mlsanityData to 14 Oct 2026

Functional Natural Policy Gradients

arXiv 2603.28681 ↗First posted 30 Mar 2026stat.ML cs.LGcited by 1

On the map: Deep reinforcement learning methods → RL theory and sample complexity

Abstract (arXiv)

We propose a cross-fitted debiasing device for policy learning from offline data. A key consequence of the resulting learning principle is √ N regret even for policy classes with complexity greater than Donsker, provided a product-of-errors nuisance remainder is O(N-1/2). The regret bound factors into a plug-in policy error factor governed by policy-class complexity and an environment nuisance factor governed by the complexity of the environment dynamics, making explicit how one may be traded against the other.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in RL theory and sample complexity, one of 15 directions in Deep reinforcement learning methods. Builds on 14 of its 30 references that are in the map. First posted on arXiv 30 Mar 2026.

Builds on

14 of 30 references.

References · 14

  1. 2015Doubly Robust Off-policy Value Evaluation for Reinforcement Learningcited by 339
  2. 2019Orthogonal Statistical Learningcited by 151
  3. 2019Double Reinforcement Learning for Efficient Off-Policy Evaluation in Markov Decision Processescited by 119
  4. 2019Efficiently Breaking the Curse of Horizon in Off-Policy Evaluation with Double Reinforcement Learningcited by 75
  5. 2018Offline Multi-Action Policy Learning: Generalization and Optimizationcited by 65
  6. 2020Fast Rates for Contextual Linear Optimizationcited by 43
  7. 2022Doubly Robust Distributionally Robust Off-Policy Evaluation and Learningcited by 34
  8. 2024Applied Causal Inference Powered by ML and AIcited by 30
  9. 2019Semi-Parametric Efficient Policy Learning with Continuous Actionscited by 26
  10. 2019More Efficient Policy Learning via Optimal Retargetingcited by 23
  11. 2024Functional Bilevel Optimization for Machine Learningcited by 20
  12. 2020Efficient Policy Learning from Surrogate-Loss Classification Reductionscited by 10
2 more
  1. 2025Learning Theory for Kernel Bilevel Optimizationcited by 2
  2. 2024Contextual Linear Optimization with Partial Feedbackcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Semiparametric Efficient Bilevel Gradient Estimationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Robust Offline Policy Learning with Observational Data from Multiple Sources
  2. 2023Federated Offline Policy Learning
  3. 2021Offline Neural Contextual Bandits: Pessimism, Optimization and Generalization
  4. 2025CAWR: Corruption-Averse Advantage-Weighted Regression for Robust Policy Optimization
  5. 2020Efficient Evaluation of Natural Stochastic Policies in Offline Reinforcement Learning
  6. 2023Offline Data Enhanced On-Policy Policy Gradient with Provable Guarantees
  7. 2023Bi-Level Offline Policy Optimization with Limited Exploration
  8. 2025Learning More with Less: A Dynamic Dual-Level Down-Sampling Framework for Efficient Policy Optimization