CS 705: BELIEF SYSTEMS IN ARTIFICIAL INTELLIGENCE Fall 1984 William J. Rapaport Room & time to be arranged Bell 214 Office Hours: M 10-10:50 636-3193 (office) WF 1-2 & by appointment 884-3289 (home) TEXTS: L. Linsky (ed.), Reference & Modality, corrected edition (Oxford: Oxford Univ. Press, 1977). Other reading materials to be distributed in class. TOPICS: This seminar will focus on recent work in AI on the representation of belief and knowledge and on reasoning systems (e.g., deductive question-answering systems) that deal with the relations between belief and action. We shall begin by examining the philosophical background of these systems, as found in the writings of Frege, Meinong, Quine, Kaplan, and Hintikka. We shall then turn to discussions of these issues by research- ers in AI and cognitive science: McCarthy, Moore, Creary, Barnden, Cohen and Perrault, Clark and Marshall, Maida and Shapiro, Wilks and Bien, and Konolige. If time permits, we shall also look at AI work on belief revision (Doyle, Martins) and on applications of philosophical work by Castaneda to AI systems (Rapaport and Shapiro). The seminar is open to all interested participants, but should be of special interest to those in the cognitive sciences: Computer Science, Philosophy, Linguistics, Psychology, Communi- cative Disorders, etc. REQUIREMENTS: No previous background in either philosophy or AI is necessary. Each student will be expected to attend all classes, do all the reading, and write several brief (1-2 page) "position papers" summarizing their reading. There will also be a required term project consisting of a 10-15 page essay or a programming project. The term project will be worth approxi- mately 50% of the final grade, and the position papers will be worth approximately 50% of the final grade.