Estimating Heterogeneous Causal Effect on Networks via Orthogonal Learning

2025-09-22

AI Summary

This paper addresses the estimation of heterogeneous causal effects under network interference, where the primary challenges are disentangling direct effects from spillover effects and mitigating structural confounding induced by network homophily. To this end, we propose a two-stage orthogonal learning framework: in Stage I, graph neural networks model complex network dependencies to estimate nuisance parameters; in Stage II, an attention-based interference model—combined with Neyman orthogonality and cross-fitting—enables robust, unbiased estimation of unit-level direct and spillover effects. The estimator remains consistent under model misspecification and strong network dependence. It accurately identifies critical neighbor subgroups and recovers the directionality of spillovers. Empirically validated across epidemiological, political, and economic applications, the method offers both strong practical applicability and interpretable causal insights.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• Economics, Online Markets and Human Computation: Social networks and social learning

Abstract

Estimating causal effects on networks is important for both scientific research and practical applications. Unlike traditional settings that assume the Stable Unit Treatment Value Assumption (SUTVA), interference allows an intervention/treatment on one unit to affect the outcomes of others. Understanding both direct and spillover effects is critical in fields such as epidemiology, political science, and economics. Causal inference on networks faces two main challenges. First, causal effects are typically heterogeneous, varying with unit features and local network structure. Second, connected units often exhibit dependence due to network homophily, creating confounding between structural correlations and causal effects. In this paper, we propose a two-stage method to estimate heterogeneous direct and spillover effects on networks. The first stage uses graph neural networks to estimate nuisance components that depend on the complex network topology. In the second stage, we adjust for network confounding using these estimates and infer causal effects through a novel attention-based interference model. Our approach balances expressiveness and interpretability, enabling downstream tasks such as identifying influential neighborhoods and recovering the sign of spillover effects. We integrate the two stages using Neyman orthogonalization and cross-fitting, which ensures that errors from nuisance estimation contribute only at higher order. As a result, our causal effect estimates are robust to bias and misspecification in modeling causal effects under network dependencies.

Problem

• Estimating heterogeneous causal effects under network interference and dependencies
• Addressing confounding between structural correlations and causal effects
• Balancing expressiveness and interpretability in network causal inference

Innovation

• Graph neural networks estimate network topology nuisances
• Attention-based interference model infers causal effects
• Neyman orthogonalization ensures robust estimation under dependencies