Probabilistic logic in dynamic domains: Particle filter with distributional clauses

We introduce a probabilistic logic programming framework to handle continuous distributions as well as dynamic domains for use in fields like robotics. The framework is based on the recently introduced notion of distributional clauses, an extension of Sato’s distribution semantics. The key contribution of this paper is the introduction of a particle filter for this formalism. The particle filter recursively updates its beliefs about the current state, where states correspond to logical interpretations and may contain continuous variables. A further contribution of this work is extending distributional clauses with stratification to support negation.status: publishe