The Tactical Language Training Project W Lewis Johnson

  • Slides: 20
Download presentation
The Tactical Language Training Project W. Lewis Johnson, Sunhee Choi, Stacy Marsella, Nicolaus Mote,

The Tactical Language Training Project W. Lewis Johnson, Sunhee Choi, Stacy Marsella, Nicolaus Mote, Shrikanth Narayanan, Hannes Vilhjálmsson, Shumin Wu University of Southern California Micro Analysis & Design US Military Academy UCLA CRESST Center for Advanced Research in Technology for Education, USC / ISI

Major Research Themes n Study role of simulation games and intelligent tutoring in promoting

Major Research Themes n Study role of simulation games and intelligent tutoring in promoting communicative skill ä Simulation motivates, builds fluency · Intelligent scaffolding built into simulation ä n n n Tutoring provides form feedback Tailor speech recognition to learner language Emulate free dialog within narrative structure Manage learner motivation ä ä Strong task orientation Promote early communicative success Employ motivational tutorial tactics Adjust level of challenge to maximize engagement Center for Advanced Research in Technology for Education, USC / ISI

Scenario n n Your special ops unit is deployed to a town to assist

Scenario n n Your special ops unit is deployed to a town to assist with postwar recovery The objectives: ä ä ä n Establish rapport with locals, describe mission Ask for directions to local leader Meet leader and discuss recovery needs Initiate mission (organize reconstruction) Overcome barriers to mission (factions, sabotage, etc. ) Impact of communication failure: ä ä Failure of mission Hostile encounters with locals Center for Advanced Research in Technology for Education, USC / ISI

Demo Center for Advanced Research in Technology for Education, USC / ISI

Demo Center for Advanced Research in Technology for Education, USC / ISI

Overall Architecture Center for Advanced Research in Technology for Education, USC / ISI

Overall Architecture Center for Advanced Research in Technology for Education, USC / ISI

Interaction with Mission Skill Builder Center for Advanced Research in Technology for Education, USC

Interaction with Mission Skill Builder Center for Advanced Research in Technology for Education, USC / ISI

Authoring Issues n Story-oriented authoring poses particular challenges: ä Authoring is a multidisciplinary team

Authoring Issues n Story-oriented authoring poses particular challenges: ä Authoring is a multidisciplinary team activity · A mixture of top-down and bottom-up design Curriculum content is introduced incrementally and opportunistically ä Needed: ä · Tools to facilitate incremental construction of materials · Tools to track curriculum coverage · Tools to check for validity and completeness Scene script - HTML Center for Advanced Research in Technology for Education, USC / ISI Scene script - XML

Speech Processing n n Dual goals of robust recognition and spoken language skill evaluation

Speech Processing n n Dual goals of robust recognition and spoken language skill evaluation Recognition subsystem goals: ä ä n Robustness to non native speech Lightweight operation Pronunciation evaluation system goals: ä ä Multilevel evaluation ranging from utterance level prosody to segmental cues Feedback requirements differ for learner and game AI Center for Advanced Research in Technology for Education, USC / ISI

Speech Recognition- System Design n HTK based system ä n Recognition networks loaded dynamically

Speech Recognition- System Design n HTK based system ä n Recognition networks loaded dynamically ä n Networks optimized for different stages in schoolhouse and mission training Mission training mode ä n Bootstrapped from modern standard Arabic and adapted to Levantine speech and lexicon Large vocabulary (LVCSR) like operation Schoolhouse mode for skill acquisition ä Limited vocabulary with pronunciation variants and hypothesis rejection Center for Advanced Research in Technology for Education, USC / ISI

Natural Language Processing n Functions of NLP: Validate authored learning materials ä Link dialog

Natural Language Processing n Functions of NLP: Validate authored learning materials ä Link dialog examples to authored explanations, translations ä Predict & recognize common learner errors ä Center for Advanced Research in Technology for Education, USC / ISI

Managing Pedagogical Drama Want an engaging pedagogical drama n n n Unfold based on

Managing Pedagogical Drama Want an engaging pedagogical drama n n n Unfold based on trainee's language & cultural skills Give trainee a sense of control and responsibility Provides pedagogically appropriate consequences Respond robustly to learner communicative actions Avoid computationally expensive approaches Center for Advanced Research in Technology for Education, USC / ISI

MPE Architecture Center for Advanced Research in Technology for Education, USC / ISI

MPE Architecture Center for Advanced Research in Technology for Education, USC / ISI

Technical Details Director uses POMDPs to model trainee and characters n ä Student model

Technical Details Director uses POMDPs to model trainee and characters n ä Student model informs trainee capabilities POMDPs used to predict possible story paths Fit parameters of character POMDPs to increase likelihood of desirable story path n n ä Example: Adjust importance of being hospitable Fitting sensitive to cultural & situational factors n ä Example: hospitality for strangers, support for US involvement important goals to adjust Center for Advanced Research in Technology for Education, USC / ISI

Evaluation Sessions, May 2004 n n 7 learners 6 some prior knowledge of TLTS

Evaluation Sessions, May 2004 n n 7 learners 6 some prior knowledge of TLTS All had prior L 2 language experience All experienced gamers n n n Practice w/ SB ~ 1 hour MPE session Post session interview Center for Advanced Research in Technology for Education, USC / ISI

General comments n n TLTS viewed as potentially much better than class instruction learners

General comments n n TLTS viewed as potentially much better than class instruction learners readily identify best features (one-on-one, ME), can anticipate benefits n n no major interface or navigation concerns many problems identified are already being solved Center for Advanced Research in Technology for Education, USC / ISI ä ä MSB · speech recognition · learner feedback MPE · avoidance

Skill Builder Conclusions n Pronunciation accuracy threshold too high for beginners ä Since has

Skill Builder Conclusions n Pronunciation accuracy threshold too high for beginners ä Since has been lowered n Tutor ä Feedback selection algorithm adjusted n Lack ä feedback sometimes repetitive of transfer from MSB to MPE New exercises added to MSB to emulate MPE dialog Center for Advanced Research in Technology for Education, USC / ISI

Practice Environment Conclusions n Learners ä reluctant to enter MPE Learners now explicitly encouraged

Practice Environment Conclusions n Learners ä reluctant to enter MPE Learners now explicitly encouraged at multiple points n Learners not sure what to do Provide detailed mission briefings ä Aide helps provide guidance ä Center for Advanced Research in Technology for Education, USC / ISI

Design for Next Evaluation n ~25 subjects, non-Arabic speakers n Four experimental conditions: Complete

Design for Next Evaluation n ~25 subjects, non-Arabic speakers n Four experimental conditions: Complete system ä MPE only ä MSB without pronunciation feedback ä Center for Advanced Research in Technology for Education, USC / ISI

Backup Slides Center for Advanced Research in Technology for Education, USC / ISI

Backup Slides Center for Advanced Research in Technology for Education, USC / ISI

Role of Gesture and Culture n Animate gestures in virtual locals: To promote learner

Role of Gesture and Culture n Animate gestures in virtual locals: To promote learner understanding ä To expose learner to unfamiliar gestures ä n Allow ä learner to select gestures: To learn appropriate gestures for different social situations · E. g. , to build rapport, manage crowds n Culture training involves: Training in face threats, face threat mitigation strategies ä Expectations (etiquette) in different social situations ä Center for Advanced Research in Technology for Education, USC / ISI