Probabilistic Logic Programming: Causal Order Uniqueness Conditions

A new paper on arXiv (2608.07230) addresses the challenge of determining causal order in probabilistic logic programming, a formalism within statistical relational artificial intelligence. The authors derive conditions under which the probabilistic information in a program uniquely determines its causal order, exploiting the relationship with Bayesian networks. They also incorporate constraints from relational structure via prescribed causal symmetries. The method verifies when interventional reasoning is unambiguous.