Rocky Duan
Papers
1
Total Citations
24
H-Index
1
About
Rocky Duan is a leading researcher in artificial intelligence and robotics, with a focus on deep reinforcement learning, domain randomization, and generative models for robotic manipulation. His most-cited work, "Domain Randomization and Generative Models for Robotic Grasping" (2018, 24 citations), addresses a critical challenge in robotics: enabling deep learning models to generalize from limited training data. Duan pioneered the use of domain randomization—systematically varying visual and physical parameters in simulation—combined with generative models to train robust grasping policies that transfer effectively to real-world environments. This approach reduces the need for extensive real-world data collection, accelerating the deployment of AI-driven robotic systems. His contributions have influenced subsequent research in sim-to-real transfer and data-efficient robot learning. Beyond this paper, Duan has advanced reinforcement learning algorithms and contributed to open-source tools that democratize AI research. His work bridges simulation and reality, making robots more adaptable and capable in unstructured settings—a key step toward practical, general-purpose automation.
Research Focus
Key Achievements
Top Papers
- 1Domain Randomization and Generative Models for Robotic Grasping24 citations · 2018