betaFeedback

WorksComputer Science

Distributionally-Robust Policy Learning with Observational Data

Abstract

Policy learning from observational data is fundamental to applications ranging from personalized medicine to welfare-program targeting, but standard approaches optimize empirical performance and may fail when the deployment distribution differs from the training distribution. We study distributionally-robust policy learning: given observational data, we seek a policy π∈Ππ\in Π that maximizes the worst-case policy value over an ff-divergence ball of radius ρρ around the data-generating distribution. We propose an estimator built on cross-fitted doubly-robust scores for the conditional treatment effect, paired with a DRO solver that handles KL, χ2χ^2, and CVaR ambiguity sets. We establish two convergence guarantees. First, the estimated policy attains the standard O~(n−1/2)\widetilde{O}(n^{-1/2}) rate on its DRO regret, and inherits the doubly-robust property – the rate continues to hold whenever either the outcome model or the propensity model is consistently estimated, even at slow nonparametric rates. We then propose a sample-splitting bias-correction algorithm derived from the Lagrangian dual, and show that under a Tsybakov-style margin condition on the treatment-effect, the estimator attains a fast rate of O~(n−(1+α)/(2+α))\widetilde{O}(n^{-(1+α)/(2+α)}) , interpolating between O~(1/n)\widetilde{O}(1/\sqrt{n}) and O~(1/n)\widetilde{O}(1/n) as the margin parameter αα ranges over [0,∞)[0,\infty). To our knowledge, this is the first work to incorporate a margin condition into the analysis of distributionally-robust policy learning, and the first to obtain fast rates in this setting. Empirically, on synthetic and semi-synthetic experiments calibrated to a welfare-to-work program, distributional robustness yields meaningful improvements over empirical welfare maximization when training and test distributions differ.

Introduction

Personalized decision-making from observational data is a problem faced across medicine, public policy, education, and economics. Given historical records of treatments assigned and outcomes observed, the goal is to learn a policy π:X→{0,1}\pi: \mathcal{X} \to\{0,1\} that prescribes treatment as a function of an individual’s covariates so as to maximize population welfare. The growing availability of large observational datasets has driven substantial methodological interest in this problem at the intersection of causal inference and statistical learning theory. Kitagawa and Tetenov [ [21], Athey and Wager [ [2] introduced the empirical welfare maximization framework, and designed policy learning methods with regret O(n−1/2)O(n^{-1/2}), which is also statistically optimal.

A central premise of these results, however, is that the distribution P0P_0 generating the observational data coincides with the distribution under which the learnt policy will be deployed. In practice this assumption is rarely satisfied. Welfare programs are designed in one set of counties and deployed in others Kleinberg et al. [ [22]. Clinical treatment rules are derived from one patient population and applied to another, and so on. In each case, the deployment distribution can differ from the training Preprint. distribution along covariate marginals. A policy that is optimal under P0P_0 may incur arbitrarily poor performance under such shifts. This concern about distribution shift has received substantial attention in the broader machine-learning literature on out-of-distribution generalization Liu et al. [ [26] but has been largely overlooked in the theoretical foundations of policy learning. Indeed, the bulk of theoretical work on policy learning from observational data [21, 2, 45, 19] characterizes the regret of an estimated policy under the same distribution that generated the training data. While such guarantees are valuable, they cannot certify performance under deployment-time distributional shift, and they leave open the question of how to learn policies under distributional shifts.

A natural framework for this problem is distributionally robust optimization (DRO), in which the learner optimizes against the worst-case loss over a neighbourhood of distributions, typically defined by an ff-divergence ball or a Wasserstein ball around the data-generating distribution. DRO has been studied extensively in operations research and statistics [24, 25, 42] (Duchi and Namkoong 2021) and provides a principled way to certify performance against bounded distributional shift. However, this literature has focused primarily on supervised learning objectives such as classification and regression. Its application to policy learning from observational data — where the objective involves unobserved counterfactual outcomes and must be reconstructed via an estimator (e.g. doubly-robust scores Kennedy [ [19]) — has been taken up only recently [18, 1, 20, 44], and the finite-sample rates available in that literature are all of order O~(n−1/2)\widetilde{O}(n^{-1/2}). The doubly-robust correction has remained unanalyzed.

