# Distributionally Robust Policy Learning under Concept Drifts

arXiv: https://arxiv.org/abs/2412.14297 · HTML: https://mlsanity.com/paper/2412.14297/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 63 references that are in the map.
- First posted on arXiv 18 Dec 2024.

## Builds on (most cited first, 20 of 20)

- [Do ImageNet Classifiers Generalize to ImageNet?](https://mlsanity.com/paper/1902.10811/) (2019)
- [Learning under Concept Drift: A Review](https://mlsanity.com/paper/2004.05785/) (2020)
- [Is Pessimism Provably Efficient for Offline RL?](https://mlsanity.com/paper/2012.15085/) (2020)
- [Robust Validation: Confident Predictions Even When Distributions Shift](https://mlsanity.com/paper/2008.04267/) (2020)
- [Bounds on the conditional and average treatment effect with unobserved confounding factors](https://mlsanity.com/paper/1808.09521/) (2018)
- [Diagnosing Model Performance Under Distribution Shift](https://mlsanity.com/paper/2303.02011/) (2023)
- [Doubly Robust Distributionally Robust Off-Policy Evaluation and Learning](https://mlsanity.com/paper/2202.09667/) (2022)
- [Policy learning "without" overlap: Pessimism and generalized empirical Bernstein's inequality](https://mlsanity.com/paper/2212.09900/) (2022)
- [Localized Debiased Machine Learning: Efficient Inference on Quantile Treatment Effects and Beyond](https://mlsanity.com/paper/1912.12945/) (2019)
- [Policy Learning with Adaptively Collected Data](https://mlsanity.com/paper/2105.02344/) (2021)
- [Distributionally Robust Batch Contextual Bandits](https://mlsanity.com/paper/2006.05630/) (2020)
- [Not all distributional shifts are equal: Fine-grained robust conformal inference](https://mlsanity.com/paper/2402.13042/) (2024)
- [Dynamic Assortment Personalization in High Dimensions](https://mlsanity.com/paper/1610.05604/) (2016)
- [Learning from a Biased Sample](https://mlsanity.com/paper/2209.01754/) (2022)
- [Optimal Multi-Distribution Learning](https://mlsanity.com/paper/2312.05134/) (2023)
- [Treatment Effect Risk: Bounds and Inference](https://mlsanity.com/paper/2201.05893/) (2022)
- [Sensitivity analysis under the f-sensitivity models: a distributional robustness perspective](https://mlsanity.com/paper/2203.04373/) (2022)
- [Risk Minimization from Adaptively Collected Data: Guarantees for Supervised and Policy Learning](https://mlsanity.com/paper/2106.01723/) (2021)
- [Diagnosing the role of observable distribution shift in scientific replications](https://mlsanity.com/paper/2309.01056/) (2023)
- [Distributionally Robust Policy Evaluation under General Covariate Shift in Contextual Bandits](https://mlsanity.com/paper/2401.11353/) (2024)

## Built on by (most cited first, 6 of 6)

- [Learning Robust Spectral Dynamics for Temporal Domain Generalization](https://mlsanity.com/paper/2505.12585/) (2025)
- [Evaluating and Learning Robust Bandit Policies Under Uncertain Causal Mechanisms](https://mlsanity.com/paper/2508.02812/) (2025)
- [Data Fusion for High-Resolution Estimation](https://mlsanity.com/paper/2508.14858/) (2025)
- [Comparison of neural network training strategies for the simulation of dynamical systems](https://mlsanity.com/paper/2512.03851/) (2025)
- [FreKoo++: Learning Continuous Spectral Dynamics for Temporal Domain Generalization](https://mlsanity.com/paper/2608.22224/) (2026)
- [Adaptive Spectral-Koopman Dynamics Modeling for Temporal Domain Generalization](https://mlsanity.com/paper/2610.02822/) (2026)

## Abstract (arXiv)

Distributionally robust policy learning aims to find a policy that performs well under the worst-case distributional shift, and yet most existing methods for robust policy learning consider the worst-case joint distribution of the covariate and the outcome. The joint-modeling strategy can be unnecessarily conservative when we have more information on the source of distributional shifts. This paper studies a more nuanced problem -- robust policy learning under the concept drift, when only the conditional relationship between the outcome and the covariate changes. To this end, we first provide a doubly-robust estimator for evaluating the worst-case average reward of a given policy under a set of perturbed conditional distributions. We show that the policy value estimator enjoys asymptotic normality even if the nuisance parameters are estimated with a slower-than-root-n rate. We then propose a learning algorithm that outputs the policy maximizing the estimated policy value within a given policy class Pi, and show that the sub-optimality gap of the proposed algorithm is of the order kappa(Pi)n-1/2, where kappa(Pi) is the entropy integral of Pi under the Hamming distance and n is the sample size. A matching lower bound is provided to show the optimality of the rate. The proposed methods are implemented and evaluated in numerical studies, demonstrating substantial improvement compared with existing benchmarks.