Representation and Reasoning for Deeper Natural Language Understanding in a Physics Tutoring System

Students' natural language (NL) explanations in the domain of qualitative mechanics lie in-between unrestricted NL and the constrained NL of proper domain statements. Analyzing such input and providing appropriate tutorial feedback requires extracting information relevant to the physics domain and diagnosing this information for possible errors and gaps in reasoning. In this paper we will describe two approaches to solving the diagnosis problem: weighted abductive reasoning and assumption-based truth maintenance system (ATMS). We also outline the features of knowledge representation (KR) designed to capture relevant semantics and to facilitate computational feasibility. Presented at the Proceedings of the International Florida Artificial Intelligence Research Society Conference (FLAIRS) (19th) held in Melbourne Beach, FL on 11-13 May 2006. Published in the Proceedings of the International Florida Artificial Intelligence Research Society Conference, May 2006.