Chenze Yu

1,000 hours of human touch data unlocks a 37% jump in vision-only dexterous manipulation success, revealing touch as a powerful pretraining signal beyond its direct use in policies.

This work introduces DexTouch-WM, an action-conditioned world model that learns from scalable human touch to jointly predict future RGB observations and bilateral tactile dynamics, and shows that scalable human interaction provides a complementary data axis for learning dexterous robot world models.

Trustworthiness in embodied intelligence isn't just about performance; it's about managing risk across a multi-layered framework that ensures safety and reliability in real-world applications.