Offline Policy Optimization with Posterior Sampling

• URL: http://arxiv.org/abs/2605.07393v1
• Date: Fri, 08 May 2026 07:48:21 GMT
• Title: Offline Policy Optimization with Posterior Sampling
• Abstract summary: A fundamental challenge in model-based offline reinforcement learning (RL) lies in the trade-off between generalization and robustness.<n>We propose Posterior Sampling-based Policy Optimization (PSPO), which formulates dynamics modeling as a Bayesian inference process.<n> Experiments on standard benchmarks validate PSPO achieves superior performance compared to state-of-the-art baselines.
• Score: 18.74544103863437
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: A fundamental challenge in model-based offline reinforcement learning (RL) lies in the trade-off between generalization and robustness against exploitation errors in out-of-distribution (OOD) regions. While OOD samples may capture valid underlying physical dynamics, they also introduce the risk of model exploitation. Existing methods typically address this risk through excessive pessimistic regularization, which ensures robustness but often sacrifices generalization. To overcome this limitation, we propose Posterior Sampling-based Policy Optimization (PSPO), which formulates dynamics modeling as a Bayesian inference process to derive a posterior that explicitly quantifies model fidelity. Through the integration of posterior sampling and constrained policy optimization, our method leverages dynamics-consistent OOD transitions for generalization while ensuring robustness against model exploitation. Theoretically, we formulate Q-value estimation under posterior sampling as a stochastic approximation problem and establish its convergence. We decompose policy optimization into a sequence of constrained subproblems, demonstrating that solving these subproblems guarantees monotonic improvement until convergence. Experiments on standard benchmarks validate that PSPO achieves superior performance compared to state-of-the-art baselines.

Related papers

• Regularized Offline Policy Optimization with Posterior Hybrid Bayesian Belief [14.44966497240221]
  We propose Posterior Hybrid Bayesian Belief, which reformulates the expectation as a convex combination over a subset of dynamics models.<n>We develop an iterative regularized policy optimization algorithm that provides metric-agnostic guarantees for monotonic improvement until convergence.
  arXiv  Detail & Related papers  (2026-05-30T11:35:26Z)
• Too Correct to Learn: Reinforcement Learning on Saturated Reasoning Data [55.84428098924793]
  We propose Constrained Uniform Top-K Sampling (CUTS), a parameter-free decoding strategy enforcing structure-preserving exploration.<n>We integrate this into Mixed-CUTS, a training framework synergizing exploitative and exploratory rollouts to amplify intra-group advantage variance.<n> Notably, Mixed-CUTS improves Pass@1 accuracy on the challenging AIME25 benchmark by up to 15.1% over standard GRPO.
  arXiv  Detail & Related papers  (2026-04-20T16:43:28Z)
• Contrastive Weak-to-strong Generalization [50.5986177336082]
  We propose Contrastive Weak-to-Strong Generalization (ConG) to advance weak-to-strong generalization.<n>This framework employs contrastive decoding between pre- and post-alignment weak models to generate higher-quality samples.
  arXiv  Detail & Related papers  (2025-10-09T07:37:23Z)
• Unified Enhancement of the Generalization and Robustness of Language Models via Bi-Stage Optimization [2.502393972789905]
  We propose a bi-stage optimization framework to uniformly enhance both the generalization and robustness of LMs.<n>We show that our method significantly improves the generalization and robustness of LMs compared to other existing methods.
  arXiv  Detail & Related papers  (2025-03-19T13:50:36Z)
• Generalization Bounds of Surrogate Policies for Combinatorial Optimization Problems [53.03951222945921]
  We analyze smoothed (perturbed) policies, adding controlled random perturbations to the direction used by the linear oracle.<n>Our main contribution is a generalization bound that decomposes the excess risk into perturbation bias, statistical estimation error, and optimization error.<n>We illustrate the scope of the results on applications such as vehicle scheduling, highlighting how smoothing enables both tractable training and controlled generalization.
  arXiv  Detail & Related papers  (2024-07-24T12:00:30Z)
• Bayesian Nonparametrics Meets Data-Driven Distributionally Robust Optimization [29.24821214671497]
  Training machine learning and statistical models often involve optimizing a data-driven risk criterion. We propose a novel robust criterion by combining insights from Bayesian nonparametric (i.e., Dirichlet process) theory and a recent decision-theoretic model of smooth ambiguity-averse preferences. For practical implementation, we propose and study tractable approximations of the criterion based on well-known Dirichlet process representations.
  arXiv  Detail & Related papers  (2024-01-28T21:19:15Z)
• Model-Based Epistemic Variance of Values for Risk-Aware Policy Optimization [59.758009422067]
  We consider the problem of quantifying uncertainty over expected cumulative rewards in model-based reinforcement learning. We propose a new uncertainty Bellman equation (UBE) whose solution converges to the true posterior variance over values. We introduce a general-purpose policy optimization algorithm, Q-Uncertainty Soft Actor-Critic (QU-SAC) that can be applied for either risk-seeking or risk-averse policy optimization.
  arXiv  Detail & Related papers  (2023-12-07T15:55:58Z)
• Model-Based Offline Reinforcement Learning with Pessimism-Modulated Dynamics Belief [3.0036519884678894]
  Model-based offline reinforcement learning (RL) aims to find highly rewarding policy, by leveraging a previously collected static dataset and a dynamics model. In this work, we maintain a belief distribution over dynamics, and evaluate/optimize policy through biased sampling from the belief. We show that the biased sampling naturally induces an updated dynamics belief with policy-dependent reweighting factor, termed Pessimism-Modulated Dynamics Belief.
  arXiv  Detail & Related papers  (2022-10-13T03:14:36Z)
• DROMO: Distributionally Robust Offline Model-based Policy Optimization [0.0]
  We consider the problem of offline reinforcement learning with model-based control. We propose distributionally robust offline model-based policy optimization (DROMO)
  arXiv  Detail & Related papers  (2021-09-15T13:25:14Z)
• Safe Continuous Control with Constrained Model-Based Policy Optimization [0.0]
  We introduce a model-based safe exploration algorithm for constrained high-dimensional control. We also introduce a practical algorithm that accelerates policy search with model-generated data.
  arXiv  Detail & Related papers  (2021-04-14T15:20:55Z)
• COMBO: Conservative Offline Model-Based Policy Optimization [120.55713363569845]
  Uncertainty estimation with complex models, such as deep neural networks, can be difficult and unreliable. We develop a new model-based offline RL algorithm, COMBO, that regularizes the value function on out-of-support state-actions. We find that COMBO consistently performs as well or better as compared to prior offline model-free and model-based methods.
  arXiv  Detail & Related papers  (2021-02-16T18:50:32Z)
• Control as Hybrid Inference [62.997667081978825]
  We present an implementation of CHI which naturally mediates the balance between iterative and amortised inference. We verify the scalability of our algorithm on a continuous control benchmark, demonstrating that it outperforms strong model-free and model-based baselines.
  arXiv  Detail & Related papers  (2020-07-11T19:44:09Z)

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.