Temporal Reasoning for RDF(S): A Markov Logic based Approach

In this work, we propose a formalism that is suitable to carry out temporal reasoning for probabilistic knowledge bases. In particular, we focus on detecting erroneous statements by exploiting temporal relations of facts. Therefore, we rely on RDF(S) and its associating entailment rules which provide a data representation model as well as a basic logical expressiveness. Moreover, we use Allen 19s interval algebra to express the relations of facts based on their associated temporal information. We carry out reasoning by transforming the statements and constraints to Markov Logic and compute the most probable consistent state (MAP inference) with respect to the defined constraints. Moreover, we evaluate the proposed approach in order to demonstrate its practicality and flexibility