhang
http://acl.ldc.upenn.edu/N/N07/N07-2052.pdf	125	Comparing Wikipedia and German Wordnet by Evaluating Semantic Relatedness on Multiple Datasets	aaai adjectives akademischer alexander also analysis annual appear applications arti automatic automatically based better between budanitsky cambridge cantly category chapter christiane classical claudia combining communications complementary computational computerlinguistik computing conceptual conclusion cone conference consider contains contextual correlates cream creating database datasets deciding dictionaries disambiguation distance distances documentation electronic evaluate evaluating fellbaum focus from future germanet goodenough graeme graph gurevych hand have herbert hirst human ijcnlp important improves improving international involved iryna jane john knowledge kunze lesk level lexical lexikalisch linguistic linguistics machine major measure measures michael morris naacl ndings need network nouns pages paolo performance performs pine pointed points ponzetto press proc range readable references regarded relatedness relations remarkably requires research rubenstein semantic semantics semantische sense should shown signi simone simple source sources speci spektrum sprachtechnologie strategies strube structure synonymy systems task tell textgraphs that them these thus torsten uence underlying used using validate verbs verlag well whether which while wider wikipedia wikirelate wordnet workshop wortnetze zesch
http://acl.ldc.upenn.edu/N/N07/N07-2009.pdf	82	Generalized Graphical Abstractions for Statistical Machine Translation	aaai advances alignment almost arbor artif bacchus backtracking based bayesia bayesian between bilmes brown cambridge canada coling comparison computational conclusion conf dalmao david dean decoding della described doing dynamic edmonton elimination estimation european extensible filali framework francisco future geiger germann graphical greedy have heckerman inference information intell intelligent jeff june kanazawa karim kaufmann knowledge koehn language languages lauritzen linear linguistics machine marcu mathematics mercer modeling models monz morgan multi multinets naacl networks neural oxford pages parameter pearl phrasebased pietra pitassi plausible press probabilistic proc processing publications reasoning references representation science search shared similarity spoken srilm statistical stolcke systematic systems task temporal tillmann time toolkit translation ulrich value various vogel word work
http://acl.ldc.upenn.edu/N/N07/N07-1011.pdf	11	First-Order Probabilistic Models for Coreference Resolution	algorithm algorithms amir amit andrew applications approach approximate aron baldridge baldwin bonn braz brian bunescu calculi canada censor chains christopher collective collins concerning conference coreference crammer culotta daniel daume denis detection effective emnlp entity exploration extraction eyal features finkel gaifman germany gibbs global grenager high icml ijcai incorporating incremental inference information into israel jason jenny jmlr joint koby large learning lifted local manning marcu margin markov math mccallum measures message methods michael model mooney multiclass networks online open optimization order oxford pages parallel parsing pascal perceptron pittsburgh prediction press probabilistic problems proceedings pronoun ranking raymond razvan references relational representations resolution roark rodrigo rose roth salvo sampling scale scoring search seventh singer statistical structured systems theory tracking tractable trond ultraconservative understanding university vancouver with workshop yoram zenios
http://acl.ldc.upenn.edu/N/N07/N07-2044.pdf	117	The Effects of Word Prediction on Communication Rate for AAC	acceleration advanced alan alternative anson assets assistive augmentative basic between cantly christopher ciency communication completion conclusions content csun debra december demonstrates denis device difference effects engineering enhance enhances experimental fringe graeme gregory hantz heather heidi hickey higginbotham hirst horabail horstmann jeffery jianhua kathleen keith keyboards keystroke knowledge koester langer language lesher level levine lexical marianne mary maxime mccoy method modeling models moist natural newell pages penni pennington performance prediction proceedings processing przywars rate rates references role saylor screen semantic september signi simon simulated stefan technique technology topic trnka typing user users using venkatagiri wells with word yarrington
