Discrete Causal Representations from Heterogeneous Domains: A Bayesian Approach with Social Survey Applications

2026-06-04

AI Summary

This work addresses the challenge of learning discrete, high-dimensional causal representations from heterogeneous multi-environment data under distributional shift. The authors propose a novel Bayesian hierarchical modeling approach that encodes causal assumptions and interpretability requirements through graph priors and structured parameterization. For the first time, they introduce an uncertainty-aware sequential Monte Carlo sampling scheme to infer posterior distributions under soft interventions across multiple nodes. Experiments on cross-national and cross-state social survey data demonstrate that the method effectively identifies semantically meaningful latent causal concepts—such as cultural values and political attitudes—and recovers plausible causal relationships among them, thereby significantly advancing both the theory and practical application of discrete causal representation learning.

Abstract

Causal representation learning aims to infer the high-level latent causal concepts that give rise to observed low-level measurements. This is particularly relevant for heterogeneous data from different environments or domains since distribution shifts often arise through sparse, localized changes in some of the underlying causal mechanisms, while other parts of the generative process remain unchanged. Whereas identifiability of causal representations has been studied extensively, practical uncertainty-aware methods and real-world use cases remain less explored. In this work, we propose a Bayesian approach to learning causal representations from multi-environment data, focusing on the case of discrete causal concepts and unknown multi-node soft interventions. To this end, we translate causal assumptions and interpretability desiderata into suitable priors and parametric choices within a hierarchical model. We then devise an inference scheme based on sequential Monte Carlo sampling to approximate the resulting multimodal posterior. We showcase our approach through case studies on social survey data, where latent causal concepts correspond to cultural values or political opinions, measurements to survey responses, and environments to different countries or states. Our model infers meaningful high-level concepts and plausible causal relations among them, demonstrating its utility for learning causal representations of complex real-world data.

Problem

• causal representation learning
• heterogeneous domains
• discrete latent variables
• distribution shift
• social survey data

Innovation

• Bayesian causal representation learning
• discrete latent variables
• multi-environment data
• soft interventions
• sequential Monte Carlo