Jusuk Lee

Abstracting human demonstrations into scene graphs unlocks zero-shot generalization for dexterous manipulation policies, achieving 71% better performance on unseen tasks than imitation-heavy methods.

Bridging semantics and motion dynamics, BooST achieves superior skill transfer and adaptation in robotics, outperforming traditional methods.

3D traces can serve as a powerful, scalable representation for robot learning, outperforming traditional action-based models without requiring specific action labels.

Encoding dynamics directly into visual representations via DynaFLIP yields up to 22.5% better robot manipulation performance in out-of-distribution scenarios.