http://acl.ldc.upenn.edu/N/N07/N07-1059.pdf	59	Automatic Assessment of Student Translations for Foreign Language Tutoring	acknowledge acknowledgements adaptive agreement aided also amta annotating anonymous applications assisted authors automatic baptist based beijing belis bulletin call cambridge china chinese collection collins combining comments computational computer context conversational cyphers darwars data database decision department design dialogue directional diversity domain ehsani engineering english eskenazi evaluation foreign fosler framework generalization generation genesis glass grateful hardison helpful high hovy icassp icslp ight initiative instil interaction interrater interspeech issues itri itsec johnson king knodt language learning learnings like limitations linguistic linguistics lussier machine making mandarin marsella measurement method methods michael models natural ndings papineni paradigm part polifroni popescu potamianos principles probabilistic proc processing pronunciation prosody prototype psychological qualitative quality quantitative recognition references research reviews roukos segmentbased seide seneff solsona some speaker speech spoken statistical strengths suggestions supported system systems tactical technology telephone telephoneaccess their thesis this tina training translation tseng tutoring uebersax university using validation versatile vihjalmsson wang ward witt would yushi zhang zitouni
http://acl.ldc.upenn.edu/N/N07/N07-4012.pdf	152	A Conversational In-car Dialog System	activities administration alex annie architecture association australia automatique automotive badri barry bratt brian calendar cavedon chat christine collaborative computational conference content conversational danilo december development dialogue evaluation florin fuliang galaxy gruenstein harry hauke heather helper highway hurley icslp international interspeech january japan jeff language langues lathrop lawrence lemon lien linguistics madhuri management matthew meeting meng mirkovic mishra multi national nhtsa oliver paci pacling page pages peters philipp pittsburgh play plug practical presentation priorities proceedings processing purver raghunathan ratiu raya raymond reference references research rohit rulemaking russel safety scheideck schmid schmidt sebastian seneff september shriberg spoken stanley stephanie strategies supporting sydney system systems tasking tasks tobias tokyo traf traitement varges vehicle victor weng years zhang
http://acl.ldc.upenn.edu/N/N07/N07-1017.pdf	17	The Domain Restriction Hypothesis: Relating Term Similarity and Semantic Consistency	american analysis applications arti automatic badulescu bompiani buitelaar cafarella canada cial cimiano computational constraints cours deerwester discovery domains downey dumais edmonton etzioni evaluation experimental extraction fabula from furnas generale girju gliozzo harshman indexing information intelligence journal july landauer latent learning lector linguistics linguistique magnini methods moldovan naacl namedentity ontology pages paris part payot popescu press proceedings references relations saussure science semantic shaked society soderland study texts thesis trento university unsupervised weld whole yates
http://acl.ldc.upenn.edu/N/N07/N07-4010.pdf	150	OMS-J: An Opinion Mining System for Japanese Weblog Reviews Using a Combination of Supervised and Unsupervised Approaches	academic analysis annual applied applying approaches araki between cascaded cation chaovalit chunking classi collecting comparison decision dependency down evaluative expressions fukushima guangwei hicss ieice information inui ipsj japan japanese journals kenji kentaro kobayashi kudoh lina matsumoto meeting mining movie note notes nozomi orientation peter pimwadee proceedings processing references review reviews semantic society structure supervised support system taku tateishi technique technology text thumbs toshikazu turney unsupervised using wang yuji zhou
http://acl.ldc.upenn.edu/N/N07/N07-2022.pdf	95	Discriminative Alignment Training without Annotated Data for Machine Translation	abraham alexander aligner alignment alignments analysis arabic asilomar australia ayan banchs based beyond bonnie boxing brown canada chen church coling combination computational correspondences daniel darpa della dorr emnlp english entropy estimation extensive federico finland fintal fraser gale going identifying impact improving ittycheriah iwslt japan kyoto lambert language linguistics machine marcello marcu mathematics maximum mercer natural necip pages parallel parameter parameters patrik peter phrase pietra proc rafael references robert roukos salim semisupervised speech spsa statistical stephen sydney texts their through training translation tuning turku vancouver vincent with word workshop
http://acl.ldc.upenn.edu/N/N07/N07-1010.pdf	10	Coreference or Not: A Twin Model for Coreference Resolution	adam algorithm amer anaphora anaphoric approach association based bean bell berger boguraev branimir british canada cation christopher coling columbia computational conditional conference copenhagen corpus daniel daume david december della detection effective ellen emnlp entity entropy everyone exploration exponential features generalized global goodman herbert identi international iterative joint joshua kennedy language lappin large leass linguistics march marcu math maximum model monthly natural noun numbers october pages parser phrases pietra proc proceedings processing pronominal references resolution riloff scale scaling sequential shalom stephen tracking vancouver vincent without
