Papers
15
Total Citations
202
H-Index
6
About
Jun Jin is a robotics and AI researcher whose work spans autonomous navigation, embodied AI, assistive robotics, and imitation learning. His research is unified by a commitment to enabling robots to operate intelligently and safely alongside humans in real-world environments. Jin's most influential contribution, "Mapless Navigation among Dynamics with Social-safety-awareness" (2020, 71 citations), introduced a reinforcement learning framework that balances a robot's own collision avoidance with its social impact on surrounding pedestrians — a meaningful step forward in human-aware robot navigation. His 2023 work, EmbodiedGPT (41 citations), reflects his engagement with cutting-edge foundation models, presenting an end-to-end multimodal system that enables embodied agents to plan and execute long-horizon tasks through chain-of-thought reasoning. Beyond navigation and language-grounded AI, Jin has made notable contributions to assistive robotics, visual imitation learning, and eye-hand coordination, consistently exploring how robots can learn meaningful behaviors from human demonstrations without extensive supervision. His counterfactual perception and predictive representation methods demonstrate a sophisticated approach to bridging simulation and real-world deployment. With over 190 citations across his published work, Jun Jin has established himself as a versatile and impactful researcher shaping the future of intelligent, human-centered robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2EmbodiedGPT: Vision-Language Pre-Training via Embodied Chain of Thought41 citations · 2023
- 3
- 4
- 5
- 6
- 7A Generative Model-Based Predictive Display for Robotic Teleoperation6 citations · 2021
- 8
- 9
- 10A Geometric Perspective on Visual Imitation Learning4 citations · 2020