Our Contributions: In this paper we bring these two strands of work together. We study the distributionally-robust policy learning problem where the goal is to learn a policy that maximizes the worst-case advantage over an ff-divergence ball of radius ρ\rho around the data-generating distribution. Our algorithms combine doubly-robust scores with a tractable DRO solver, and analyze their regret under three commonly used ambiguity sets: KL, χ2\chi^2, and CVaR. Finite-sample regret guarantees of order O~(n−1/2)\widetilde{O}(n^{-1/2}) for distributionally robust policy learning from observational data have been obtained before, notably by Kallus et al. [ [18] for KL and general ff-divergence ambiguity sets and by Wang et al. [ [44] under concept drift; our first result recovers such a rate in a unified treatment of the KL, χ2\chi^2, and CVaR sets. Our main contribution is to go beyond the O~(n−1/2)\widetilde{O}(n^{-1/2}) barrier: to our knowledge this is the first work to obtain fast rates of convergence for any distributionally robust learning problem under a wide-margin condition Tsybakov [ [43]. Our precise contributions are the following.

  • We propose an Algorithm that combines doubly-robust scores with an inner DRO solver to learn optimal distributionally-robust policies for KL, χ2\chi^2, and CVaR ambiguity sets. We show that the resulting policy π^n\hat{\pi}_n attains the regret bound of O(n−(β+γ))+O~(d/n)O(n^{-(\beta+\gamma)}) + \widetilde{O}(\sqrt{d/n}), under standard assumptions on the nuisance estimators1 and a smooth-parametrized policy class with dimension dd. The first term decays at the doubly-robust rate n−(β+γ)n^{-(\beta+\gamma)} and the second at O~(n−1/2)\widetilde{O}(n^{-1/2}).

  • We extend the analysis to a group-robust formulation in which the ambiguity set is defined over proportions of kk predefined subgroups rather than over the full covariate distribution. We establish corresponding regret bound that scales as O~(klog⁡k/n)\widetilde{O}(\sqrt{k\log k/n}) in the statistical error term and decomposes the doubly-robust error along subgroup sample sizes.

  • We then propose a new algorithm that exploits the Lagrangian dual of the inner DRO problem to perform sample-splitting bias correction. Under a Tsybakov-style wide-margin condition on the treatment-effect contrast, we establish a fast rate of O~(n−(1−α)/(2+α))\widetilde{O}(n^{-(1-\alpha)/(2+\alpha)}) for KL-divergence and χ2\chi^2-divergence. As α\alpha varies in [0,∞)[0,\infty) the rate can interpolate between O~(n−1/2)\widetilde{O}(n^{-1/2}) and O~(n−1)\widetilde{O}(n^{-1}) at α→∞\alpha\to\infty. To our knowledge this is the first guarantee for any distributionally-robust learning problem under a margin condition.

  • We carry out experiments2 on a semi-synthetic dataset calibrated to the California GAIN welfare-to-work program with a deliberate shift between training and test counties. The results illustrate that distributional robustness yields meaningful improvements over standard empirical welfare maximization in the shifted setting, while remaining competitive when no shift is present.

Policy learning from observational data. The problem of choosing a treatment rule from sample data goes back to Manski [ [29], who framed it as a statistical decision problem and studied minimax-regret rules. The empirical welfare maximization framework was introduced by Kitagawa and Tetenov [ [21], who established O~(n−1/2)\widetilde{O}(n^{-1/2}) regret bounds and fast rates of order O~(n−(1+α)/(2+α))\widetilde{O}(n^{-(1+\alpha)/(2+\alpha)}) under a margin condition. Athey and Wager [ [2] extended this analysis to the observational setting by incorporating doubly-robust scores into the welfare objective, demonstrating that the same O~(n−1/2)\widetilde{O}(n^{-1/2}) rate holds under product nuisance errors of order o(n−1/2)o(n^{-1/2}). Zhou et al. [ [45] considered multi-action and multi-armed extensions, and Kennedy [ [19] developed optimal doubly-robust estimators for the heterogeneous treatment effect itself. None of these works addresses distributional shift, and the regret bounds they establish are with respect to the training distribution.