http://acl.ldc.upenn.edu/N/N07/N07-2024.pdf	97	Detection of Non-native Sentences using Machine-translated Training Data	absolute accuracy advances also approach arboretum articles assistance assisted assistive augmented automated automatic available baldwin based bayes bender better between beyond both boundary brockett burges burstein call careful carry cation chance checking chodorow classi classifying clickthrough collins combining comparison computer conclusions conference contrast corpora corpus correcting corrections corston criterion dale data deaf dekker detecting detection dolan domains eamt effective engines english enough error errors essay evaluation examined exists experts explore explored extensible features feedback flickinger framework from future gamon generalize generative goal grammar handbook harmful harold have heidorn here hermann higher however icall icslp identify indicates inequalities inherently instil intelligent international isahara ishioka izumi japanese joachims jones kameda kawai kernel language large latter leacock learned learners learning less level lexicalised like long machine magazine making marcel maximally mccoy method methods michaud mismatch modeling models moisi more morihiro naacl nagata native natural nature nonnative oepen oliver online optimizing other over parsing pennington perform performing phrasal plan practical precision press proc processing propose publicly ranking reference references relative replace results robert round saiga salient scale scholkopf scores scoring search seen sentence sentencelevel sentences separate service shown sigkdd size smaller smets smola some somers speci spoken srilm statistical still stolcke study such suggests supnithi support symposium system table tagged task tasks techniques technologies test text than that this those three through tomokiyo toolkit train trained training translation translations tutoring uchimoto uency ultimate using utterance walsh weights well with without work would writing written yield
http://acl.ldc.upenn.edu/N/N07/N07-1002.pdf	2	To Memorize or to Predict: Prominence Labeling in Conversational Speech	academic accent annotating baayen beaver bolinger brenier brown calhoun chafe cieri contrastive contrastiveness corpus development discourse distinction distributions english features fisher framework frequency given givenness godfrey graff holliman icassp ieee information jurafsky kevin kimball kluwer kothari language linguistic mcdaniel measurements miller models nenkova nissim niteness pages part point predicting proceedings prominence prosodic prosody publishers references research speech spoken spontaneous steedman stress structure subject subjects switchboard technology telephone topic topics training transcripts utility view walker whitton word workshop
http://acl.ldc.upenn.edu/N/N07/N07-2006.pdf	79	Translation Model Pruning via Usage Statistics for Statistical Machine Translation	able according acknowledgements actual advances after akiba alex algorithm alignment almut also always amount amta ando approach arbitrarily arbor automatic available bach baltimore bannard based baseline beam beneficial best bilingual bing bleu brain broad budapest burch calculate callison campaign certain china chris colin compare comparison conclusions conversation corpora corpus could count coverage damage darpa data decoder denker development done drops each eamt efficient eiichiro element evaluation even exchange extraction federico figure franz fumiaki future gale hanazawa having heavily hermann hewavitharana hildebrand hirofumi hiromi hongkong hungary ichi idea ideas improve improved information interesting introduced ishikawa isotani iwslt japan john josef josh kando kaufmann kishore koehn kolss kyoshi kyoto lane large larger lattice long longer loss lrec machine marcello matthias memory method metric metricbest michael model models morgan mtsummit muntsin nakaiwa necessary network neural nguyen noriko number occurring optimal option outperform over overview pages pair pairs palmas papineni partly path paul pdas performance pesa philadelphia philipp phrase phrases phuket previously proceedings processing programs proposed prune pruned pruning real recalculates reference references remains removing require research roukos ryosuke saliency salim sanjika sara scaling schroeder score scoring search seiichi sentence sequentially shin shinichi significant silja situations smaller software solla spain speech splitting statistical statistics stephan strong sugaya sumita supported system systems takezawa tell thailand there this threshold todd tokyo toshiyuki toward translation transtac travel tsujii under used using very viable vogel waibel ward weijing which while with without work world would yamabana yamamoto yann yasuhiro ying zhang zhao
