Consistent Labeling Across Group Assignments: Variance Reduction in Conditional Average Treatment Effect Estimation
This addresses a specific issue in causal inference for researchers and practitioners, but it is incremental as it builds on existing CATE estimation methods.
The paper tackles the problem of inconsistent learning behavior in Conditional Average Treatment Effect (CATE) estimation algorithms across different group assignments, proposing a method called CLAGA that eliminates this inconsistency and shows significant performance improvements in experiments.
Numerous algorithms have been developed for Conditional Average Treatment Effect (CATE) estimation. In this paper, we first highlight a common issue where many algorithms exhibit inconsistent learning behavior for the same instance across different group assignments. We introduce a metric to quantify and visualize this inconsistency. Next, we present a theoretical analysis showing that this inconsistency indeed contributes to higher test errors and cannot be resolved through conventional machine learning techniques. To address this problem, we propose a general method called \textbf{Consistent Labeling Across Group Assignments} (CLAGA), which eliminates the inconsistency and is applicable to any existing CATE estimation algorithm. Experiments on both synthetic and real-world datasets demonstrate significant performance improvements with CLAGA.