PrASP

A tool for probabilistic reasoning based on logic programming and first-order theories under stable model semantics. This system description paper describes the software framework PrASP (“probabilistic answer set programming”). PrASP is both an uncertainty reasoning and machine learning software and a probabilistic logic programming language based on answer set programming (ASP). Besides serving as a research software platform for non-monotonic (inductive) probabilistic logic programming, our framework mainly targets applications in the area of uncertainty stream reasoning. PrASP programs can consist of ASP (AnsProlog) as well as first-order logic formulas (with stable model semantics), annotated with conditional or unconditional probabilities or probability intervals. A number of alternative inference algorithms allow to attune the system to different task characteristics (e.g., whether or not independence assumptions can be made).