Disentangled Graph Autoencoder for Treatment Effect Estimation

2024-12-19

AI Summary

Unobserved confounding in networked observational data leads to biased causal effect estimation. Method: We propose a decoupled causal representation learning framework that explicitly disentangles four semantically distinct factors—instrumental variables, confounders, adjustment variables, and noise—unlike conventional proxy-based approaches that treat observed variables and network structure as mere surrogates for latent confounders. To enforce statistical independence among these factors, we introduce Hilbert–Schmidt Independence Criterion (HSIC) constraints. Technically, the framework integrates variational graph autoencoders with graph neural networks to achieve network-structure-aware causal representation disentanglement. Contribution/Results: Evaluated on multiple benchmark datasets, our method reduces mean absolute error by 12.7%–23.4% over state-of-the-art baselines, significantly improving causal identifiability and overcoming the fundamental limitations of proxy-based confounding modeling.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Social networks and social learning

Abstract

Treatment effect estimation from observational data has attracted significant attention across various research fields. However, many widely used methods rely on the unconfoundedness assumption, which is often unrealistic due to the inability to observe all confounders, thereby overlooking the influence of latent confounders. To address this limitation, recent approaches have utilized auxiliary network information to infer latent confounders, relaxing this assumption. However, these methods often treat observed variables and networks as proxies only for latent confounders, which can result in inaccuracies when certain variables influence treatment without affecting outcomes, or vice versa. This conflation of distinct latent factors undermines the precision of treatment effect estimation. To overcome this challenge, we propose a novel disentangled variational graph autoencoder for treatment effect estimation on networked observational data. Our graph encoder disentangles latent factors into instrumental, confounding, adjustment, and noisy factors, while enforcing factor independence using the Hilbert-Schmidt Independence Criterion. Extensive experiments on multiple networked datasets demonstrate that our method outperforms state-of-the-art approaches.

Problem

• Estimates treatment effects from observational data.
• Addresses influence of latent confounders in networks.
• Disentangles latent factors to improve estimation precision.

Innovation

• Disentangled Graph Autoencoder
• Hilbert-Schmidt Independence Criterion
• Instrumental Confounding Adjustment