edith@ai.toronto.edu (Edith Fraser) (02/06/90)
Department of Computer Science, University of Toronto (GB = Gailbraith Building, 35 St. George Street) ------------------------------------------------------------- ARTIFICIAL INTELLIGENCE/THEORY SEMINAR GB305, at 11:00 a.m., Thursday 15 February 1990 Professor Joe Halpern U of T and IBM Almaden Rsch. Ctr. "Two views of belief: Belief as generalized probability and belief as evidence" Belief functions are mathematical objects defined to satisfy three axioms that look somewhat similar to the Kolmogorov axioms defining probability functions. We argue that there are (at least) two useful and quite different ways of understanding belief functions. The first is as a generalized probability function (which technically corresponds to the inner measure induced by a probability function). The second is as a way of representing evidence. Evidence, in turn, can be understood as a mapping from probability functions to probability functions. It makes sense to think of updating a belief if we think of it as a generalized probability. We provide a way of updating beliefs that is different from and, we claim, more appropriate than that proposed by Dempster (in terms of his rule of combination), and argue that it is more appropriate. On the other hand, it does makes sense to combine two beliefs (using, say, Dempster's rule of combination) if we think of the belief functions as representing evidence. Many researchers have pointed out problems with the belief function approach; we show how all these problems can be explained as a consequence of confounding these two views of belief functions. This is joint work with Ron Fagin. The talk is completely self-contained; no previous knowledge of belief functions will be assumed.