Learning probabilistic logic programs

Reasoning, learning and decision making in complex, uncertain domains is central to many applications of artificial intelligence, e.g., in robotics, natural language processing, social networks, bioinformatics, smart sensor networks, and automatic knowledge acquisition and integration from structured and unstructured sources such as databases and webpages. Statistical relational AI languages and systems provide an ideal framework for this. A prominent example of such a system is the probabilistic logic programming language ProbLog, which supports high level reasoning, low level perception through neural networks, and decision theory. However, ProbLog users currently have to manually encode all relevant information, which is cumbersome especially in rich domains. In this project, we will therefore develop techniques to automatically learn probabilistic logic programs from data, which will further broaden the practical applicability of the system.

Keywords:probabilistic logic programming, statistical relational learning

Disciplines:Knowledge representation and reasoning, Machine learning and decision making