Deyu Yang
Papers
3
Total Citations
51
H-Index
2
About
Deyu Yang is a robotics researcher whose work focuses on bridging the gap between robotic perception and manipulation in complex, real-world environments. His primary research areas include robotic grasping, reinforcement learning (RL), and safe robot manipulation. Yang’s major contributions lie in developing data-driven methods that enable robots to perform sophisticated tasks, such as searching for and grasping specific objects in cluttered scenes. His most cited work, "REGRAD: A Large-Scale Relational Grasp Dataset for Safe and Object-Specific Robotic Grasping in Clutter" (2022, 43 citations), provides a comprehensive dataset that captures object relationships, significantly advancing robots’ ability to perceive and interact with their surroundings. In offline reinforcement learning, his paper "Improving Offline Reinforcement Learning With in-Sample Advantage Regularization for Robot Manipulation" (2024, 6 citations) introduces a novel regularization technique to enhance policy learning from fixed datasets, improving both efficiency and safety. Additionally, his work on "Density-based Curriculum for Multi-goal Reinforcement Learning with Sparse Rewards" (2021) addresses the challenge of reward engineering in multi-goal tasks. Through these contributions, Yang is shaping the future of autonomous robotic systems capable of operating reliably in unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3