|
Abstract : |
Abstract. First order probabilistic logics combine a rst order logic with a probabilistic knowledge representation. In this context, we introduce continuous Bayesian logic programs, which extend the recently introduced Bayesian logic programs to deal with continuous random variables. Bayesian logic programs tightly integrate denite logic programs with Bayesian networks. The resulting framework nicely seperates the qualitative (i.e. logical) component from the quantitative (i.e. the probabilistic) one. We also show how the quantitative component can be learned using a gradient-based maximum likelihood method. 1, |
