凍結VLAポリシーへの検索トリガー型行動投影補正：実証研究と失敗分析
Retrieval-triggered Action Projection Adjustment for Frozen VLA Policies: An Empirical Study and Failure Analysis
凍結したOpenVLAに外部メモリで検索した補正を加えるRTPCをCALVINで評価したが、効果は再現されず失敗要因を分析した研究。
著者: Jinming Wang, Ruiheng Zhang, Yong Wang, Zipan Xiong, Keyuan Luo
分類: cs.RO
原文アブストラクト
Frozen Vision-Language-Action (VLA) policies cannot incorporate previously encountered failures without updating the underlying model. We present retrieval-triggered projection correction (RTPC), an auditable external-memory wrapper for frozen OpenVLA policies that retrieves stress-collected local-depth records and applies translation-only projection correction while keeping rotation and gripper commands unchanged. RTPC is evaluated on CALVIN single-task manipulation using primary and independent replication cohorts. In the primary evaluation, RTPC reduced a depth-based near-collision or stall proxy event rate without changing task-success count. However, this reduction was not reproduced in the independent replication, and retrieved correction directions were not distinguishable from random same-family directions. Analysis identifies endpoint coupling, batch sensitivity, and the lack of a trigger-matched memory-free baseline as limitations for attributing effects to retrieval-based correction. These findings do not establish physical safety improvement or memory-specific benefits, but provide an empirical evaluation and failure analysis of retrieval-conditioned intervention for frozen VLA policies.