SOURCE-LINKED INTELLIGENCE

Learning Multimodal One-step Flow Policy via Value-weighted Optimal Transport

arXiv · AI, language, vision and robotics · article · Sep 14, 2026 · UTC

Offline reinforcement learning aims to learn a policy solely from fixed datasets, which often contain multimodal action distributions. Flow policies can naturally represent such multimodal behaviors, but learning an efficient one-step flow policy remains challenging: standard value guidance often leads to mode collapse or exploits overestimation bias in out-of-distribution regions. To address this, we introduce One-step Flow policy via Optimal Transport (OptiFlow), a framework for one-step flow policy learning as a structured sample-allocation problem. OptiFlow jointly trains a value-aware ref

Read original source ↗ Open in workspace

recordType

paper

region

Global