VLAモデルの信頼性指標としての輸送不一致
Transport Discrepancy as a Reliability Signal for Vision-Language-Action Models
フローマッチング型VLAモデルにおいて、観測特徴と行動表現間の輸送コストを信頼性指標として利用し、推論時に行動を反復修正する軽量プラグインDiGを提案した論文。
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: Wanpeng Zhang, Ye Wang, Hao Luo, Haoqi Yuan, Yicheng Feng, Chaoyi Xu, Sipeng Zheng, Qin Jin, Zongqing Lu
分類: cs.RO
原文アブストラクト
Vision-language-action (VLA) models that generate continuous action chunks via flow matching lack an internal signal for judging whether a given prediction is reliable. Distribution shift and long-horizon rollouts can push backbone representations away from the region the action head decodes reliably, yet the policy has no mechanism to detect or react to this drift. We observe that the cost of transporting observation features to the action representation in a shared feature space rises precisely when such drift occurs, providing a per-step reliability estimate without extra supervision. Building on this observation, we propose DiG (Discrepancy Gate), a lightweight plug-in module for flow-matching VLA policies. DiG computes a sliced Wasserstein transport cost between backbone features and the action expert's own input projection, maps it through an exponential gate, and uses the gate to modulate both a residual feature refinement and the training loss. At inference time, the gate enables DiG-Refinefine, an iterative refinement process that corrects action chunks before execution. Experiments on both simulation and real-world scenarios show that DiG consistently improves success rates, with the largest gains under distribution shift and on long-horizon tasks.
