paper-with-me

Papers

Flow-Based Policy for Online Reinforcement Learning

2025-06-15 · Lei Lv, Yunfei Li, Yu Luo, Fuchun Sun, Tao Kong, Jiafeng Xu, Xiao Ma

We present \textbf{FlowRL}, a novel framework for online reinforcement learning that integrates flow-based policy representation with Wasserstein-2-regularized optimization. We argue that in addition to training signals, enhancing the expressiveness of the policy class is crucial for the performance gains in RL. Flow-based generative models offer such potential, excelling at capturing complex, multimodal action distributions. However, their direct application in online RL is challenging due to a fundamental objective mismatch: standard flow training optimizes for static data imitation, while RL requires value-based policy optimization through a dynamic buffer, leading to difficult optimization landscapes. FlowRL first models policies via a state-dependent velocity field, generating actions through deterministic ODE integration from noise. We derive a constrained policy search objective that jointly maximizes Q through the flow policy while bounding the Wasserstein-2 distance to a behavior-optimal policy implicitly derived from the replay buffer. This formulation effectively aligns the flow optimization with the RL objective, enabling efficient and value-aware policy learning despite the complexity of the policy class. Empirical evaluations on DMControl and Humanoidbench demonstrate that FlowRL achieves competitive performance in online reinforcement learning benchmarks.

📄 PDF Abstract BibTeX arXiv:2506.12811

Code (0)

등록된 구현이 없습니다.

Tasks

reinforcement-learningReinforcement Learning

Similar Papers 제목 키워드 기반

Flow Matching with Injected Noise for Offline-to-Online Reinforcement Learning

2026-02-20 · Yongjae Shin, Jongseong Chae, Jongeui Park, Youngchul Sung arxiv

Generative models have recently demonstrated remarkable success across diverse domains, motivating their adoption as expressive policies in reinforcement learning (RL). While they have shown strong performance in offline…

Reinforcement LearningOffline RL

Max-Entropy Reinforcement Learning with Flow Matching and A Case Study on LQR

2025-12-29 · Yuyang Zhang, Yang Hu, Bo Dai, Na Li arxiv

Soft actor-critic (SAC) is a popular algorithm for max-entropy reinforcement learning. In practice, the energy-based policies in SAC are often approximated using simple policy classes for efficiency, sacrificing the expr…

Reinforcement Learning

One-Step Flow Policy Mirror Descent

2025-07-31 · Tianyi Chen, Haitong Ma, Na Li, Kai Wang 외 arxiv

Diffusion policies have achieved great success in online reinforcement learning (RL) due to their strong expressive capacity. However, the inference of diffusion policy models relies on a slow iterative sampling process,…

Reinforcement Learning

Reinforcement Fine-Tuning of Flow-Matching Policies for Vision-Language-Action Models

2025-10-11 · Mingyang Lyu, Yinqian Sun, Erliang Lin, Huangrui Li 외 arxiv

Vision-Language-Action (VLA) models such as OpenVLA, Octo, and $π_0$ have shown strong generalization by leveraging large-scale demonstrations, yet their performance is still fundamentally constrained by the quality and …

Reinforcement Learning

Score-Based One-step MeanFlow Policy Optimization

2026-05-22 · Kyungyoon Kim, Donghyeon Ki, Hee-Jun Ahn, Byung-Jun Lee arxiv

Diffusion and flow matching have emerged as expressive policy classes in reinforcement learning, but their reliance on multi-step denoising imposes substantial computational overhead at inference time, which is particula…

Reinforcement Learning