Model-agnostic meta-learners for estimating heterogeneous treatment effects over time

2024-07-07

arXiv.org

AI Summary

This paper addresses the problem of estimating time-varying heterogeneous treatment effects (HTE) in sequential settings. We propose the first model-agnostic dynamic meta-learner framework, distinct from existing methods that rely on specific base models; it supports plug-and-play integration of arbitrary black-box models (e.g., Transformers). Our method directly and efficiently estimates time-varying HTE via weighted pseudo-outcome regression combined with doubly robust estimation. Key contributions include: (1) the first systematic construction of a model-agnostic meta-learner family tailored for temporal HTE estimation; (2) theoretical error bounds characterizing convergence rates and bias–variance trade-offs; and (3) overcoming a critical bottleneck in extending static HTE meta-learning to dynamic, time-series settings. Empirical evaluation on real-world sequential data—including electronic health records—demonstrates both superior estimation accuracy and seamless compatibility with diverse predictive models.

Technology Category

• Machine Learning: Causal Learning
• Search and Optimization: Metareasoning and Metaheuristics
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Estimating heterogeneous treatment effects (HTEs) over time is crucial in many disciplines such as personalized medicine. For example, electronic health records are commonly collected over several time periods and then used to personalize treatment decisions. Existing works for this task have mostly focused on model-based learners (i.e., learners that adapt specific machine-learning models). In contrast, model-agnostic learners -- so-called meta-learners -- are largely unexplored. In our paper, we propose several meta-learners that are model-agnostic and thus can be used in combination with arbitrary machine learning models (e.g., transformers) to estimate HTEs over time. Here, our focus is on learners that can be obtained via weighted pseudo-outcome regressions, which allows for efficient estimation by targeting the treatment effect directly. We then provide a comprehensive theoretical analysis that characterizes the different learners and that allows us to offer insights into when specific learners are preferable. Finally, we confirm our theoretical insights through numerical experiments. In sum, while meta-learners are already state-of-the-art for the static setting, we are the first to propose a comprehensive set of meta-learners for estimating HTEs in the time-varying setting.

Problem

• Estimating heterogeneous treatment effects over time
• Model-agnostic meta-learners for HTEs
• Theoretical and numerical analysis of meta-learners

Innovation

• Model-agnostic meta-learners
• Weighted pseudo-outcome regressions
• Time-varying treatment effects estimation