40
Making Sense of Multimodal Learning Analytics (MMLA) Marcelo Worsley Assistant Professor, Learning Sciences & Computer Science

Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

  • Upload
    others

  • View
    6

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Making Sense of Multimodal Learning Analytics (MMLA) Marcelo Worsley AssistantProfessor,LearningSciences&ComputerScience

Page 2: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

-  What is MMLA? -  Why use MMLA? -  How to use MMLA? -  Examples of MMLA research -  On-going challenges and future directions -  Resources

Agenda

Page 3: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

What is MMLA?

LearningAnaly9cs—asetofmul9-modalsensoryinputs,thatcanbeusedtopredict,understandandquan9fystudentlearning.

(WorsleyandBlikstein,2011)

Page 4: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

What is MMLA?

Mul9modallearninganaly9cs(MMLA)(Blikstein&Worsley,inpress.;Blikstein,2013;Worsley,2012)sitsattheintersec9onofthreeideas:mul9modalteachingandlearning,mul9modaldata,andcomputer-supportedanalysis.Atitsessence,MMLAu9lizesandtriangulatesamongnon-tradi9onalaswellastradi9onalformsofdatainordertocharacterizeormodelstudentlearningincomplexlearningenvironments.However,aswedescribelater,thewaysthatresearchersu9lizemul9modaldatavarywidely.

(Worsley,Abrahamson,Blikstein,Grover,SchneiderandTissenbaum,2016)

Page 5: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

What is MMLA?

Page 6: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

What is MMLA?

Page 7: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

What is MMLA?

Page 8: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

VIDEO

Pose

Gaze

Gestures

ObjectManipula9on

FacialExpressions

Proxemics

SceneUnderstanding

Movement

Heartrate

ObjectTracking

AUDIO

SpeakerDiariza9on

Prosody

Spectrum

Emo9on/Affect

Intensity

SpeakingRate

Collabora9on/Turn-Taking

VoiceQuality

TEXT

Emo9on/Affect

Complexity

Cohesion

Seman9cs

Syntac9cs

Content

EYETRACKINGDATA

Fixa9ons

Fixa9onPaeern

PupilDila9on

Eyeloca9on

Aeen9on

HeadPose

DEPTHCAMERA

Gestures

BodyPosi9on

JointTracking

SMARTPHONES

Signg/Standing

Loca9on

NoiseLevel

LEAPMOTION

Gestures

ToolUsage

HandMovement

FingerMovement

EDAWATCH

Arousal

Cogni9veLoad

BodyTemperature

Gestures

Heartrate

Stress

EVENTLOGS

Ac9vity

OTHERMODALITIES

fMRI

ELECTROCARDIOLOGY

ELECTROENCEPHALOGY

ELECTROMYOGRAPHY

ELECTROGASTROGRAPHY

Page 9: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Why use MMLA? -  Teachingandlearningaremul9modal-  Studyandsupportcomplexlearning

environments-  Seethehardtosee

-  Informdesignofmul9modaltechnologies-  Expandno9onsoflearningtonon-

tradi9onalmodali9es-  Improveaccessibilityandinclusivity

-  Triangulateacrossmodali9es

Page 10: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Why use MMLA? -  Visualizing/Represen9nginforma9onfor

humaninference-  Predic9onofindicators-  Data-driveninterven9ons-  Evalua9ngconjecture-basedlearning

designs

Page 11: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 12: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 13: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA? DATACAPTURESOFTWARE

OpenSocialSignalInterpreta9on

LabStreamingLayer

OpenPipeKit

OpenBroadcasterSohware

Mul9sense(AVRecorder)

iMo9onsAeen9onTool

Page 14: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA? DATACAPTURESOFTWARE

OpenSocialSignalInterpreta9on

LabStreamingLayer

OpenPipeKit

OpenBroadcasterSohware

Mul9sense(AVRecorder)

iMo9onsAeen9onTool

Page 15: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA? DATACAPTURESOFTWARE

OpenSocialSignalInterpreta9on

LabStreamingLayer

OpenPipeKit

OpenBroadcasterSohware

Mul9sense(AVRecorder)

iMo9onsAeen9onTool

Page 16: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA? DATACAPTURESOFTWARE

