Systems and learning algorithms for probabilistic logical knowledge bases

In real world domains the information is often uncertain, hence it is of foremost importance to be able to model uncertainty and to reason over it. In this paper we show tools and learning systems under development for probabilistic structured data. Four systems will be considered and an overview of the related issues and of future work will be given. The first described system is cplint on SWISH, a web application that allows the user to write Probabilistic Logic Programs and submit the computation of the probability of queries with a web browser. Then two distributed structure learning algorithm are illustrated: SEMPRE ("distributed Structure lEarning by MaPREduce") and LEAPMR ("LEArning Probabilistic description logics by MapReduce"), the former learns new clauses of Probabilistic Logic Programs, the latter is used in the context of Probabilistic Description Logics. The last system taken into account is SML-Bench, developed by the research group AKSW of Leipzig, a benchmarking tool for structured data that has been extended to deal with algorithms for probabilistic structured data.