Dialog Move Generation and Conversation Management in AutoTutor
Natalie K. Person, Arthur C. Graesser, Derek Harter, Eric Mathews · 2000
AutoTutor is an automated computer literacy tutor that participates in a conversation with the student. AutoTutor simulates the discourse patterns and pedagogical dialog moves of human tutors. This paper describes how the Dialog Advancer Network (DAN) manages AutoTutor’s conversations and how AutoTutor generates pedagogically effective dialog moves that are sensitive to the quality and nature of the learner’s dialog contributions. Two versions of AutoTutor are discussed. AutoTutor-1 simulates the dialog moves of normal, untrained human tutors, whereas AutoTutor-2 simulates dialog moves that are motivated by more ideal tutoring strategies. Background Human one-to-one tutoring is second to no other instructional method in yielding positive student learning gains. This particular claim has been supported in numerous research studies and is not particularly controversial. However, when the tutors of typical tutoring situations are considered, this claim becomes somewhat perplexing. Most tutors in school settings are older students, parent volunteers, or teachers ’ aides that possess some knowledge about particular topic domains and virtually no knowledge about expert tutoring techniques. Given their limited knowledge, it is somewhat impressive that these untrained tutors are responsible for the considerable learning gains that have been reported in the tutoring literature. Effect sizes ranging from.5 to 2.3 standard deviations have been reported for untrained tutors versus other comparable learning conditions (Bloom, 1984; Cohen, Kulik, & Kulik, 1982). In order to identify the mechanisms that produce such positive learning gains, several members of the Tutoring Research Group (TRG) extensively analyzed a large corpus of tutoring interactions that occurred between