Yaosheng Fu
Papers
1
Total Citations
3
H-Index
1
About
Yaosheng Fu is a researcher whose work lies at the intersection of computer architecture and machine learning systems, with a particular focus on the hardware-software co-design challenges of distributed reinforcement learning (RL). His most-cited paper, "The Architectural Implications of Distributed Reinforcement Learning on CPU-GPU Systems" (2020), provides a critical analysis of how modern RL training workloads interact with heterogeneous computing platforms. Fu identifies fundamental performance bottlenecks and power inefficiencies that arise when scaling RL across CPU-GPU architectures, offering architectural insights that are essential for building more efficient next-generation AI accelerators. While his citation count is still growing, this work has already established him as a thoughtful voice in the emerging field of systems for deep RL. His contributions are particularly valuable for students and engineers seeking to understand why simply adding more GPUs does not always translate to linear training speedups in RL, and what hardware innovations are needed to bridge that gap.
Research Focus
Key Achievements
Top Papers
- 1