Learning Exposure Mapping Functions for Inferring Heterogeneous Peer Effects

2025-03-03

AI Summary

This work addresses the estimation of heterogeneous peer effects in network causal inference—i.e., variation in counterfactual outcomes under identical peer exposure levels, attributable to individual-level background differences. To this end, we propose, for the first time, an end-to-end differentiable exposure mapping function, moving beyond conventional fixed-rule exposure definitions (e.g., neighborhood treatment proportion) and explicitly modeling how local topological structure and edge attributes differentially shape individual exposure. Leveraging graph neural networks, we design EgoNetGNN to jointly encode node features, neighborhood topology, and edge attributes for exposure representation learning. Evaluated on synthetic and semi-synthetic network datasets, our method outperforms state-of-the-art baselines, reducing heterogeneous peer effect estimation error by 32% on average. Moreover, it demonstrates superior robustness to unknown or misspecified influence mechanisms.

Technology Category

• Machine Learning: Causal Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Causality

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: Incentives in network design for Web infrastructures and ecosystems

Abstract

In causal inference, interference refers to the phenomenon in which the actions of peers in a network can influence an individual's outcome. Peer effect refers to the difference in counterfactual outcomes of an individual for different levels of peer exposure, the extent to which an individual is exposed to the treatments, actions, or behaviors of peers. Estimating peer effects requires deciding how to represent peer exposure. Typically, researchers define an exposure mapping function that aggregates peer treatments and outputs peer exposure. Most existing approaches for defining exposure mapping functions assume peer exposure based on the number or fraction of treated peers. Recent studies have investigated more complex functions of peer exposure which capture that different peers can exert different degrees of influence. However, none of these works have explicitly considered the problem of automatically learning the exposure mapping function. In this work, we focus on learning this function for the purpose of estimating heterogeneous peer effects, where heterogeneity refers to the variation in counterfactual outcomes for the same peer exposure but different individual's contexts. We develop EgoNetGNN, a graph neural network (GNN)-based method, to automatically learn the appropriate exposure mapping function allowing for complex peer influence mechanisms that, in addition to peer treatments, can involve the local neighborhood structure and edge attributes. We show that GNN models that use peer exposure based on the number or fraction of treated peers or learn peer exposure naively face difficulty accounting for such influence mechanisms. Our comprehensive evaluation on synthetic and semi-synthetic network data shows that our method is more robust to different unknown underlying influence mechanisms when estimating heterogeneous peer effects when compared to state-of-the-art baselines.

Problem

• Automatically learn exposure mapping functions for peer effects.
• Estimate heterogeneous peer effects considering individual contexts.
• Develop GNN-based method to capture complex peer influence mechanisms.

Innovation

• Automated learning of exposure mapping functions.
• Graph neural networks for peer effect estimation.
• Incorporates local structure and edge attributes.