DFW: A Novel Weighting Scheme for Covariate Balancing and Treatment Effect Estimation

2025-08-07

AI Summary

Causal effect estimation from observational data is often compromised by selection bias, leading to covariate imbalance; conventional inverse probability weighting (IPW) suffers from sensitivity to propensity score estimation errors and high weight variance, limiting its stability and accuracy. This paper proposes Deconfounding Factor Weighting (DFW), a novel method that introduces learnable deconfounding factors to generate bounded, low-variance weights, explicitly mitigating confounding bias. Unlike IPW, DFW does not require precise propensity score estimation, naturally accommodates multiple treatment groups, and reconstructs a pseudo-population via weighting to approximate randomized trial conditions. Extensive experiments on multiple real-world and synthetic datasets demonstrate that DFW achieves significantly superior covariate balance and causal effect estimation accuracy compared to state-of-the-art methods including IPW and Covariate Balancing Propensity Score (CBPS).

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Computer Vision: Bias, Fairness & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Estimating causal effects from observational data is challenging due to selection bias, which leads to imbalanced covariate distributions across treatment groups. Propensity score-based weighting methods are widely used to address this issue by reweighting samples to simulate a randomized controlled trial (RCT). However, the effectiveness of these methods heavily depends on the observed data and the accuracy of the propensity score estimator. For example, inverse propensity weighting (IPW) assigns weights based on the inverse of the propensity score, which can lead to instable weights when propensity scores have high variance-either due to data or model misspecification-ultimately degrading the ability of handling selection bias and treatment effect estimation. To overcome these limitations, we propose Deconfounding Factor Weighting (DFW), a novel propensity score-based approach that leverages the deconfounding factor-to construct stable and effective sample weights. DFW prioritizes less confounded samples while mitigating the influence of highly confounded ones, producing a pseudopopulation that better approximates a RCT. Our approach ensures bounded weights, lower variance, and improved covariate balance.While DFW is formulated for binary treatments, it naturally extends to multi-treatment settings, as the deconfounding factor is computed based on the estimated probability of the treatment actually received by each sample. Through extensive experiments on real-world benchmark and synthetic datasets, we demonstrate that DFW outperforms existing methods, including IPW and CBPS, in both covariate balancing and treatment effect estimation.

Problem

• Addresses selection bias in observational causal effect estimation
• Improves covariate balance across treatment groups via weighting
• Provides stable bounded weights to reduce estimation variance

Innovation

• Novel weighting scheme using deconfounding factors
• Ensures bounded weights and lower variance
• Extends to multi-treatment settings naturally