arXiv cs.AI· Seongheon Park, Heecheol Kim, Shulin Tian, Lilika Makabe, Namiko Saito, Katsushi Ikeuchi, Sharon Li, Yasuyuki Matsushita·· 14 小时前AI 评分31
DiVeR：面向 VLA 测试时扩展的决策关键验证器学习
DiVeR: Decision-Critical Verifier Learning for VLA Test-Time Scaling
AI 导读
DiVeR 通过采样动作表示的离散度估计决策关键性，将验证器学习重新加权到动作选择影响最大的状态，无需步级标注或额外环境交互。在 LIBERO、RoboCasa 及 Franka Research 3 真机实验中，DiVeR 通过更有效的验证器引导动作选择持续提升任务成功率，验证器推理开销可忽略。
正文 · 原文
Abstract:Scaling robot data and model capacity has improved Vision-Language-Action (VLA) policies, but further progress is constrained by the high cost of robotic data. Verifier-guided test-time scaling offers an efficient alternative by sampling multiple action candidates and selecting the one most likely to lead to task success at inference time. Existing classification-based verifiers learn from trajectory-level outcomes but treat all visited states equally, even though their value for candidate discrimination can vary across a trajectory. At many states, plausible actions are similar and provide limited discrimination signal, while only a sparse set of decision-critical states admits meaningfully different actions that can substantially affect downstream outcomes. To address this, we propose DiVeR, which estimates decision criticality from the dispersion of sampled action representations. DiVeR then uses this signal to reweight verifier learning toward states where action selection is most consequential, without requiring step-level annotations or additional environment interaction. Across LIBERO, RoboCasa, and real-world experiments on a Franka Research 3 robot, DiVeR consistently improves task success through more effective verifier-guided action selection, while adding negligible verifier inference overhead.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.04933 [cs.RO] |
| (or arXiv:2610.04933v2 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.04933 arXiv-issued DOI via DataCite |
Submission history
From: Seongheon Park [view email]
[v1]
Sun, 4 Oct 2026 04:25:40 UTC (10,653 KB)
[v2]
Sun, 25 Oct 2026 00:46:00 UTC (10,653 KB)
来源：arXiv cs.AI · arxiv.org