Distributionally robust optimization. The DRO literature has developed a rich toolkit for solving worst-case stochastic optimization problems over ff-divergence balls and Wasserstein balls. Duchi and Namkoong [ [9] established a duality between χ2\chi^2-divergence DRO and variance-regularized empirical risk minimization which we also exploit. The convex duality that makes ff-divergence ambiguity sets tractable, and which underlies the dual reformulation our second algorithm exploits, is due to Ben-Tal et al. [ [5]; Duchi and Namkoong [2021] provided uniform convergence guarantees for ff-divergence DRO over general hypothesis classes, and Duchi et al. [2021] characterised the asymptotics of the plug-in DRO value through a generalized empirical likelihood lens, which explains why its bias is Θ(n−1/2)\Theta(n^{-1/2}). Namkoong and Duchi [ [34] and Levy et al. [ [25] developed efficient large-scale stochastic methods for DRO, the latter also giving matching lower bounds on the bias of empirical-plug-in estimators of the DRO value. A parallel line of work replaces the divergence ball with an optimal-transport ball [33, 6]; Sinha et al. [ [42] studied Wasserstein DRO with applications to adversarial robustness in supervised learning. The CVaR ambiguity set, in particular, has a long history in risk-sensitive optimization [38]. While this literature provides both algorithmic templates and statistical foundations for solving the inner DRO problem, the resulting bounds are typically of order O~(n−1/2)\widetilde{O}(n^{-1/2}), and we are aware of no prior work that establishes a fast rate for any DRO problem under a margin-style condition.

Distributionally-robust causal inference. A small but growing body of work has studied robustness to distributional shift specifically in the causal-inference setting. Mo et al. [ [32] consider distributionally-robust treatment assignment under covariate shift, where the test-time covariate marginal lies in a χ2\chi^2 ball around the training marginal, and analyze a plug-in estimator. Kallus and Zhou [ [17] study robustness to unmeasured confounding via sensitivity analysis, which can be viewed as DRO over the propensity score rather than the covariate distribution. Si et al. [ [41] consider distributionally-robust policy evaluation in offline contextual bandit settings. Qi et al. [ [35] estimate individualized decision rules that control the average lower tail of the outcome distribution, a CVaR-type criterion. Closest to our setting, Kallus et al. [ [18] develop doubly-robust estimators for distributionally-robust off-policy evaluation and learning under KL ambiguity sets, and obtain an O(n−1/2)O(n^{-1/2}) regret rate with nonparametrically estimated propensities; Wang et al. [ [44] study robust policy learning under concept drift, where only the conditional law of the outcome given the covariates is perturbed, and establish a sub-optimality gap of order O(n−1/2)O(n^{-1/2}) together with a matching lower bound. However, they only consider KL-constrained ambiguity sets and make additional assumptions regarding the asymptotic estimation rate of empirical risk minimizer. In econometrics, Adjaho and Christensen [ [1] study externally valid policy choice, characterising which welfare-maximizing policies remain robust to shifts between the experimental and target populations using optimal-transport ambiguity sets, and Kido [ [20] analyzes policy learning over Wasserstein balls. Compared to these works, our contribution is to (i) treat the KL, χ2\chi^2, and CVaR ambiguity sets for policy learning (rather than evaluation) from observational data with cross-fitted doubly-robust scores in a single framework, and (ii) establish the first fast rates in this setting under a wide-margin condition. We emphasise that the n−1/2n^{-1/2} lower bound of Wang et al. [ [44] is a worst-case statement over policy classes and distributions; the margin condition we impose (Assumption 6) is precisely what lets us move past it, exactly as the Tsybakov condition does in classification.

Group and subpopulation robustness. When the ambiguity set is indexed by the mixture proportions of kk known subgroups rather than by a divergence ball over the whole covariate distribution, the DRO objective reduces to a worst-case-over-groups criterion. This formulation was popularised in machine learning by Hashimoto et al. [ [15], who use a χ2\chi^2 ball to control the risk of latent minority groups without demographic labels, and by Sagawa et al. [ [39], who optimize the worst-group training loss directly and show that regularization is essential for worst-group generalization. Duchi et al. [ [12] give a convex procedure with finite-sample guarantees that controls the worst-case loss over all subpopulations above a given size, and Sahoo et al. [ [40] study learning under Γ\Gamma-biased sampling, where the training sample over- or under-represents groups relative to the deployment population, reducing the problem to an augmented convex risk minimization via the CVaR duality of Rockafellar et al. [ [38]. Martinez et al. [ [30] treat the worst-group criterion as a multi-objective problem and characterise minimax-Pareto-optimal classifiers. All of this work concerns supervised losses. Our group-robust extension (Equation (3)) complements this line by combining group-level ambiguity with cross-fitted doubly-robust score estimation for the causal objective, and by giving a regret bound of order O~(klog⁡k/n)\widetilde{O}(\sqrt{k\log k/n}) whose nuisance-error term decomposes along the per-group sample sizes ngn_g.

