Max Robotics — Global Service Robot Directory + US Certification

About

Jiangeng Li is a researcher whose work lies at the intersection of reinforcement learning, imitation learning, and robotics, with a particular focus on solving challenging problems in robotic control. His major contributions include developing model-free approaches for reinforcement learning with sparse rewards, notably through his work on "Efficient hindsight reinforcement learning using demonstrations for robotic tasks with sparse rewards" (2020, 22 citations), which addresses the critical challenge of learning from limited feedback. Li has also advanced adversarial imitation learning, creating methods that work with low-quality and mixed demonstrations from multiple sources, as seen in his papers "Off-policy adversarial imitation learning for robotic tasks with low-quality demonstrations" (2020, 13 citations) and "Adversarial imitation learning with mixed demonstrations from multiple demonstrators" (2021, 10 citations). His earlier work on inverse kinematics for manipulators using fuzzy logic (2002, 8 citations) demonstrates a long-standing interest in robotic manipulation. More recently, Li has explored offline reinforcement learning with Anderson acceleration (2022, 2 citations), pushing the boundaries of data-efficient robot learning. With over 55 total citations, his research is valuable for students and researchers seeking practical, sample-efficient methods for real-world robotic applications.