Fangxin Liu
Papers
2
Total Citations
17
H-Index
2
About
Fangxin Liu is a rising leader at the intersection of brain-inspired computing and hardware acceleration, whose work is redefining how machines process high-dimensional data. His primary research areas span hyperdimensional computing (HDC), in-memory computing architectures, and energy-efficient AI hardware. Liu’s major contribution is pioneering frameworks that bridge the gap between cognitive algorithms and physical hardware. His paper “L3E-HD” (12 citations) introduced a novel framework enabling efficient ensemble learning in high-dimensional space for language tasks, demonstrating how HDC can achieve robust learning by mapping data to neural activity patterns—a paradigm shift from traditional deep learning. Complementing this, his work “RTSA” (5 citations) tackles the critical bottleneck of collision detection in robotics, achieving sub-100 microsecond computation speeds using RRAM-TCAM based in-memory search accelerators. This breakthrough addresses a long-standing challenge where collision detection consumes over 90% of motion planning time. Liu’s impact is evident not only in his citation counts but in his ability to translate abstract cognitive principles into tangible, real-time hardware solutions. His research holds profound implications for edge computing, autonomous systems, and energy-constrained AI applications.
Research Focus
Key Achievements
Top Papers
- 1
- 2