Zicheng Duan
Papers
2
Total Citations
13
H-Index
2
About
Zicheng Duan is a researcher at the forefront of efficient deep learning and embodied AI, with a focus on bridging the gap between resource-constrained hardware and high-performance perception. His work centers on model compression, particularly through automated network pruning, and on developing robust perception systems for physical robotics. Duan’s most cited paper, "ABCP: Automatic Blockwise and Channelwise Network Pruning via Joint Search" (2022, 9 citations), introduces a novel method that jointly optimizes blockwise and channelwise pruning, enabling deep models to run efficiently on devices with limited computational power—a critical need for real-time robotic detection. His second highly cited work, "Neurons Perception Dataset for RoboMaster AI Challenge" (2022, 4 citations), provides a specialized dataset for physical robot competitions, supporting the development of perception algorithms that transition from virtual simulations to real-world, dynamic environments. By tackling the challenges of deploying AI on resource-constrained platforms and advancing perception for physical robotics, Duan’s contributions are paving the way for more practical, autonomous systems in competitive and industrial settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2Neurons Perception Dataset for RoboMaster AI Challenge4 citations · 2022