Structural Agnostic Modeling: Adversarial Learning of Causal Graphs

TL;DR

Structural Agnostic Modeling (SAM) is a novel adversarial learning approach that discovers causal graphs from observational data by combining distributional and independence cues within a neural network framework.

Contribution

SAM introduces an innovative adversarial learning method that jointly estimates causal structures and distributions, integrating sparsity and acyclicity constraints.

Findings

Successfully recovers causal graphs on synthetic data

Performs well on real observational datasets

Outperforms existing causal discovery methods

Abstract

A new causal discovery method, Structural Agnostic Modeling (SAM), is presented in this paper. Leveraging both conditional independencies and distributional asymmetries, SAM aims to find the underlying causal structure from observational data. The approach is based on a game between different players estimating each variable distribution conditionally to the others as a neural net, and an adversary aimed at discriminating the generated data against the original data. A learning criterion combining distribution estimation, sparsity and acyclicity constraints is used to enforce the optimization of the graph structure and parameters through stochastic gradient descent. SAM is extensively experimentally validated on synthetic and real data.