Margin conditions and fast rates. The Tsybakov margin condition originates in classification, with foundational contributions by Mammen and Tsybakov [ [28] and Tsybakov [ [43], and was developed further in the analyses of Massart and Nédélec [ [31] (ERM regret under margin via weighted empirical processes), Audibert and Tsybakov [ [3] (plug-in classifiers achieving fast and super-fast rates), and Koltchinskii [ [23] (local Rademacher complexities and oracle inequalities). Kitagawa and Tetenov [ [21] adapted the condition to policy learning to obtain fast regret rates in the empirical-welfare-maximization framework. Luedtke and Van Der Laan [ [27] invoke a closely related margin assumption in their work on inference for the value of the optimal treatment rule, where it ensures pathwise differentiability of the value parameter and enables n\sqrt{n}-consistent inference. Our Assumption 6 follows the Kitagawa–Tetenov formulation, applied to the distributionally-robust treatment-effect contrast under the worst-case distribution.

Doubly-Robust Methods in Causal Inference. Doubly-robust estimation has its origins in the semiparametric statistics and missing-data literature, where Robins et al. [ [37] introduced the augmented inverse propensity weighted (AIPW) estimator and showed that it remains consistent if either the outcome model or the propensity model is correctly specified. Bang and Robins [ [4] extended this idea to causal inference under the no-unmeasured-confounders assumption and formalised the doubly-robust property as a desirable design criterion when both nuisance components are necessarily approximated. The modern double/debiased machine learning framework of Chernozhukov et al. [ [8] reformulated this property through the lens of Neyman-orthogonal moment conditions, showing that combining orthogonal scores with cross-fitting allows arbitrary first-stage machine-learning estimators of the nuisances — converging at slow nonparametric rates — to be plugged in while preserving n−1/2n^{-1/2}-consistency and asymptotic normality of the target parameter. Kennedy [ [19] subsequently developed optimal doubly-robust estimators specifically for the heterogeneous treatment effect τ(x)\tau(x), characterising the rates achievable for this conditional functional. In parallel, the contextual-bandit and off-policy-learning literature has adapted these techniques to the sequential decision-making setting: Dudík et al. [ [13] introduced the doubly-robust estimator for contextual-bandit policy evaluation and learning, and Dudík et al. [ [14] provided a refined statistical analysis. The policy-learning theory of Athey and Wager [ [2] that we build on combines the AIPW score of Robins et al. [ [37] with empirical welfare maximization in the style of Kitagawa and Tetenov [ [21] to obtain O(n−1/2)O(n^{-1/2}) regret bounds under product-rate nuisance error o(n−1/2)o(n^{-1/2}). Our work extends this line by replacing the empirical welfare objective with a distributionally-robust one while preserving the same doubly-robust structure: Theorem 1 shows that the regret of π^\hat{\pi} inherits the standard doubly-robust decomposition, with a nuisance-product term decaying at n−(β+γ)n^{-(\beta+\gamma)} and a statistical term at O~(n−1/2)\widetilde{O}(n^{-1/2}).

Setting

We assume access to observational data D={(Xi,Yi,Wi)}i=1n\mathcal{D} = \{(X_i,Y_i,W_i)\}_{i=1}^{n} where XiX_i is co-variate of an individual, WiW_i is the treatment, YiY_i is the outcome. The goal of policy learning is to learn a policy π∈Π:X→{0,1}\pi\in\Pi: \mathcal{X} \to\{0,1\} which maximizes its value. Following Kitagawa and Tetenov [ [21], Athey and Wager [ [2] we consider the value of a policy to be the utility relative to treating no one.

