个人简历

屠可伟，信息学院助理教授，博士生导师，目前在学院担任视觉与数据智能中心主任、毕业工作委员会主任。
屠可伟博士，上海科技大学信息科学与技术学院助理教授、博士生导师。于2002和2005年在上海交通大学计算机科学与工程系获学士和硕士学位；2012年于美国爱荷华州立大学获计算机科学博士学位；2012至2014年在美国加州大学洛杉矶分校统计系与计算机系从事博士后研究工作。研究方向包括自然语言处理、机器学习、知识表示等人工智能领域，目前侧重于研究文法的表示、学习与应用

研究领域

自然语言处理
机器学习
知识表示
计算机视觉
人工智能

近期论文

WenjuanHan,GeWang,YongJiangandKeweiTu,“MultilingualGrammarInductionwithContinuousLanguageIdentification”,inProceedingsofthe2019ConferenceonEmpiricalMethodsinNaturalLanguageProcessingand9thInternationalJointConferenceonNaturalLanguageProcessing(EMNLP-IJCNLP2019),HongKong,China,November3–7,2019.
YongJiang,WenjuanHanandKeweiTu,“ARegularization-BasedFrameworkforBilingualGrammarInduction”,inProceedingsofthe2019ConferenceonEmpiricalMethodsinNaturalLanguageProcessingand9thInternationalJointConferenceonNaturalLanguageProcessing(EMNLP-IJCNLP2019),HongKong,China,November3–7,2019.
LiwenZhang,KeweiTuandYueZhang,“LatentVariableSentimentGrammar”,inProceedingsofthe57thAnnualMeetingoftheAssociationforComputationalLinguistics(ACL2019),Florence,Italy,July28–August2,2019.
WenjuanHan,YongJiangandKeweiTu,“EnhancingUnsupervisedGenerativeDependencyParserwithContextualInformation”,inProceedingsofthe57thAnnualMeetingoftheAssociationforComputationalLinguistics(ACL2019),Florence,Italy,July28–August2,2019.
YunzheYuan,YongJiangandKeweiTu,“BidirectionalTransition-BasedDependencyParsing”,inProceedingsoftheThirty-ThirdAAAIConferenceonArtificialIntelligence(AAAI2019),Honolulu,Hawaii,USA,January27-February1,2019.
YanpengZhao,LiwenZhangandKeweiTu,“GaussianMixtureLatentVectorGrammars”,inProceedingsofthe56thAnnualMeetingoftheAssociationforComputationalLinguistics(ACL2018),Melbourne,Australia,July15–20,2018.
JunMei,YongJiangandKeweiTu,“MaximumAPosterioriInferenceinSum-ProductNetworks”,inProceedingsoftheThirty-SecondAAAIConferenceonArtificialIntelligence(AAAI2018),NewOrleans,Lousiana,USA,February2–7,2018.
YongJiang,YangZhouandKeweiTu,“LearningandEvaluationofLatentDependencyForestModels”,toappearinNeuralComputingandApplications.
YongJiang,WenjuanHanandKeweiTu,“CombiningGenerativeandDiscriminativeApproachestoUnsupervisedDependencyParsingviaDualDecomposition”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2017),Copenhagen,Denmark,September7–11,2017.
WenjuanHan,YongJiangandKeweiTu,“DependencyGrammarInductionwithNeuralLexicalizationandBigTrainingData”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2017),Copenhagen,Denmark,September7–11,2017.
JiongCai,YongJiangandKeweiTu,“CRFAutoencoderforUnsupervisedDependencyParsing”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2017),Copenhagen,Denmark,September7–11,2017.
XiaoZhang,YongJiang,HaoPeng,KeweiTuandDanGoldwasser,‘Semi-supervisedStructuredPredictionwithNeuralCRFAutoencoder”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2017),Copenhagen,Denmark,September7–11,2017.
ShanboChu,YongJiangandKeweiTu,“LatentDependencyForestModels”,inProceedingsoftheThirty-FirstAAAIConferenceonArtificialIntelligence(AAAI2017),SanFrancisco,California,USA,February4–9,2017.
LinQiu,KeweiTuandYongYu,“Context-DependentSenseEmbedding”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2016),Austin,Texas,USA,November1-5,2016.
YongJiang,WenjuanHanandKeweiTu,“UnsupervisedNeuralDependencyParsing”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2016),Austin,Texas,USA,November1-5,2016.
KeweiTu,“ModifiedDirichletDistribution:AllowingNegativeParameterstoInduceStrongerSparsity”,inProceedingsoftheConferenceonEmpiricalMethodsinNaturalLanguageProcessing(EMNLP2016),Austin,Texas,USA,November1-5,2016.
KeweiTu,“StochasticAnd-OrGrammars:AUnifiedFrameworkandLogicPerspective”,inProceedingsofthe25thInternationalJointConferenceonArtificialIntelligence(IJCAI2016),NewYorkCity,USA,July9-15,2016.
KeweiTu,MengMeng,MunWaiLee,TaeEunChoe,andSong-ChunZhu,“JointVideoandTextParsingforUnderstandingEventsandAnsweringQueries”,inProceedingsofIEEEMultiMedia,vol.21,no.2,pp.42-70,2014.