Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
NaturalLanguageProcessingandBigDataforPsychologyandComputationalLinguistics__________________________________________________________________________________________________________________
OverviewOver the past decade, there has been a remarkable cross-fertilization between the fields ofpsycholinguisticsandnatural languageprocessingwith insights fromeachfieldenhancingthedevelopmentoftheory,methodology,andresourcesintheother.
DevelopmentsinstatisticalNLPhaveleadtoamuchlargeravailabilityoftextualresourcesforpsycholinguistic research and have made it possible for psycholinguists to quickly developspecific resources such as corpora and word-frequency lists. As a result, there has been anexplosive growth in behavioral and linguistic data that can be leveraged for psycholinguisticresearch.
Our increasingly networked and technological society has also spurred development oftechniquessuchasnatural languageunderstandingandmachinetranslation.Thishasledtoare-birthofartificial intelligenceknownasdeeplearning,whichcanbetracedbacktolearningtheory developed in psychology in (Rescorla &Wagner, 1977) and can again be applied toresearchonhumanlanguageprocessing(e.g.,Mandera,Keuleers,&Brysbaert,2017).
These developments are quickly changing psycholinguistic research. From a field that fordecades has beendominatedby small-scale one-time controlled laboratory experiments it isbecomingadynamic researchenterprise relyingon reusable anddistributeddata generationprocessesleveragingcrowdparticipation(Keuleers&Balota,2015).
Knowledgeof thesedevelopments and techniques is becoming indispensable for researchersworking in any area of cognitive science that deals directly or indirectly with language orlinguisticresources.
Objectives
Theprimaryobjectivesforthiscourseare:
1) Toexplore themutual connectionsbetween the fieldsofnatural languageprocessingandlanguagepsychology,fromahistoricalperspective.
2) To teach participants natural language processing techniques that can be applied tobehavioralresearch.
3)To instill inparticipantsthemindsettodocreativepsycholinguistic researchthat takes fulladvantageofNLPtechniquesandbigdata.
Modules Dates:December10th,2018–December15th,2018.
Day1(10thDecember)ModuleA:HumanandMachineLearningofLanguageLecture1(EK,1.5hrs)Language,behaviorism,andcognitive-science.Lecture 2 (EK, 1.5 hrs)Classical conditioning, operant conditioning, theRescorla-Wagnermodel,theperceptron,recurrentneuralnetworks,deeplearning.Day2(11thDecember)Lecture 3 (EK, 1.5 hrs)Discriminative and associative learning of humanlanguage,distributionalsemanticsModuleB:NaturalLanguageProcessingTutorial1(EK,2hrs)Basicpython,regularexpressions,tokenizing.Day3(12thDecember)Tutorial2(EK,2hrs)Lemmatizing,named-entityrecognition.Tutorial3(EK,2hrs)Counting,wordfrequencydistributions.Day4(13thDecember)Tutorial 4 (EK, 2 hrs)TFxIDF, Feature Extraction, Classification, MachineLearningalgorithms
Tutorial5(EK,2hrs)Vector-SpacemodelsDay5(14thDecember)ModuleC:CrowdsourcingandBigDataLecture4(EK,1.5hrs)SocialIntelligence,SocialMediaAnalytics2Lecture5(EK,1.5hrs)CrowdsourcingofbehavioraldataDay6(15thDecember)ModuleD:TheoreticalapplicationsLecture6(EK,1.5hrs)Frequency,visualwordrecognition,grammaticalityjudgments.Lecture7(EK,1.5hrs)Aging,multilingualism,transactionrecords.
Whocanattend? Thecoursecanbeattendedbystudentsfromdiversefieldsascomputerscience, linguistics, cognitive psychology, cognitive sciences,psycholinguistics,NaturalLanguageProcessing,DataMining,BigDataandsoon.StudentsatalllevelsBTech/MTech,BSc./MSc.andfromearlylevelPhDs across these disciplines and also interested young faculty fromreputeduniversitiesandtechnicalinstitutionsarewelcometoattendthecourse.
Fees Theparticipationfeesfortakingthecourseisasfollows:Participantsfromabroad:US$200Industry/ResearchOrganizations:15000INRAcademicInstitutions:5000INRStudents:2000INRThe above fee include all instructional materials, computer use fortutorialsandassignments,laboratoryequipmentusagecharges,24hrfreeinternetfacility.Theparticipantswillbeprovidedwithaccommodationonpaymentbasis.
TheFaculty
Dr.EmmanuelKeuleers isanAssistantProfessor inDepartment of Cognitive Science and ArtificialIntelligenceatTilburgUniversity(TheNetherlands).Hisresearchissituatedattheintersectionofnaturallanguage processing, machine learning, theoreticallinguistics, and psychology. He is the author ofnumerous well-cited publications on newtechniques, datasets, and theoretical findings inthese fields.His researchhas been instrumental inestablishing the use of techniques from naturallanguage processing and computational linguisticsin psycholinguistic investigation.Hewas the editorofarecentspecial issueoftheQuarterlyJournalofExperimental Psychology on "Megastudies,crowdsourcing, and big datasets inpsycholinguistics". He is a consulting editor forBehavior Research Methods, a member of theeditorialboardoftheMentalLexiconandaregularcontributortotopjournals inhisfieldsofresearch.His current work focuses on the computationalmodeling of human language processing followingdesign principles from information theory,discriminationlearning,andnetworkscience.
CourseCoordinatorDr.ArkVerma,AssistantProfessorofPsychology,IITKanpur.Phone:+91-512-6847(O),+91-9506304191E-mail:[email protected]...........................................................................