検証のスケーリングは視覚-言語-行動の整合性において方策学習のスケーリングより効果的になり得る
Scaling Verification Can Be More Effective than Scaling Policy Learning for Vision-Language-Action Alignment
視覚-言語-行動モデルの指示と行動のずれを減らすため、テスト時に言い換え指示と行動候補を増やして検証する手法CoVerを提案し、方策事前学習のスケーリングより高い性能を示した。
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: Jacky Kwok, Xilun Zhang, Mengdi Xu, Yuejiang Liu, Azalia Mirhoseini, Chelsea Finn, Marco Pavone
分類: cs.RO, cs.AI, cs.SY, eess.SY
原文アブストラクト
The long-standing vision of general-purpose robots hinges on their ability to understand and act upon natural language instructions. Vision-Language-Action (VLA) models have made remarkable progress toward this goal, yet their generated actions can still misalign with the given instructions. In this paper, we investigate test-time verification as a means to shrink the "intention-action gap." We first characterize the test-time scaling laws for embodied instruction following and demonstrate that jointly scaling the number of rephrased instructions and generated actions greatly increases test-time sample diversity, often recovering correct actions more efficiently than scaling each dimension independently. To capitalize on these scaling laws, we present CoVer, a contrastive verifier for vision-language-action alignment, and show that our architecture scales gracefully with additional computational resources and data. We then introduce CoVer-VLA, a hierarchical test-time verification pipeline using the trained verifier. At deployment, our framework precomputes a diverse set of rephrased instructions from a Vision-Language-Model (VLM), repeatedly generates action candidates for each instruction, and then uses the verifier to select the optimal high-level prompt and low-level action chunks. Compared to scaling policy pre-training on the same data, our verification approach yields 22% gains in-distribution and 13% out-of-distribution on the SIMPLER benchmark, with a further 45% improvement in real-world experiments. On the PolaRiS benchmark, CoVer-VLA achieves 14% gains in task progress and 9% in success rate.
