Max Robotics — Global Service Robot Directory + US Certification

About

Qishen Zhao is a researcher at the intersection of robotics, reinforcement learning, and advanced manufacturing, with a primary focus on enabling robots to learn complex tasks from limited or imperfect data. His most impactful work tackles the fundamental challenge of sparse rewards in robotic reinforcement learning. In his highly cited 2020 paper, "Efficient hindsight reinforcement learning using demonstrations for robotic tasks with sparse rewards" (22 citations), Zhao proposed a general, model-free approach that leverages demonstrations to dramatically improve learning efficiency when rewards are rare. He further advanced the field by addressing the practical problem of low-quality demonstrations, introducing off-policy adversarial imitation learning techniques that allow robots to learn effectively from imperfect human teachers. His research on adversarial imitation learning with mixed demonstrations from multiple demonstrators (2021, 10 citations) provides a robust framework for learning from heterogeneous sources. Beyond learning algorithms, Zhao has also contributed to manufacturing process control, notably in modeling the peeling front geometry in roll-to-roll thin film transfer—a critical process for scalable production of 2D materials and flexible electronics. Through his work, Zhao is bridging the gap between theoretical reinforcement learning and real-world robotic applications, making autonomous systems more practical and data-efficient.