factor-graph-reduction

This paper presents a method to reduce the size of chemical reaction networks (CRNs) implementing probabilistic inference by leveraging factor graph reduction techniques, resulting in smaller CRNs while preserving belief propagation fixed points on surviving variables.