V(π)=EXi∼P0[Yi(π(Xi))−Yi(0)]V(\pi) = \mathbb{E}_{X_i \sim P_0}[Y_i(\pi(X_i)) - Y_i(0)]

Let us now write τ(Xi)=E[Yi(1)−Yi(0)∣Xi]\tau(X_i)=\mathbb{E}[Y_i(1)-Y_i(0)\mid X_i] to be the conditional average treatment effect. Then it is easy to see the following expression of V(π)V(\pi), which we will also refer to as the advantage of policy π\pi.

A(π)=EX∼P0[(2π(X)−1)τ(X)]A(\pi)=\mathbb{E}_{X\sim P_0}[(2\pi(X)-1)\tau(X)]

The goal of policy learning from observation data is to learn a policy π(⋅)\pi(\cdot) that maximizes the advantage A(⋅)A(\cdot), or equivalently minimizes the following notion of regret.3

R(π)=max⁡π′∈ΠA(π′)−A(π)R(\pi)=\max_{\pi'\in\Pi} A(\pi')-A(\pi)

It is usually assumed that the underlying distribution P0P_0 over the covariates is fixed and the observed covariates {Xi}i=1n\{X_i\}_{i=1}^n are drawn from the same distribution. However, in many applications, the distribution during deployment of the policy differs significantly from the distribution P0P_0. In order to capture this phenomenon, we consider a distributionally robust version of policy learning problem. Let us define a distributionally robust version of advantage A(π)A(\pi) where the target distribution belongs to the set UP0={Q:Df(Q∥P0)≤ρ}U_{P_0}=\{Q:D_f(Q\mathbin\Vert P_0)\le\rho\}.

Af(π)=min⁡Q:Df(Q∥P0)≤ρEX∼Q[(2π(X)−1)τ(X)]A_f(\pi)=\min_{Q:D_f(Q\mathbin\Vert P_0)\le\rho}\mathbb{E}_{X\sim Q}[(2\pi(X)-1)\tau(X)]

We aim to maximize Af(π)A_f(\pi) i.e. consider the following distributionally robust policy learning problem.

max⁡π∈Πmin⁡Q:Df(Q∥P0)≤ρEX∼Q[(2π(X)−1)τ(X)](1)\max_{\pi\in\Pi}\min_{Q:D_f(Q\mathbin\Vert P_0)\le\rho}\mathbb{E}_{X\sim Q}[(2\pi(X)-1)\tau(X)] \tag*{(1)}

Our first method i.e. Algorithm 1 solves a finite-sample version of the above distributionally robust optimization problem.

max⁡π∈Πmin⁡Q:Df(Q∥P^n)≤ρ∑i=1nqi⋅(2π(Xi)−1)Γ^i.(2)\max_{\pi\in\Pi}\min_{Q:D_f(Q\mathbin\Vert\widehat P_n)\le\rho}\sum_{i=1}^{n}q_i\cdot(2\pi(X_i)-1)\widehat{\Gamma}_i. \tag*{(2)}

Here Γ^i\widehat{\Gamma}_i is a doubly robust score for estimating the treatment effect τ(Xi)\tau(X_i). Algorithm 1 builds an estimator of conditional outcomes (m^(x,w))(\widehat m(x,w)) and propensity scores (e^(x))(\widehat e(x)) and uses them to construct the doubly robust scores Γ^i\widehat{\Gamma}_i. 4 The optimization problem defined in eq. (2) might be nonconvex in π\pi, however the inner minimization problem is a convex optimization problem. We first analyze the statistical properties of the policy π^\widehat\pi.

Algorithm 1 Distributionally Robust Policy Learning from Observational Data

Input: Observational data D={(Xi,Yi,Wi)}i=13n\mathcal{D}=\{(X_i,Y_i,W_i)\}_{i=1}^{3n}.

