Multi-Domain Causal Representation Learning via Weak Distributional Invariances

TL;DR

This paper introduces a method for causal representation learning across multiple domains by exploiting stable distributional properties of certain latent variables, even under complex interventions, using autoencoders.

Contribution

It relaxes previous assumptions by allowing multi-node imperfect interventions and demonstrates that autoencoders can identify stable latent sets across domains.

Findings

Autoencoders can identify stable latents under relaxed assumptions.

Stable distributional properties persist across domains with complex interventions.

The method improves causal representation learning in multi-domain datasets.

Abstract

Causal representation learning has emerged as the center of action in causal machine learning research. In particular, multi-domain datasets present a natural opportunity for showcasing the advantages of causal representation learning over standard unsupervised representation learning. While recent works have taken crucial steps towards learning causal representations, they often lack applicability to multi-domain datasets due to over-simplifying assumptions about the data; e.g. each domain comes from a different single-node perfect intervention. In this work, we relax these assumptions and capitalize on the following observation: there often exists a subset of latents whose certain distributional properties (e.g., support, variance) remain stable across domains; this property holds when, for example, each domain comes from a multi-node imperfect intervention. Leveraging this observation, we show that autoencoders that incorporate such invariances can provably identify the stable set of latents from the rest across different settings.