Fewer Steps, Better Actions: Rethinking Flow-Matching Inference for VLA Policies

Published September 21, 2026 · Category: Robotics

Overview

arXiv:2609.21216v1 Announce Type: new Abstract: Vision-language-action (VLA) policies based on flow matching generate action chunks through repeated evaluations of an action expert. Increasing the number of integration steps raises inference cost, but does not necessarily improve closed-loop success. We propose Coda, which reallocates part of this integration budget to a single learned endpoint correction. A frozen policy first completes a few-step noise-to-action trajectory; a lightweight Transformer then predicts a demonstration-supervised residual using the candidate action, source noise, and shared observation-prefix cache. Only the corrector is trained. On 50 RoboTwin Easy tasks, five-step Coda improves success from 71.64% to 74.68% over the matched five-step baseline, while reducing forward latency by 30.2% relative to the default ten-step policy. A two-step configuration achieves 71.88% success with a 2.12$\times$ speedup. An independent 13-task control shows a 5.69-percentage-point gain at nearly equal latency, supporting correction as an effective alternative to additional integration. The same design also improves frozen official SmolVLA, raising two-step success from 60.8% to 69.4%. These results show that endpoint correction improves the quality-latency trade-off of frozen flow-matching policies.

Source

Originally published at arxiv.org.

Related Articles

• MA-LIPP: Cooperative Multi-Agent Load-Aware Informative Path Planning for Heterogeneous Robot Teams
• OpenRoIS: A Community-Driven Open-Source Middleware Implementing the Robotic Interaction Service (RoIS) Framework for Physical Robots and Virtual Agents
• Stochastic Neural Signed Swept Volume for Real-time Chance-Constrained Trajectory Optimization