Slicing of probabilistic programs based on specifications

This paper presents the first slicing approach for probabilistic programs based on specifications. We show that when are accompanied by their specifications in form of pre- and post-condition, we can exploit this semantic information to produce specification-preserving slices strictly more precise than yielded conventional techniques data/control dependency. To achieve goal, our technique is backward propagation post-conditions via greatest pre-expectation transformer—the counterpart Dijkstra weakest pre-condition transformer. The termination-sensitive, allowing preserve partial as well total correctness w.r.t. It modular, featuring a local reasoning principle, formally proved correct. As fundamental technical ingredients technique, design prove sound verification condition generators establishing programs, which interest own be exploited elsewhere other purposes. On practical side, demonstrate applicability means few illustrative examples case study from modelling field. also describe an algorithm computing least among space derived technique.