Papers
5
Total Citations
31
H-Index
3
About
Juhyoung Lee is a researcher specializing in energy-efficient hardware design for artificial intelligence, with a particular focus on deep reinforcement learning (DRL) processors and accelerators for edge and mobile autonomous systems. His most notable contribution is the development of OmniDRL, a highly efficient DRL processor achieving 4.18 TFLOPS performance at an impressive 29.3 TFLOPS/W energy efficiency — a landmark design that introduced dual-mode weight compression, group-sparse training, and an on-chip sparse weight transposer to enable practical DRL training on resource-constrained edge devices. Published across both conference and journal venues in 2021 and 2022, the OmniDRL work has garnered over 20 citations and addresses a critical challenge in deploying adaptive AI in applications such as autonomous driving, robotics, and drones. Lee has also contributed to FPGA-based DRL acceleration featuring selective mixed-precision retraining for fast environmental adaptation. His earlier work on the POSTECH Hand 5 robotic end effector demonstrates a foundational interest in autonomous and robotic systems that has clearly shaped his hardware-algorithm co-design research trajectory. Collectively, Lee's work offers meaningful advances toward making intelligent, self-adapting systems viable at the edge.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5The development of Postech Hand 53 citations · 2004