世界モデルのように考え、VLAのように行動する:世界モデル表現をコンパクトなロボット方策に蒸留
Think Like a World Model, Act Like a VLA: Distilling World-Model Representations into Compact Robot Policies
凍結した世界モデルの内部特徴をVLA訓練に特徴整合項として一度だけ蒸留し、推論時コストを増やさずにロボット方策の堅牢性と成功率を向上させた。
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: Trung Dao, Sankalp Yamsani, Jaden Park, Joohyung Kim, Yong Jae Lee
分類: cs.RO, cs.CV
原文アブストラクト
Vision-Language-Action (VLA) models map observations to actions with no objective that accounts for how the world responds, so their robustness is bounded primarily by data coverage. World models carry precisely that missing objective and are better grounded for it, yet rolling the future forward costs seconds per decision and rules them out of the control loop. We show the two can be separated. What a world model knows about physical scenes lives in its \emph{internal features}; generating the future is merely the objective that produced them, so the grounding can be inherited while the generative machinery is left behind. We add one feature-alignment term to ordinary VLA training: a frozen world model is run over the training frames once and cached, and the student learns to agree with that cache. No teacher is loaded during training, the projector is discarded after it, and the deployed policy is identical to the undistilled baseline, running in $32$~ms and $1.86$~GB on a consumer RTX~5090, so every gain is attributable to the representation rather than to added capacity or test-time compute. A $0.8$B student reaches $97.9\%$ on LIBERO, improves from $48.2\%$ to $50.5\%$ on RoboCasa-GR1 humanoid manipulation, and the same objective carries over to real hardware, on both a single-arm and a bimanual platform. The gain survives changes of student scale, backbone, alignment layer, and teacher, indicating a broad representational prior rather than a fragile alignment between two particular networks. Project page: https://thaw-vla.trung-dt.com/.