GFlowCausal: Generative Flow Networks for Causal Discovery

TL;DR

GFlowCausal introduces a generative flow network approach to causal discovery, transforming DAG search into a sequential generation process that efficiently scales and guarantees acyclicity.

Contribution

It proposes a novel generative flow network method for causal discovery, enabling scalable and effective DAG learning from observational data.

Findings

Outperforms existing methods on synthetic datasets.

Effective in large-scale causal discovery tasks.

Ensures acyclicity through a transitive closure module.

Abstract

Causal discovery aims to uncover causal structure among a set of variables. Score-based approaches mainly focus on searching for the best Directed Acyclic Graph (DAG) based on a predefined score function. However, most of them are not applicable on a large scale due to the limited searchability. Inspired by the active learning in generative flow networks, we propose a novel approach to learning a DAG from observational data called GFlowCausal. It converts the graph search problem to a generation problem, in which direct edges are added gradually. GFlowCausal aims to learn the best policy to generate high-reward DAGs by sequential actions with probabilities proportional to predefined rewards. We propose a plug-and-play module based on transitive closure to ensure efficient sampling. Theoretical analysis shows that this module could guarantee acyclicity properties effectively and the consistency between final states and fully-connected graphs. We conduct extensive experiments on both synthetic and real datasets, and results show the proposed approach to be superior and also performs well in a large-scale setting.