Approximate Inference in Probabilistic Answer Set Programming for Statistical Probabilities

Abstract “Type 1” statements were introduced by Halpern in 1990 with the goal to represent statistical information about a domain of interest. These are form “x% elements share same property”. The recently proposed language PASTA (Probabilistic Answer set programming for STAtistical probabilities) extends Probabilistic Logic Programs under Distribution Semantics and allows definition this type statements. To perform exact inference, programs converted into probabilistic answer Credal Semantics. However, algorithm is infeasible scenarios when more than few random variables involved. Here, we propose several algorithms both conditional unconditional approximate inference test them on different benchmarks. results show that scale hundreds thus can manage real world domains.