Accuracy of tracking student's natural language in operation ARIES!, a serious game for scientific methods

  • Authors:

  • Zhiqiang Cai;Carol Forsyth;Mae-Lynn Germany;Arthur Graesser;Keith Millis

  • Affiliations:

  • Institute for Intelligent Systems, University of Memphis, Memphis, TN;Institute for Intelligent Systems, University of Memphis, Memphis, TN;Institute for Intelligent Systems, University of Memphis, Memphis, TN;Institute for Intelligent Systems, University of Memphis, Memphis, TN;Department of Psychology, Northern Illinois University, Dekalb, IL

  • Venue:
  • ITS'12 Proceedings of the 11th international conference on Intelligent Tutoring Systems
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

OperationARIES! is an ITS that uses natural language conversations in order to teach research methodology to students in a serious game environment. Regular expressions and Latent Semantic Analysis (LSA) are used to evaluate the semantic matches between student contributions, expected good answers and misconceptions. Current implementation of these algorithms yields accuracy comparable to human ratings of student contributions. The performance of LSA can be further perfected by using a domain-specific rather than a generic corpus as a space for interpreting the meaning of the student generated contributions. ARIES can therefore accurately compute the quality of student answers during natural language tutorial conversations.