生成ロボットポリシーにおけるシミュレーションと実機の共学習のメカニズム解析
A Mechanistic Analysis of Sim-and-Real Co-Training in Generative Robot Policies
シミュレーションと実機データを組み合わせた共学習が有効に機能するメカニズムを理論と実験で解析し、性能を左右する2つの効果（構造的表現アライメントと重要度再重み付け）を特定した。
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: Yu Lei, Minghuan Liu, Abhiram Maddukuri, Zhenyu Jiang, Yuke Zhu
分類: cs.RO, cs.AI, cs.LG
原文アブストラクト
Co-training, which combines limited in-domain real-world data with abundant surrogate data such as simulation or cross-embodiment robot data, is widely used for training generative robot policies. Despite its empirical success, the mechanisms that determine when and why co-training is effective remain poorly understood. We investigate the mechanism of sim-and-real co-training through theoretical analysis and empirical study, and identify two intrinsic effects governing performance. The first, \textbf{``structured representation alignment"}, reflects a balance between cross-domain representation alignment and domain discernibility, and plays a primary role in downstream performance. The second, the \textbf{``importance reweighting effect"}, arises from domain-dependent modulation of action weighting and operates at a secondary level. We validate these effects with controlled experiments on a toy model and extensive sim-and-sim and sim-and-real robot manipulation experiments. Our analysis offers a unified interpretation of recent co-training techniques and motivates a simple method that consistently improves upon prior approaches. More broadly, our aim is to examine the inner workings of co-training and to facilitate research in this direction.
