FORCE: 高效VLA强化学习微调框架，提升79%成功率
FORCE: Efficient VLA Reinforcement Fine-Tuning via Value-Calibrated Warm-up and Self-Distillation
新框架FORCE让机器人学动作更快更稳，成功率飙升79%，比现有RL方法还快32.5%，不用人插手。
FORCE是一个三阶段框架，通过价值校准热身和自蒸馏来稳定VLA模型的强化学习微调。它解决了Q函数不稳定导致的初期遗忘和低质量探索数据导致的策略更新低效问题。
在模拟和真实任务上，FORCE取得了79%的绝对成功率提升，比此前RL方法高出10%，同时训练速度加快32.5%。
该框架无需人工干预即可实现稳健性能。
FORCE: Efficient VLA Reinforcement Fine-Tuning via Value-Calibrated Warm-up and Self-Distillation
Vision-Language-Action (VLA) models are often constrained by the imitation ceiling imposed by sub-optimal data. While Reinforcement Learning (RL) fine-tuning can surpass this limit, it is notoriously sample inefficient. This challenge arises from two core issues: (1) catastrophic initial unlearning due to an unstable Q-function and (2) inefficient policy updates caused by low-quality exploration data, often forcing a reliance on costly human interventions. We introduce FORCE, a 3-stage framework that stabilizes fine-tuning by tackling both issues. FORCE first incorporates a Value-Calibrated Warm-Up phase, utilizing on-policy rollouts to mitigate the distributional shift of the Q-function. Subsequently, during the online stage, this calibrated Q-function acts as a filter for both the policy's own action proposals and expert data, ensuring only high-value actions are used for the policy update. We evaluate FORCE on various simulation and real-world tasks, and the result shows that FORCE achieves a 79% absolute improvement in success rates and outperform prior RL methods by 10%, while accelerating training by 32.5%. Critically, it mitigates the common success rate drop and achieves this robust performance without human intervention, marking a significant step towards deploying capable and autonomous robotic agents.
This report is categorized under AITOP's "Models" desk. Key entities and technical topics under continuous tracking include FORCE, VLA, 强化学习, 机器人.
Source attribution: curated from arXiv cs.AI (@Shuyi Zhang, Yunfan Lou, Hongyang Cheng, Yichen Guo, Chuyao Fu, Yaoxu Lyu, Xiaojie Zhang, Haoran Li, Pengwei Wang, Zhongyuan Wang, Shanghang Zhang).
