Treating Autism with the Help of Artificial Intelligence: A Value Proposition
by William L. Jarrold
to be published in the Proceedings of Agent-Based Systems for Human Learning and Entertainment (ABSHLE) Workshop at AAMAS, May 2007.
Abstract
A growing number of individuals afflicted with autism need improved psychoeducational materials to remedy difficulties with affective mindreading – i.e., inferring the emotions of other people. One approach to this problem is to build an AI-based tutoring system that teaches appropriate mindreading
skills. A review of three studies explains how a rule-based model of simple affective mindreading was constructed based on exercises from a clinical workbook as well as data from empirical studies involving humans. This model was tested in the third study and was seen to produce inferences that were of comparable believability to humangenerated
appraisals.
