Duzhen Zhang
Papers
2
Total Citations
39
H-Index
2
About
Duzhen Zhang is a leading researcher at the intersection of neuromorphic computing and reinforcement learning, pioneering the integration of biologically plausible spiking neural networks (SNNs) into deep reinforcement learning (DRL) frameworks. His work addresses a critical gap in artificial intelligence: while DNNs have achieved remarkable success in complex tasks from games to robotic control, they lack the biological realism and energy efficiency of SNNs. Zhang’s major contributions include the development of the Multi-Scale Dynamic Coding Improved Spiking Actor Network (2022, 32 citations), which introduces multi-scale temporal dynamics to enhance SNN-based policy learning. He also advanced the field with his Population-coding and Dynamic-neurons improved Spiking Actor Network (2021, 7 citations), demonstrating how diverse neuronal populations and dynamic spiking mechanisms can improve performance on robotic control tasks. By bridging the gap between neuroscience-inspired models and practical DRL applications, Zhang is laying the groundwork for more efficient, brain-like intelligent systems. His work is particularly notable for showing that SNNs can match or exceed traditional DNNs in reinforcement learning while offering superior biological plausibility and potential for low-power hardware implementation.
Research Focus
Key Achievements
Top Papers
- 1
- 2