NeoMaPy: calcul de MAP inference sur des graphs de connaissance temporels
Résumé
Markov Logic Networks (MLN) are used for reasoning on uncertain and inconsistent temporal data. We proposed the TMLN (Temporal Markov Logic Network) which extends them with sorts/types, weights on rules and facts, and various temporal consistencies. The NeoMaPy framework integrates it in a knowledge graph based on conflict graphs, which offers flexibility for reasoning with parametric Maximum A Posteriori (MAP) inferences, efficiency thanks to an optimistic heuristic and interactive graph visualization for results explanation.