Decision-Aware Proximal Bridge Learning for Optimal Treatment Selection

2026-05-16

AI Summary

This work addresses the challenge of causal response estimation under unmeasured confounding in continuous treatment settings, where conventional methods—focused on global fit—often overlook regions critical for optimal decision-making, thereby limiting treatment efficacy. The authors propose a decision-oriented weighted bridge learning framework that uniquely integrates proximal causal inference with decision-aware mechanisms. By employing policy-driven weights to emphasize treatment regions pivotal for optimal decisions, the approach alternately optimizes weighted bridge functions, response surfaces, policies, and adaptive weights while preserving global stability. Theoretical analysis demonstrates that the proposed weighted bridge loss effectively reduces regret in treatment selection by controlling the weighted ill-posedness constant. Empirical evaluations confirm that this method significantly enhances individualized decision performance across multiple proximal solvers.

Abstract

Individualized treatment selection with continuous actions requires accurate causal response estimation in decision-relevant regions, rather than uniformly over the entire action space. Estimating a global causal response surface and then choosing the treatment that maximizes it can therefore be suboptimal, since standard estimation objectives allocate modeling effort according to the observed treatment distribution rather than the regions that determine the optimal decision. While decision-aware approaches have been studied in unconfounded settings, this problem remains underexplored in proximal causal inference, where proxy variables and bridge functions enable identification under suitable assumptions even in the presence of hidden confounding. Despite recent progress, proximal methods have primarily focused on treatment-effect and potential-outcome estimation rather than treatment selection and optimal decision-making. To bridge this gap, we introduce a policy-targeted weighted bridge loss that emphasizes decision-relevant treatment regions while retaining global stabilization. We prove a regret bound showing that the proposed weighted bridge loss controls treatment-selection regret through a weighted ill-posedness constant. We instantiate the framework in decision-aware variants of several proximal bridge solvers, yielding practical algorithms that alternate between weighted bridge estimation, response-surface projection, policy update, and weight refinement. Empirically, we find that decision-aware weighting reduces regret across several bridge solvers, suggesting improved treatment selection in proximal settings.

Problem

• treatment selection
• proximal causal inference
• decision-aware learning
• continuous actions
• hidden confounding

Innovation

• decision-aware learning
• proximal causal inference
• bridge functions
• treatment selection
• weighted regret minimization