aiqbench analysis
This paper introduces PReFlow, a novel policy extraction method for offline reinforcement learning that overcomes the expressiveness limits of standard Gaussian or deterministic editors by utilizing conditional refinement flows. By combining critic-based proposal selection with this flow architecture, the approach effectively models multiple high-value action modes while maintaining computational efficiency through simulation-free adjoint matching. The method demonstrates superior aggregate performance on OGBench tasks, particularly after online fine-tuning, marking a significant step forward in handling complex behaviors within offline RL frameworks.
Diffusion Policy Improvement with Proposal-Conditioned Refinement Flows
arXiv:2609.36812v1 Announce Type: cross Abstract: Diffusion and flow policies can model complex behaviors in offline reinforcement learning (RL). However, penalizing their KL divergence from the behavior policy can discourage actions having high critic values with low behavior density. Directly refining behavior proposals may be an alternative, yet Gaussian or deterministic editors limit expressiveness to represent multiple separated modes for the same proposal. In this work, we introduce Proposal-Conditioned Refinement Flows (PReFlow), a policy extraction method combining critic-based proposal selection with a conditional refinement flow. To optimize proposal selection and refinement together, we formulate a KL-regularized objective whose optimum induces a Gibbs policy over final actions under a Gaussian-smoothed behavior prior. The refinement flow can represent multiple high value modes, while a proposal-centered Gaussian reference regulates large action changes. This Gaussian reference further enables us to make use of simulation-free, closed form adjoint matching targets from sampled endpoints and critic gradients, yielding a single velocity regression loss without a backward adjoint solve. On 50 OGBench tasks, PReFlow achieves competitive offline performance and the highest aggregate score among the compared methods after online fine-tuning, reaching 91\% after 500K environment steps.
Abstract:Diffusion and flow policies can model complex behaviors in offline reinforcement learning (RL). However, penalizing their KL divergence from the behavior policy can discourage actions having high critic values with low behavior density. Directly refining behavior proposals may be an alternative, yet Gaussian or deterministic editors limit expressiveness to represent multiple separated modes for the same proposal. In this work, we introduce Proposal-Conditioned Refinement Flows (PReFlow), a policy extraction method combining critic-based proposal selection with a conditional refinement flow. To optimize proposal selection and refinement together, we formulate a KL-regularized objective whose optimum induces a Gibbs policy over final actions under a Gaussian-smoothed behavior prior. The refinement flow can represent multiple high value modes, while a proposal-centered Gaussian reference regulates large action changes. This Gaussian reference further enables us to make use of simulation-free, closed form adjoint matching targets from sampled endpoints and critic gradients, yielding a single velocity regression loss without a backward adjoint solve. On 50 OGBench tasks, PReFlow achieves competitive offline performance and the highest aggregate score among the compared methods after online fine-tuning, reaching 91\% after 500K environment steps.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Robotics (cs.RO)
arXiv:2609.36812 [cs.LG]
(or arXiv:2609.36812v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.36812 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 282 words
Related coverage
Unified Condition-Action Modeling for Accurate One-Step Action Generation
arXiv · Aug 18, 2026
Where Should Action Generation Begin? A Learnable Source Prior for Generative Robot Policies
arXiv · Jun 17, 2026
Stock Market Today Highlights, Sept 2: Sensex ends 373 pts lower at 76, 570, Nifty below 24, 000; Asian Paints, HDFC Bank fall most
thehindubusinessline.com · Sep 2, 2026
Restoration Flow Matching-Based Channel Refinement and Equalization Correction for MIMO Semantic Communications
arXiv · Jul 28, 2026