References

  1. [1]Christopher Adjaho and Timothy Christensen. Externally valid policy choice. arXiv preprint arXiv:2205.05561, 2022.arxiv.org/abs/2205.05561
  2. [2]Susan Athey and Stefan Wager. Policy learning with observational data. Econometrica, 89(1): 133–161, 2021.arxiv.org/abs/1702.02896
  3. [3]Jean-Yves Audibert and Alexandre B. Tsybakov. Fast learning rates for plug-in classifiers. The Annals of Statistics, 35(2):608–633, 2007.DOI
  4. [4]Heejung Bang and James M Robins. Doubly robust estimation in missing data and causal inference models. Biometrics, 61(4):962–973, 2005.pubmed.ncbi.nlm.nih.gov/16401269
  5. [5]Aharon Ben-Tal, Dick den Hertog, Anja De Waegenaere, Bertrand Melenberg, and Gijs Renne. Robust solutions of optimization problems affected by uncertain probabilities. Management Science, 59(2):341–357, 2013.
  6. [6]Jose Blanchet and Karthyek Murthy. Quantifying distributional model risk via optimal transport. Mathematics of Operations Research, 44(2):565–600, 2019.arxiv.org/abs/1604.01446
  7. [7]S Boucheron, G Lugosi, and P Massart. A non asymptotic theory of independence, 2013.
  8. [8]Victor Chernozhukov, Denis Chetverikov, Mert Demirer, Esther Duflo, Christian Hansen, Whitney Newey, and James Robins. Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal, 21(1):C1–C68, 2018.
  9. [9]John Duchi and Hongseok Namkoong. Variance-based regularization with convex objectives. Journal of Machine Learning Research, 20(68):1–55, 2019.
  10. [12]John C Duchi, Tatsunori Hashimoto, and Hongseok Namkoong. Distributionally robust losses for latent covariate mixtures. Operations Research, 71(2):649–664, 2023.arxiv.org/abs/2007.13982
  11. [13]Miroslav Dudík, John Langford, and Lihong Li. Doubly robust policy evaluation and learning. In International Conference on Machine Learning, pages 1097–1104, 2011.arxiv.org/abs/1103.4601
  12. [14]Miroslav Dudík, Dumitru Erhan, John Langford, and Lihong Li. Doubly robust policy evaluation and optimization. Statistical Science, 29(4):485–511, 2014.
  13. [15]Tatsunori Hashimoto, Megha Srivastava, Hongseok Namkoong, and Percy Liang. Fairness without demographics in repeated loss minimization. In International Conference on Machine Learning, volume 80, pages 1929–1938. PMLR, 2018.
  14. [16]V Joseph Hotz, Guido W Imbens, and Jacob A Klerman. Evaluating the differential effects of alternative welfare-to-work training components: A reanalysis of the california gain program. Journal of Labor Economics, 24(3):521–566, 2006.DOI
  15. [17]Nathan Kallus and Angela Zhou. Confounding-robust policy improvement. Advances in neural information processing systems, 31, 2018.arxiv.org/abs/1805.08593
  16. [18]Nathan Kallus, Xiaojie Mao, Kaiwen Wang, and Zhengyuan Zhou. Doubly robust distributionally robust off-policy evaluation and learning. In International Conference on Machine Learning, volume 162, pages 10598–10632. PMLR, 2022.arxiv.org/abs/2202.09667
  17. [19]Edward H Kennedy. Towards optimal doubly robust estimation of heterogeneous causal effects. Electronic Journal of Statistics, 17(2):3008–3049, 2023.arxiv.org/abs/2004.14497
  18. [20]Daido Kido. Distributionally robust policy learning with wasserstein distance. arXiv preprint arXiv:2205.04637, 2022.arxiv.org/abs/2205.04637
  19. [21]Toru Kitagawa and Aleksey Tetenov. Who should be treated? empirical welfare maximization methods for treatment choice. Econometrica, 86(2):591–616, 2018.
  20. [22]Jon Kleinberg, Jens Ludwig, Sendhil Mullainathan, and Ziad Obermeyer. Prediction policy problems. American Economic Review, 105(5):491–495, 2015.pubmed.ncbi.nlm.nih.gov/27199498
  21. [23]Vladimir Koltchinskii. Local Rademacher complexities and oracle inequalities in risk minimization. The Annals of Statistics, 34(6):2593–2656, 2006.DOI
  22. [24]Daniel Kuhn, Soroosh Shafiee, and Wolfram Wiesemann. Distributionally robust optimization. Acta Numerica, 34:579–804, 2025.
