robotics

Let It Be Simple: One-Step Action Generation for Vision-Language-Action Models

arXiv:2606.05737

summary

The paper proposes a one-step action generation method for vision‑language‑action models, using high‑noise training and a flow‑matching loss, and demonstrates strong performance on LIBERO robot benchmarks.

Abstract

Generating diverse images from sparse text is hard; generating compact actions from rich observations is easier. From the condition-target view, Vision-Language-Action (VLA) thus aligns with image-to-text, not text-to-image. We formalize this view through the irreducible velocity loss of standard flow matching and validate it with a controlled 8-mode toy experiment and image-to-text MNIST task. We then show that high-noise training boosts one-step VLA decoding on standard LIBERO, achieving 95.6% on LIBERO-Long, and remains competitive across LIBERO-Plus, LIBERO-Pro, and real-world robot tasks, while ablations that weaken the condition or expand the horizon predictably erase the one-step gain. These results suggest that whether one-step action generation works in VLA depends not on specialized training, but on the condition-target structure.

13 pages, 10 figures

Topics & keywords

#vision-language-action#one-step action generation#flow matching#robot learning#benchmark evaluationirreducible velocity losshigh-noise trainingone-step decodingLIBEROcondition-target structure
Let It Be Simple: One-Step Action Generation for Vision-Language-Action Models · wovepaper