http://acl.ldc.upenn.edu/N/N07/N07-1020.pdf	20	High-Performance, Language-Independent Morphological Segmentation	achieves across against algorithm alignment also although analysis approach attachment baayen bangla based bayesian bengali bernhard bestperforming better brent case celex challenge clusters compares competition complete complexity computational computer concepts conclusions conll corpora corpus creutz dasgupta database demonstrating department description detection different discovering discovery distributions english evaluated experimental favorably fifth finally finnish form free freitag frequency from general gold goldsmith grounding gulikers hand harris have having helsinki icon identification incorrect indicate induction inflectional information instance international into introduction intuitive jean jurafsky keshava knowledge koskenniemi lagus language languages learning length level levels lexical linguistica linguistics lundberg mentioned minimally minimum model morfessor morph morpheme morphemes morphemic morphessor morphological morphologies morphology multimodal murthy naacl natural necessary none offers only orthographic other outperforms pages paradigm paradigms parsers participating pascal pennsylvania performance performs philadephia phoneme piepenbrock pitler poorer poorly predictability presented prior proceedings production publication recognition references relative report respectively results robust robustness rules schone science score scores segment segmentation significantly simpler snover standards statistics structures study suffixes supervised system systems table target technology term text than that those three turkish univ university unsupervised untagged using when whereas wicentowski winner winners with word words workshop yarowsky
http://acl.ldc.upenn.edu/N/N07/N07-1036.pdf	36	An Exploration of Eye Gaze in Spoken Language Processing for Multimodal Conversational Interfaces	aaai about acous action advanced alignment anaphoric applications approach attention automatic baldridge based behavior bock bolt breath campana carpenter cascaded chai christoudias classen cognition cognitive collaboration communicating communication component comprehension computational computer computers context conversational conversing cooperative data deictic design determine dialogue display dowding driven duchowski eberhard editors eisenstein emnlp environments estimation event expressions eyes ferreira flippo form francis from furness gacovski gaze gesture grif groupware henderson hockey huang huls human icmi ieee ihde information input instruments integration interaction interface interpretation jacob joint just katz kaur kehler knowlton language linguistic linguistics look magic mantravadi manual mediating methods modalities model modeling models morimoto movement movements mukherjee multimodal multiparty naacl nonverbal oxford pages patterns perceptive pointing position pragmatics press priming probabilities problem proceedings processes processing psychological psychology qvarfordt recogniser reference references referent referents remington research resolution responsive robust salience science sedivy selfdisclosing situated solving sparse speaking speech spivey spoken starker status stone survey synchronization system systems tanenhaus taylor techniques termaine towards tracking trans transfer understanding university user using velichkovsky vertegaal virtual vision visual what where wilder with workshop world xations york zhai
http://acl.ldc.upenn.edu/N/N07/N07-1021.pdf	21	Probabilistic Generation of Weather Forecast Texts	acquired approximations arpa automatic automatically belz brighton cahill cation charniak compared comparing context doddington eacl enlg evaluated evaluation free genabith generation gram human itri knight language machine methods models nltg occurrence pages pcfgbased pcru probabilistic proc proceedings quality references reiter report representational robust statistical statistics summit syntaxbased systems technical technology three translation underspeci university using workshop yamada
http://acl.ldc.upenn.edu/N/N07/N07-2019.pdf	92	Implicitly Supervised Language Model Adaptation for Meeting Transcription	about acoustics adapt adaptation adapted agenda also amount analysis annealing april arctic automatically banerjee baron based bhagat bilmes black boosting cache cant carvey cation classi clustering collected component compression conclusions conference content conversation data databases described deterministic dhillon difference distribution edwards entropy equivalent estimation extracting folding found from gelbart given have hofmann hong human icassp icsi identify ieee implicit improvement increases intelligence intelligent interfaces international into isca items iteration janin january katz kominek kong krupski kuhn language latent machine mass meeting meetings method model modeling morgan mori much multi multiparty naacl natural nearly note notes obtain optimization original over participants pattern perplexity peskin pfau pittsburgh plsa probabilistic probabilities probability problematic problems procedure proceedings processing pruning recognition recognizer reduces references regression related renormalized research results rose rudnicky running segmenting semantic show shown shriberg signal signi simply situation sparse speaker speech statistically stockholm stolcke structured subsequent supervision synthesis table taking that their then these this thomas though transactions unigram updated used user values vocabulary while with wooters words workshop
http://acl.ldc.upenn.edu/N/N07/N07-2003.pdf	76	Conquest – an Open-Source Dialog System for Conferences	about access accommodation acts adreani agenda alleva allows annotated approach arpa aruba award based beach been best bohus built call cepstral collect collected collecting communicator computer conference conquest constructing corpora corpus costs darpa decomposition deployed developed developing dialog dialogue differ dimensions discoh domain during edinburgh efforts enable eurospeech expectation fabbrizio facilitates fledged footprint full functionality general geneva gilbert gillick goal goals hakkani have help helpdesk hierarchical http huang human hwang icsi ieee improvements information infrastructure initial instance interfaces issar january language lemon manage management modular more natural number offers olympus open operates options other overview page pages paper participants plainsboro plan primary proceedings project provides publish ravenclaw recent recognition references releasing relies research researchers reward rosenfeld routing rudnicky schedule signals similar small source speech sphinx spoken submission swifttm switzerland synthesizer system task technical technology text that their there this tospeech transparent understanding university using venue vote ward well while wise with within workshop
http://acl.ldc.upenn.edu/N/N07/N07-2015.pdf	88	Are Very Large N-best Lists Useful for SMT?	actual advantages algorithm american annual arbor around assoc association balance based bertoldi best bleu boston building canada candidates cantly cattoni cettolo chapter chen cient complex computational computing conclusion conf current currently described discriminating does easily edmonton emnlp empirical eppstein evaluation experiments extraction features federico finding fraser from generation gildea given good graphs hasan have help human hypotheses improve indicate international irst iwslt jain japan july june khudanpur koehn kumar kyoto language large larger linear linguistics list lists machine marcu matusov mauser meeting method methods models more much naacl natural north november oracle pages parallel paths performance philadelphia phrasebased power proc processing quality radev references representation reranked rescoring results rwth sarkar saturate scores seems separate shen shortest should siam signi smith smorgasbord space spoken statistical stay system technology testing texts that translation translations using whereas with word workshop yamada zens
http://acl.ldc.upenn.edu/N/N07/N07-1025.pdf	25	Using Wikipedia for Automatic Word Sense Disambiguation	aaai acapulco accurate acquisition across adafre agirre algorithms analysis approach automatic automatically barcelona based bias bigrams bootstrapping boston bottleneck brittleness building bunescu cambridge canary categorization cation chaining chan chklovski chodorow computational computers computing cone conference corpora corpus cream data decision diab dictionaries directions disambiguate disambiguation distinguishing eacl emnlp empirical encyclopedic engineering english enhancing entity evaluation examplar examples expert exploiting finding framework from future gabrilovich galley generating humanities identi ieee ijcai importance improving integrating intelligence interconnections islands italy japan july june kilgarriff knowledge language languages large leacock learning lesk lexical linguistics lrec machine markovitch martinez mckeown method methods mexico mihalcea miller mind moldovan multiple naacl named natural navigli open orlando overcoming pami parallel pasca pattern pedersen philadelphia pine pittsburgh ponzetto predictor proceedings readable recent references relatedeness relations relieving resnik results retrieved rijke rivaling rosenzweig sapporo semantic sense senses sentences sigdoc similar sources spain statistics strube structural study successes supervised systems tagged tagging tell text texts toronto transactions tree trento unsupervised using velardi wang wikipedia wikirelate with word wordnet workshop yarowsky
http://acl.ldc.upenn.edu/N/N07/N07-2042.pdf	115	Virtual Evidence for Training Speech Recognizers using Partially Labeled data	acoustic activities adda architectures arti bartels based bayesian bilmes boundary cation chang cial classi clsp computer conference context dept discrimination eurospeech evidence explicity feature fosler framework gauvian glass graphical greece greenberg halberstadt heterogeneous hopkins icslp ieee inference integrating intelligence intelligent interspeech into johns kaufmann lamel language lightly lussier magazine mccandless measurements model morgan networks pages pearl philadelphia phonetic plausible probabilistic proc processing project publishers reasoning recognition recognizing references report research rhodes sensors september signal soft spatial speech subramanya summer supervised switchboard systems technical training transcription uncertainty university unsupervised using uweetr volume wang washington wearable workshop