ON ACCEPTING ACCEPTANCE
暂无分享,去创建一个
“Believing on the Basis of the Evidence” (BBE) makes too many points to address them all here, so I limit this critique to what I view as the main problems of the approach advocated in BBE. In particular, I show that the idea of “acceptance” is a confused combination of belief and utility and so leads to problems where both need to be separately specified. I also show that the criticisms of Bayesian probabilistic inference are unfounded. The Bayesian equivalent of “acceptance” is a decision to condition belief in the target proposition on another proposition (“evidence”) that is not certain. This local does not commit the user to believe that the evidence proposition is true in any other context. Most of the problems of the “acceptance” approach in BBE come from unnecessarily making a decision about acceptance before it is known how the resulting accepted statement will be used.
[1] Peter C. Cheeseman,et al. An inquiry into computer understanding , 1988, Comput. Intell..
[2] Peter C. Cheeseman,et al. A Method of Computing Generalized Bayesian Probability Values for Expert Systems , 1983, IJCAI.
[3] A. Hasman,et al. Probabilistic reasoning in intelligent systems: Networks of plausible inference , 1991 .
[4] Judea Pearl,et al. Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.