OpenSocialSignalInterpreta9on

LabStreamingLayer

OpenPipeKit

OpenBroadcasterSohware

Mul9sense(AVRecorder)

iMo9onsAeen9onTool

Page 17: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA? DATACAPTURESOFTWARE

OpenSocialSignalInterpreta9on

LabStreamingLayer

OpenPipeKit

OpenBroadcasterSohware

Mul9sense(AVRecorder)

iMo9onsAeen9onTool

Page 18: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 19: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Pre-processingisimportantfordatasynchroniza9on,accoun9ngforindividualdifferencesbetweenpar9cipants,andgegngdataintheappropriateformatfordataextrac9on.Ittendstovarybydatatype.

Page 20: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 21: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

VIDEOOpenFace(gaze)OpenFace(facerecogni9on)Emo9entFACETNoldusFaceReaderOpenCVAffec9vaMicrosohFaceAPIMicrosohEmo9onAPIIntrafaceMul9senseEularianMagnifica9onCodeiMo9ons

AUDIOHTKPraatOpenSmileOpenEarCovarepELANICSIDiarizerMatlabLIUMDiarizerCMUSphinxGoogleASRAT&TWatsonASRBingSpeechAPIAudacityEmovoice

TEXT

NaturalLanguageToolkit

Lightside

LIWC

Cohmetrix

ToneAnalyzer

StanfordParser

Wordnet

Sen9wordnet

Mallet

Word2Vec

EYETRACKINGDATA

PyGaze

Ogama

EyeTrackingR

TobiiSDK

SMISDK

iMo9onsAeen9onTool

PupilDila9on

DEPTHCAMERA

OpenNUI

KinectforWindows

QSRLib

Libfreenect

EDAWATCH

Ledalab

DATAEXTRACTIONSOFTWARE

OpenSocialSignalInterpreta9on

Page 22: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 23: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

HumanCoders

Page 24: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Fusion Let’ssaythatyouwanttostudyengagementandhave

audio,video,bio-physiologicalandgesturedataavailableforanalysis.Howdoyouusethesetocomputea

measureofengagement?

Page 25: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 26: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among
Page 27: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

How to use MMLA?

Page 28: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Examples

Page 29: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among
Page 30: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

XboxKinect-  Audio(Talk)-  Video(Pose/

Gaze)-  Gestures(Hand

Movement)

RapidMiner-  XMeans

Researcher(s)

Page 31: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

XboxKinect-  Audio(Talk)-  Video(Pose/

Gaze)-  Gestures(Hand

Movement)

RapidMiner-  XMeans

Researcher(s)Custom

Page 32: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Challenges & Future Directions

Page 33: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Simplifying Data Capture & Analysis

Page 34: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among
Page 35: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Better Visualization/Inference Tools

Page 36: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Best Practices Around Data Fusion and Data Analysis Pipelines

Page 37: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Applications (especially as it relates to inclusive technology

and providing feedback to learners and teachers)

Page 38: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Resources Mul9modalLearningAnaly9csSpecialInterestGroup-hep://sigmla.orgCrossMMLAWorkshopsEC-TEL2017,LAK2018(tenta9ve)MMLAWorkshops@ICMI,LAK,ICLS(2012–Present)SOLARLASI–Mul9modalLearningAnaly9csTutorialWorkshopsJournalofLearningAnaly9csSpecialSec9ononMul9modalLearningAnaly9csCIRCLCyberlearningReport

Page 39: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

Berland,M.,Baker,R.S.,&Blikstein,P.(2014).Educa9onalDataMiningandLearningAnaly9cs:Applica9onstoConstruc9onistResearch.Technology,KnowledgeandLearning,19(1–2),205–220.

