Combining Logical and Probabilistic Reasoning.
Michael Gelfond, J. Nelson Rushton, Weijun Zhu · 2006
This paper describes a family of knowledge representation problems, whose intuitive solutions require reasoning about defaults, the effects of actions, and quantitative probabilities. We describe an extension of the probabilistic logic language P-log (Baral & Gelfond & Rushton 2004), which uses “consistency restoring rules ” to tackle the problems described. We also report the results of a preliminary investigation into the efficiency of our P-log implementation, as compared with ACE(Chavira & Darwiche & Jaeger 2004), a system developed by Automated Reasoning Group at UCLA.