Max Robotics — Global Service Robot Directory + US Certification

About

Zhiding Yu is a leading researcher at the intersection of computer vision, robotics, and machine learning, with a focus on bridging the gap between simulation and reality. His work is foundational for enabling robust, generalizable AI systems that operate safely in the physical world. Yu’s key contributions include pioneering domain adaptation techniques, such as his highly cited work on “Domain Stylization” (43 citations), which uses fast covariance matching to reduce the visual gap between synthetic and real data—a critical step for training autonomous driving and robotics models. He has also advanced zero-shot generalization in visual reinforcement learning with “SECANT” (14 citations), a self-expert cloning method that helps policies ignore irrelevant visual distractions. In the realm of 3D perception, his “PointDP” framework (10 citations) introduces diffusion-driven purification to defend point cloud recognition models against adversarial attacks. Most notably, Yu is the lead author of “GR00T N1” (2025), an open foundation model for generalist humanoid robots, representing a major leap toward versatile, intelligent machines capable of operating in human environments. His work consistently pushes the boundaries of how AI learns from and interacts with the real world.