Proceedings of SSST-2 Second Workshop on Syntax and Structure in Statistical Translation David Chiang and Dekai Wu (editors) ACL-08: HLT Columbus, Ohio, USA 20 June 2008 ProductionandManufacturingby OmnipressInc. 2600AndersonStreet Madison,WI53707 USA (cid:13)c 2008TheAssociationforComputationalLinguistics OrdercopiesofthisandotherACLproceedingsfrom: AssociationforComputationalLinguistics(ACL) 209N.EighthStreet Stroudsburg,PA18360 USA Tel: +1-570-476-8006 Fax: +1-570-476-0860 [email protected] ISBN978-1-932432-16-9 ii Introduction The Second Workshop on Syntax and Structure in Statistical Translation (SSST-2) was held on 20 June2008followingtheACL-08:HLTconferencehostedbyOhioStateUniversityinColumbus,Ohio. LikethefirstSSSTworkshopin2007,itaimedtobringtogetherresearchersfromdifferentcommunities workingintherapidlygrowingfieldofstatistical,tree-structuredmodelsofnaturallanguagetranslation. We selected eleven papers for this year’s workshop. There was a strong emphasis this year on the use of explicit syntactic information: constituent structures (Yamamoto, Okuma and Sumita; Zhou, Xiang, Zhu and Gao; Elming; Clark, Frederking and Levin; Lavie, Parlikar and Ambati), dependency structures (Nikoulina and Dymetman; Ma, Ozdowska, Sun and Way), part-of-speech tags (Tillmann), and combinations thereof (Ge, Ittycheriah and Papineni). These papers applied syntactic information to grammar-based models as well as to phrase-based and word-based models. The program was roundedoutbypapersdescribingnewdecodingtechniques(LiandKhudanpur)andmachine-learning techniques(Subotin)forgrammar-basedtranslationmodels. WewouldliketothankourauthorsandourProgramCommitteeformakingthisyear’sSSSTworkshop anothersuccess. DavidChiangandDekaiWu1 1ThisworkwassupportedinpartbytheDefenseAdvancedResearchProjectsAgency(DARPA)underGALEContract Nos.HR0011-06-C-0023,subcontractSRIInternational(DekaiWu)andHR0011-06-C-0022,subcontractBBNTechnologies 9500008412(DavidChiang).Anyopinions,findingsandconclusionsorrecommendationsexpressedinthismaterialarethose oftheauthor(s)anddonotnecessarilyreflecttheviewsoftheDefenseAdvancedResearchProjectsAgency. iii Organizers: David CHIANG,USCInformationSciencesInstitute,USA Dekai WU,HongKongUniversityofScienceandTechnology(HKUST),HongKong ProgramCommittee: Srinivas BANGALORE,AT&TResearch,USA Marine CARPUAT,HongKongUniversityofScienceandTechnology(HKUST),HongKong Pascale FUNG,HongKongUniversityofScienceandTechnology(HKUST),HongKong Daniel GILDEA,UniversityofRochester,USA Kevin KNIGHT,USCInformationSciencesInstitute,USA Jonas KUHN,UniversityofPotsdam,Germany Yang LIU,InstituteofComputingTechnology,ChineseAcademyofSciences,China Daniel MARCU,USCInformationSciencesInstitute,USA Yuji MATSUMOTO,NaraInstituteofScienceandTechnology,Japan Hermann NEY,RWTHAachen,Germany Owen RAMBOW,ColumbiaUniversity,USA Philip RESNIK,UniversityofMaryland,USA Stefan RIEZLER,GoogleInc.,USA Libin SHEN,BBNTechnologies,USA Christoph TILLMANN,IBMT.J.WatsonResearchCenter,USA Stephan VOGEL,CarnegieMellonUniversity,USA Taro WATANABE,NTTCommunicationScienceLaboratories,Japan Andy WAY,DublinCityUniversity,Ireland Yuk-Wah WONG,GoogleInc.,USA Richard ZENS,GoogleInc.,USA v Table of Contents ImposingConstraintsfromtheSourceTreeonITGConstraintsforSMT Hirofumi YAMAMOTO,Hideo OKUMAandEiichiro SUMITA...............................1 A Scalable Decoder for Parsing-Based Machine Translation with Equivalent Language Model State Maintenance Zhifei LIandSanjeev KHUDANPUR.....................................................10 PriorDerivationModelsForFormallySyntax-BasedTranslationUsingLinguisticallySyntacticParsing andTreeKernels Bowen ZHOU,Bing XIANG,Xiaodan ZHUandYuqing GAO .............................. 