Optimal Value Inference for Reinforcement Learning

• URL: http://arxiv.org/abs/2609.09981v1
• Date: Wed, 09 Sep 2026 10:10:55 GMT
• Title: Optimal Value Inference for Reinforcement Learning
• Abstract summary: We study offline inference for the optimal value in reinforcement learning.<n>Two new nuisances are derived as fixed points of a self-induced Bellman equation.<n>We propose a debiased estimator through the Neymanity and establish its normality under diverging horizons.
• Score: 17.717171821276427
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We study offline inference for the optimal value in reinforcement learning. Two new nuisances are derived as fixed points of a self-induced Bellman equation, in which we approximate the maximum Bellman operator by its softmax correspondence. We propose a debiased estimator through the Neyman orthogonality and establish its asymptotic normality under diverging horizons even when the behavior policy changes with time, as long as the nuisances have the statistical rates that can be achieved by many machine learning methods. We provide a concrete estimating procedure for these nuisances and show they can lead to valid inference. Synthetic experiments validate the numerical performance of our inference method, and we implement it in real-life decision-making problems, including bike repositioning and AI agentic tool use.

Related papers

• OEUVRE: OnlinE Unbiased Variance-Reduced loss Estimation [23.762163604982366]
  We introduce OEUVRE, an estimator that evaluates each incoming sample on the function learned at the current and previous time steps.<n>We use algorithmic stability, a property satisfied by many popular online learners, for optimal updates and prove consistency, convergence rates, and concentration bounds for our estimator.
  arXiv  Detail & Related papers  (2025-10-26T16:41:17Z)
• Inverse Reinforcement Learning Using Just Classification and a Few Regressions [38.71913609455455]
  Inverse reinforcement learning aims to explain observed behavior by uncovering an underlying reward.<n>We show that the population maximum-likelihood solution is characterized by a linear fixed-point equation involving the behavior policy.<n>We provide a precise characterization of the optimal solution, a generic oracle-based algorithm, finite-sample error bounds, and empirical results showing competitive or superior performance to MaxEnt IRL.
  arXiv  Detail & Related papers  (2025-09-25T13:53:43Z)
• Risk-Averse Reinforcement Learning with Itakura-Saito Loss [63.620958078179356]
  Risk-averse agents choose policies that minimize risk, occasionally sacrificing expected value.<n>We introduce a numerically stable and mathematically sound loss function based on the Itakura-Saito divergence for learning state-value and action-value functions.<n>In the experimental section, we explore multiple scenarios, some with known analytical solutions, and show that the considered loss function outperforms the alternatives.
  arXiv  Detail & Related papers  (2025-05-22T17:18:07Z)
• Accelerated zero-order SGD under high-order smoothness and overparameterized regime [79.85163929026146]
  We present a novel gradient-free algorithm to solve convex optimization problems. Such problems are encountered in medicine, physics, and machine learning. We provide convergence guarantees for the proposed algorithm under both types of noise.
  arXiv  Detail & Related papers  (2024-11-21T10:26:17Z)
• Statistical Inference for Temporal Difference Learning with Linear Function Approximation [55.80276145563105]
  We investigate the statistical properties of Temporal Difference learning with Polyak-Ruppert averaging.<n>We make three theoretical contributions that improve upon the current state-of-the-art results.
  arXiv  Detail & Related papers  (2024-10-21T15:34:44Z)
• Explicit Lipschitz Value Estimation Enhances Policy Robustness Against Perturbation [2.2120851074630177]
  In robotic control tasks, policies trained by reinforcement learning (RL) in simulation often experience a performance drop when deployed on physical hardware. We propose that Lipschitz regularization can help condition the approximated value function gradients, leading to improved robustness after training.
  arXiv  Detail & Related papers  (2024-04-22T05:01:29Z)
• In Defense of Softmax Parametrization for Calibrated and Consistent Learning to Defer [27.025808709031864]
  It has been theoretically shown that popular estimators for learning to defer parameterized with softmax provide unbounded estimates for the likelihood of deferring. We show that the cause of the miscalibrated and unbounded estimator in prior literature is due to the symmetric nature of the surrogate losses used and not due to softmax. We propose a novel statistically consistent asymmetric softmax-based surrogate loss that can produce valid estimates without the issue of unboundedness.
  arXiv  Detail & Related papers  (2023-11-02T09:15:52Z)
• Actively Learning Reinforcement Learning: A Stochastic Optimal Control Approach [3.453622106101339]
  We propose a framework towards achieving two intertwined objectives: (i) equipping reinforcement learning with active exploration and deliberate information gathering, and (ii) overcoming the computational intractability of optimal control law. We approach both objectives by using reinforcement learning to compute the optimal control law. Unlike fixed exploration and exploitation balance, caution and probing are employed automatically by the controller in real-time, even after the learning process is terminated.
  arXiv  Detail & Related papers  (2023-09-18T18:05:35Z)
• Provable Guarantees for Generative Behavior Cloning: Bridging Low-Level Stability and High-Level Behavior [51.60683890503293]
  We propose a theoretical framework for studying behavior cloning of complex expert demonstrations using generative modeling. We show that pure supervised cloning can generate trajectories matching the per-time step distribution of arbitrary expert trajectories.
  arXiv  Detail & Related papers  (2023-07-27T04:27:26Z)
• Model-Based Uncertainty in Value Functions [89.31922008981735]
  We focus on characterizing the variance over values induced by a distribution over MDPs. Previous work upper bounds the posterior variance over values by solving a so-called uncertainty Bellman equation. We propose a new uncertainty Bellman equation whose solution converges to the true posterior variance over values.
  arXiv  Detail & Related papers  (2023-02-24T09:18:27Z)
• Robust Fitted-Q-Evaluation and Iteration under Sequentially Exogenous Unobserved Confounders [9.401989343015364]
  We study robust policy evaluation and policy optimization in the presence of sequentially-exogenous unobserved confounders.<n>We provide sample complexity bounds, insights, and show effectiveness both in simulations and on real-world longitudinal healthcare data of treating sepsis.
  arXiv  Detail & Related papers  (2023-02-01T18:40:53Z)
• Reinforcement Learning with a Terminator [80.34572413850186]
  We learn the parameters of the TerMDP and leverage the structure of the estimation problem to provide state-wise confidence bounds. We use these to construct a provably-efficient algorithm, which accounts for termination, and bound its regret.
  arXiv  Detail & Related papers  (2022-05-30T18:40:28Z)
• Bellman Residual Orthogonalization for Offline Reinforcement Learning [53.17258888552998]
  We introduce a new reinforcement learning principle that approximates the Bellman equations by enforcing their validity only along a test function space. We exploit this principle to derive confidence intervals for off-policy evaluation, as well as to optimize over policies within a prescribed policy class.
  arXiv  Detail & Related papers  (2022-03-24T01:04:17Z)
• Learning with Differentiable Perturbed Optimizers [54.351317101356614]
  We propose a systematic method to transform operations into operations that are differentiable and never locally constant. Our approach relies on perturbeds, and can be used readily together with existing solvers. We show how this framework can be connected to a family of losses developed in structured prediction, and give theoretical guarantees for their use in learning tasks.
  arXiv  Detail & Related papers  (2020-02-20T11:11:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.