Score-Based One-step MeanFlow Policy Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Kyungyoon, Ki, Donghyeon, Ahn, Hee-Jun, Lee, Byung-Jun
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918518188933120
author Kim, Kyungyoon
Ki, Donghyeon
Ahn, Hee-Jun
Lee, Byung-Jun
author_facet Kim, Kyungyoon
Ki, Donghyeon
Ahn, Hee-Jun
Lee, Byung-Jun
contents 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 particularly problematic in online RL. MeanFlow offers a promising alternative by learning an average velocity field that maps noise to data in a single network evaluation. However, MeanFlow typically requires samples from the target distribution to construct its target velocity field, which are unavailable in online RL. We propose Score-Based One-step MeanFlow Policy Optimization (SOM), an actor-critic algorithm that resolves this by constructing the target velocity field directly from the Q-function via score estimation and a probability flow ODE, thereby concentrating probability mass on high-value modes. In the fully online RL setting, SOM achieves state-of-the-art performance on locomotion tasks with a single generation step, while substantially reducing both training and inference time compared to prior diffusion- and flow-matching-based policies.
format Preprint
id arxiv_https___arxiv_org_abs_2605_23365
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Score-Based One-step MeanFlow Policy Optimization
Kim, Kyungyoon
Ki, Donghyeon
Ahn, Hee-Jun
Lee, Byung-Jun
Machine Learning
Artificial Intelligence
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 particularly problematic in online RL. MeanFlow offers a promising alternative by learning an average velocity field that maps noise to data in a single network evaluation. However, MeanFlow typically requires samples from the target distribution to construct its target velocity field, which are unavailable in online RL. We propose Score-Based One-step MeanFlow Policy Optimization (SOM), an actor-critic algorithm that resolves this by constructing the target velocity field directly from the Q-function via score estimation and a probability flow ODE, thereby concentrating probability mass on high-value modes. In the fully online RL setting, SOM achieves state-of-the-art performance on locomotion tasks with a single generation step, while substantially reducing both training and inference time compared to prior diffusion- and flow-matching-based policies.
title Score-Based One-step MeanFlow Policy Optimization
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2605.23365