19 GeneralizingLocalTranslationModels Michael SUBOTIN.....................................................................28 ARule-DrivenDynamicProgrammingDecoderforStatisticalMT Christoph TILLMANN..................................................................37 SyntacticReorderingIntegratedwithPhrase-BasedSMT Jakob ELMING........................................................................46 ExperimentsinDiscriminatingPhrase-BasedTranslationsontheBasisofSyntacticCouplingFeatures Vassilina NIKOULINAandMarc DYMETMAN............................................55 MultipleReorderingsinPhrase-BasedMachineTranslation Niyu GE,Abe ITTYCHERIAHandKishore PAPINENI .....................................61 ImprovingWordAlignmentUsingSyntacticDependencies Yanjun MA,Sylwia OZDOWSKA,Yanli SUNandAndy WAY ..............................69 Inductive Detection of Language Features via Clustering Minimal Pairs: Toward Feature-Rich Gram- marsinMachineTranslation JonathanH. CLARK,Robert FREDERKINGandLori LEVIN ............................... 78 Syntax-Driven Learning of Sub-Sentential Translation Equivalents and Translation Rules from Parsed ParallelCorpora Alon LAVIE,Alok PARLIKARandVamshi AMBATI.......................................87 vii Workshop Program Friday,June20,2008 9:00–9:05 OpeningRemarks 9:05–9:30 ImposingConstraintsfromtheSourceTreeonITGConstraintsforSMT Hirofumi YAMAMOTO,Hideo OKUMAandEiichiro SUMITA 9:30–9:55 A Scalable Decoder for Parsing-Based Machine Translation with Equivalent Lan- guageModelStateMaintenance Zhifei LIandSanjeev KHUDANPUR 9:55–10:20 Prior Derivation Models For Formally Syntax-Based Translation Using Linguisti- callySyntacticParsingandTreeKernels Bowen ZHOU,Bing XIANG,Xiaodan ZHUandYuqing GAO 10:30–11:00 CoffeeBreak 11:00–11:25 GeneralizingLocalTranslationModels Michael SUBOTIN 11:25–12:15 InvitedTalk 12:15–13:45 Lunch 13:45–14:10 ARule-DrivenDynamicProgrammingDecoderforStatisticalMT Christoph TILLMANN 14:10–14:35 SyntacticReorderingIntegratedwithPhrase-BasedSMT Jakob ELMING 14:40–15:05 ExperimentsinDiscriminatingPhrase-BasedTranslationsontheBasisofSyntactic CouplingFeatures Vassilina NIKOULINAandMarc DYMETMAN 15:05–15:30 MultipleReorderingsinPhrase-BasedMachineTranslation Niyu GE,Abe ITTYCHERIAHandKishore PAPINENI 15:30–16:00 CoffeeBreak ix Friday,June20,2008(continued) 16:00–16:25 ImprovingWordAlignmentUsingSyntacticDependencies Yanjun MA,Sylwia OZDOWSKA,Yanli SUNandAndy WAY 16:25–16:50 InductiveDetectionofLanguageFeaturesviaClusteringMinimalPairs: TowardFeature- RichGrammarsinMachineTranslation JonathanH. CLARK,Robert FREDERKINGandLori LEVIN 16:50–17:15 Syntax-DrivenLearningofSub-SententialTranslationEquivalentsandTranslationRules fromParsedParallelCorpora Alon LAVIE,Alok PARLIKARandVamshi AMBATI 17:15–17:45 PanelDiscussion x
Description: