RePO-VLA: 回復駆動型ポリシー最適化による視覚言語行動モデル
RePO-VLA: Recovery-Driven Policy Optimization for Vision-Language-Action Models
成功軌跡のみの模倣学習では実行時のずれに対処できないため、失敗軌跡を活用して回復行動を学習するフレームワークを提案。成功・回復・失敗の軌跡に役割を割り当て、価値関数で進捗を評価し、ポリシーを改善する。
詳しい要約
1. どんなもの？
2. 先行研究と比べてどこがすごい？
3. 技術・手法の肝は？
4. どうやって有効だと検証した？
5. 議論はある？
6. 次に読むべき論文は？
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: Weijia Liufu, Xiaoyu Guo, Ruiyi Chen, Jingzhi Liu, Kaidong Zhang, Xiwen Liang, Jianqi Lin, Dawei Sun, Yuze Wang, Rongtao Xu, Bingqian Lin, Bowen Yang, Tongtong Cao, Bowen Peng, Dongyu Zhang, Guangrun Wang, Min Wang, Liang Lin, Xiaodan Liang
分類: cs.RO, cs.AI
原文アブストラクト
Vision-Language-Action (VLA) models remain brittle in long-horizon, contact-rich manipulation because success-only imitation provides little supervision for execution drift, while failed rollouts are often discarded. We introduce RePO-VLA, a recovery-driven policy optimization framework that assigns distinct roles to success, recovery, and failure trajectories. RePO-VLA first applies Recovery-Aware Initialization (RAI), slicing recovery segments and resetting history so corrective actions depend on the current adverse state rather than the preceding failure. It then learns a Progress-Aware Semantic Value Function (PAS-VF), aligning spatiotemporal trajectory features with instructions and successful references. The resulting labels salvage useful failure prefixes via reliability decay, while low-value labels mark drift and terminal breakdowns, teaching differences among nominal, failed, and corrective actions. The data engine turns adverse states into planner-generated or human-collected corrective rollouts, teaching recovery to the success manifold. Value-Conditioned Refinement (VCR) trains the policy to prefer high-progress actions. At deployment, a fixed high value ($v=1.0$) biases actions toward the learned success manifold without online failure detectors or heuristic retries. We introduce FRBench, with standardized error injection and recovery-focused evaluation. Across simulated and real-world bimanual tasks, RePO-VLA improves robustness, raising adversarial success from 20% to 75% on average and up to 80% in scaled real-world trials.
