Treatment Effect Estimation for Exponential Family Outcomes using Neural Networks with Targeted Regularization

2025-02-11

AI Summary

This paper addresses the limitations of Gaussian assumptions, substantial bias, and lack of double robustness in causal treatment effect estimation under exponential-family distributions (e.g., binomial, Poisson). We propose a novel neural-network-based doubly robust estimator. Methodologically, we extend the targeted regularization framework to exponential families for the first time, construct low-bias estimators via von-Mises expansions of the augmented inverse probability weighting (AIPW) functional, and establish theoretical $n^{-1/2}$ convergence rates. The approach synergizes the flexible representation capacity of neural networks with the statistical interpretability of generalized linear models. Extensive experiments—including diverse synthetic settings and real-world datasets—demonstrate that our estimator consistently outperforms existing methods in estimation accuracy, robustness to model misspecification, and generalizability. This work substantially broadens the applicability of targeted regularization to non-Gaussian causal inference.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Intelligent Robots: State Estimation

Application Category

• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Neural Networks (NNs) have became a natural choice for treatment effect estimation due to their strong approximation capabilities. Nevertheless, how to design NN-based estimators with desirable properties, such as low bias and doubly robustness, still remains a significant challenge. A common approach to address this is targeted regularization, which modifies the objective function of NNs. However, existing works on targeted regularization are limited to Gaussian-distributed outcomes, significantly restricting their applicability in real-world scenarios. In this work, we aim to bridge this blank by extending this framework to the boarder exponential family outcomes. Specifically, we first derive the von-Mises expansion of the Average Dose function of Canonical Functions (ADCF), which inspires us how to construct a doubly robust estimator with good properties. Based on this, we develop a NN-based estimator for ADCF by generalizing functional targeted regularization to exponential families, and provide the corresponding theoretical convergence rate. Extensive experimental results demonstrate the effectiveness of our proposed model.

Problem

• Extend targeted regularization to exponential family outcomes
• Develop a doubly robust neural network estimator
• Provide theoretical convergence rate for ADCF estimation

Innovation

• Neural Networks with Targeted Regularization
• Exponential Family Outcomes Extension
• Doubly Robust Estimator Development