Guangqi Jiang
Papers
2
Total Citations
14
H-Index
2
About
Guangqi Jiang is a rising researcher at the forefront of reinforcement learning (RL) and generative AI, best known for pioneering the use of video diffusion models to solve the reward specification problem in RL. His landmark work, "Diffusion Reward: Learning Rewards via Conditional Video Diffusion," introduces a novel framework that learns reward functions directly from expert demonstration videos, bypassing the need for handcrafted reward engineering. By leveraging conditional video diffusion models, Jiang’s approach enables RL agents to infer intended behaviors from visual data alone, making it both affordable and highly effective for complex tasks. The paper has already garnered significant early attention, with over 12 citations since its 2024 publication, reflecting its impact on the intersection of computer vision and reinforcement learning. Jiang’s contributions are particularly notable for bridging generative modeling and decision-making, offering a scalable path toward more autonomous and adaptable AI systems. His work stands out as a key step in making RL more accessible and practical for real-world applications, marking him as a promising innovator in the field.
Research Focus
Key Achievements
Top Papers
- 1Diffusion Reward: Learning Rewards via Conditional Video Diffusion12 citations · 2024
- 2Diffusion Reward: Learning Rewards via Conditional Video Diffusion2 citations · 2023