Kuan Fang conducts research at the intersection of robotics, machine learning, and computer vision. His research aims to enable robots to perform diverse and complex tasks in unstructured environments using deep learning. To achieve this, his lab develops scalable algorithms and systems for robot perception and control, focusing on: acquiring versatile and generalizable skills for visuomotor control by learning from massive and diverse data; continuously improving the capabilities of robots through autonomous data collection and generation; boosting generalization to novel tasks, environments, and robots by integrating knowledge from various sources.