Bipartite Graphical Causal Models Resolve Ambiguity in Cyclic Systems

Bipartite Graphical Causal Models (BGCMs) use a bipartite graph of variable and equation nodes to represent systems at equilibrium. This method resolves ambiguities in standard structural causal models when performing hard interventions on cyclic causal dependencies.