Blikstein,P.,&Worsley,M.(2016).Mul9modalLearningAnaly9csandEduca9onDataMining :usingcomputa9onaltechnologiestomeasurecomplexlearningtasks,3(2),220–238.heps://doi.org/hep://dx.doi.org/10.18608/jla.2016.32.11

Fouse,A.S.(2011).ChronoViz :Asystemforsuppor9ngnaviga9onof9me-codeddata.Chi,1–6.heps://doi.org/10.1145/1979742.1979706

Grafsgaard,J.F.(2014).Mul9modalAnalysisandModelingofNonverbalBehaviorsDuringTutoring.InProceedingsofthe16thInterna=onalConferenceonMul=modalInterac=on(pp.404–408).NewYork,NY,USA:ACM.heps://doi.org/10.1145/2663204.2667611

Leong,C.W.,Chen,L.,Feng,G.,Lee,C.M.,&Mulholland,M.(2015).U9lizingdepthsensorsforanalyzingmul9modalpresenta9ons:Hardware,sohwareandtoolkits.ICMI2015-Proceedingsofthe2015ACMInterna=onalConferenceonMul=modalInterac=on,(3),547–556.heps://doi.org/10.1145/2818346.2830605

Luz,S.(2013).Automa9ciden9fica9onofexpertsandperformancepredic9oninthemul9modalmathdatacorpusthroughanalysisofspeechinterac9on.Proceedingsofthe15thACMonInterna=onalConferenceonMul=modalInterac=on-ICMI’13,575–582.heps://doi.org/10.1145/2522848.2533788

Morency,L.-P.,Oviae,S.,Scherer,S.,Weibel,N.,&Worsley,M.(2013).ICMI2013grandchallengeworkshoponmul9modallearninganaly9cs.Proceedingsofthe15thACMonInterna=onalConferenceonMul=modalInterac=on-ICMI’13,373–378.heps://doi.org/10.1145/2522848.2534669

Oviae,S.,&Cohen,A.(2013).Wrieenandmul9modalrepresenta9onsaspredictorsofexper9seandproblem-solvingsuccessinmathema9cs.…15thACMonInterna=onalConferenceonMul=modal…,1–8.Retrievedfromhep://dl.acm.org/cita9on.cfm?id=2533793

Scherer,S.,Worsley,M.,&Morency,L.-P.(2012).1stinterna9onalworkshoponmul9modallearninganaly9cs.InICMI(pp.609–610).

Schneider,B.,&Blikstein,P.(2015).UnravelingStudents’Interac9onAroundaTangibleInterfaceusingMul9modalLearningAnaly9cs.JournalofEduca=onalDataMining.

Spikol,D.(2017).UsingMul9modalLearningAnaly9cstoIden9fyAspectsofCollabora9oninProject-BasedLearningIntroduc9onPELARSsystemandcontext.Cscl,(June),263–270.heps://doi.org/10.22318/cscl2017.37

Thompson,K.(2013).Usingmicro-paeernsofspeechtopredictthecorrectnessofanswerstomathema9csproblems:anexerciseinmul9modallearninganaly9cs.…15thACMonInterna=onalConferenceonMul=modal….Retrievedfromhep://dl.acm.org/cita9on.cfm?id=2533792

Worsley,M.(2012).Mul9modalLearningAnaly9cs-Enablingthefutureoflearningthroughmul9modaldataanalysisandinterfaces.Interna=onalConferenceonMul=modalInterac=on,12–15.heps://doi.org/10.1145/2388676.2388755

Worsley,M.,Scherer,S.,Morency,L.P.,&Blikstein,P.(2016).ExploringBehaviorRepresenta9onforLearningAnaly9cs.Proceedingsofthe2015Interna=onalConferenceonMul=modalInterac=on(ICMI),251–258.

XavierOchoa,MarceloWorsley,KatherineChiluiza,&SaturninoLuz.(2014).MLA’14:ThirdMul9modalLearningAnaly9csWorkshopandGrandChallenges.Proceedingsofthe16thInterna=onalConferenceonMul=modalInterac=on,531–532.heps://doi.org/10.1145/2663204.2668318

Page 40: Making Sense of Multimodal Learning Analytics · 2017. 10. 11. · and learning, mul9modal data, and computer-supported analysis. At its essence, MMLA u9lizes and triangulates among

39

MarceloWorsleyAssistantProfessorElectricalEngineeringandComputerScience&LearningSciencesmarcelo.worsley@northwestern.eduMarceloworsley.com9ilt.northwestern.edu