Guangliang Li
Papers
18
Total Citations
239
H-Index
7
About
Guangliang Li is a robotics and artificial intelligence researcher whose work sits at the intersection of reinforcement learning, human-robot interaction, and sim-to-real transfer. His most influential contribution, "Transferring Policy of Deep Reinforcement Learning from Simulation to Reality for Robotics" (2022, 128 citations), addresses one of the field's most pressing challenges: bridging the gap between simulated training environments and real-world robot deployment. This work has become a key reference for researchers tackling the notorious sim-to-real transfer problem. Beyond policy transfer, Li has made significant strides in interactive and imitation learning, developing frameworks that allow robots to learn from human demonstrations and evaluative feedback — including natural, implicit signals rather than cumbersome manual inputs. His work on the social robot Haru demonstrates a particular commitment to affective computing, enabling robots to express empathy, communicate emotion, and adapt behaviors through human interaction. His GAN-based and model-based adversarial imitation learning approaches push these methods into more complex, high-dimensional environments. Collectively, Li's research empowers robots to learn efficiently and safely from ordinary people, advancing the vision of socially intelligent, practically deployable robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Automating Behavior Selection for Affective Telepresence Robot10 citations · 2021
- 4Personalized Storytelling with Social Robot Haru10 citations · 2022
- 5
- 6Shaping Affective Robot Haru’s Reactive Response8 citations · 2021
- 7
- 8
- 9Realizing full-body control of humanoid robots7 citations · 2024
- 10