Richard Li
Papers
1
Total Citations
21
H-Index
1
About
Richard Li is a roboticist whose work bridges the gap between reinforcement learning and real-world manipulation. His research focuses on multi-object manipulation, relational reinforcement learning, and sample-efficient robot control—areas critical for advancing autonomous systems in logistics and manufacturing. Li’s most influential work, “Towards Practical Multi-Object Manipulation using Relational Reinforcement Learning” (2019, 21 citations), tackles the fundamental challenge of learning robotic tasks with sparse rewards, which often demand prohibitive amounts of data. By introducing relational structures that capture object interactions, his approach dramatically reduces the complexity and data requirements for multi-object tasks, enabling robots to generalize across diverse environments. This contribution has been recognized as a key step toward making reinforcement learning viable for practical robotics. Li’s work is notable for its focus on scalability and real-world applicability, earning him recognition in top robotics venues. For students and researchers, his research offers a blueprint for designing learning systems that are both data-efficient and capable of handling the messy, multi-object scenarios that define real-world automation.
Research Focus
Key Achievements
Top Papers
- 1