  23. [25]Daniel Levy, Yair Carmon, John C Duchi, and Aaron Sidford. Large-scale methods for distributionally robust optimization. Advances in neural information processing systems, 33:8847–8860, 2020.arxiv.org/abs/2010.05893
  24. [26]Jiashuo Liu, Zheyan Shen, Yue He, Xingxuan Zhang, Renzhe Xu, Han Yu, and Peng Cui. Towards out-of-distribution generalization: A survey. arXiv preprint arXiv:2108.13624, 2021.arxiv.org/abs/2108.13624
  25. [27]Alexander R Luedtke and Mark J Van Der Laan. Statistical inference for the mean outcome under a possibly non-unique optimal treatment strategy. Annals of statistics, 44(2):713, 2016.DOI
  26. [28]Enno Mammen and Alexandre B. Tsybakov. Smooth discrimination analysis. The Annals of Statistics, 27(6):1808–1829, 1999.
  27. [29]Charles F Manski. Statistical treatment rules for heterogeneous populations. Econometrica, 72(4):1221–1246, 2004.
  28. [30]Natalia Martinez, Martin Bertran, and Guillermo Sapiro. Minimax pareto fairness: A multi objective perspective. In International Conference on Machine Learning, volume 119, pages 6755–6764. PMLR, 2020.
  29. [31]Pascal Massart and Élodie Nédélec. Risk bounds for statistical learning. The Annals of Statistics, 34(5):2326–2366, 2006.DOI
  30. [32]Weibin Mo, Zhengling Qi, and Yufeng Liu. Learning optimal distributionally robust individualized treatment rules. Journal of the American Statistical Association, 116(534):659–674, 2021.
  31. [33]Peyman Mohajerin Esfahani and Daniel Kuhn. Data-driven distributionally robust optimization using the wasserstein metric: performance guarantees and tractable reformulations. Mathematical Programming, 171(1–2):115–166, 2018.arxiv.org/abs/1505.05116
  32. [34]Hongseok Namkoong and John C Duchi. Stochastic gradient methods for distributionally robust optimization with f-divergences. In Advances in Neural Information Processing Systems, volume 29, 2016.
  33. [35]Zhengling Qi, Jong-Shi Pang, and Yufeng Liu. On robustness of individualized decision rules. Journal of the American Statistical Association, 118(543):2143–2157, 2023.DOI
  34. [36]James Riccio et al. Gain: Benefits, costs, and three-year impacts of a welfare-to-work program. california’s greater avenues for independence program. Technical report, Manpower Demonstration Research Corporation, 1994.
  35. [37]James M Robins, Andrea Rotnitzky, and Lue Ping Zhao. Estimation of regression coefficients when some regressors are not always observed. Journal of the American Statistical Association, 89(427):846–866, 1994.
  36. [38]R Tyrrell Rockafellar, Stanislav Uryasev, et al. Optimization of conditional value-at-risk. Journal of risk, 2:21–42, 2000.DOI
  37. [39]Shiori Sagawa, Pang Wei Koh, Tatsunori B Hashimoto, and Percy Liang. Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization. In International Conference on Learning Representations, 2020.arxiv.org/abs/1911.08731
  38. [40]Roshni Sahoo, Lihua Lei, and Stefan Wager. Learning from a biased sample. arXiv preprint arXiv:2209.01754, 2022.arxiv.org/abs/2209.01754
  39. [41]Nian Si, Fan Zhang, Zhengyuan Zhou, and Jose Blanchet. Distributionally robust policy evaluation and learning in offline contextual bandits. In International Conference on Machine Learning, pages 8884–8894. PMLR, 2020.
  40. [42]Aman Sinha, Hongseok Namkoong, Riccardo Volpi, and John Duchi. Certifying some distributional robustness with principled adversarial training. arXiv preprint arXiv:1710.10571, 2017.arxiv.org/abs/1710.10571
  41. [43]Alexander B Tsybakov. Optimal aggregation of classifiers in statistical learning. The Annals of Statistics, 32(1):135–166, 2004.DOI
  42. [44]Jingyuan Wang, Zhimei Ren, Ruohan Zhan, and Zhengyuan Zhou. Distributionally robust policy learning under concept drifts. In International Conference on Machine Learning, volume 267. PMLR, 2025.arxiv.org/abs/2412.14297
  43. [45]Zhengyuan Zhou, Susan Athey, and Stefan Wager. Offline multi-action policy learning: Generalization and optimization. Operations Research, 71(1):148–183, 2023.arxiv.org/abs/1810.04778

Paper details

vlnc:W_9fahcvvd
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents