Moment-Matching Graph-Networks for Causal Inference

TL;DR

This paper introduces a novel unsupervised deep learning framework that uses moment-matching loss functions on causal graphs to generate accurate interventional probabilities, even outside the training data range.

Contribution

It proposes a new architecture for applying moment-matching loss to causal graph edges, enabling automated sampling of conditional distributions in causal inference.

Findings

Generates out-of-sample interventional probabilities faithful to ground truth.

Capable of modeling non-linear structural equation models from observational data.

Integrates with autoencoders to incorporate causal graph structures.

Abstract

In this note we explore a fully unsupervised deep-learning framework for simulating non-linear structural equation models from observational training data. The main contribution of this note is an architecture for applying moment-matching loss functions to the edges of a causal Bayesian graph, resulting in a generative conditional-moment-matching graph-neural-network. This framework thus enables automated sampling of latent space conditional probability distributions for various graphical interventions, and is capable of generating out-of-sample interventional probabilities that are often faithful to the ground truth distributions well beyond the range contained in the training set. These methods could in principle be used in conjunction with any existing autoencoder that produces a latent space representation containing causal graph structures.