Insik Yoon
Papers
2
Total Citations
116
H-Index
2
About
Dr. Insik Yoon is a leading innovator in energy-efficient neuromorphic computing, with a focus on enabling autonomous intelligence at the edge. His research centers on mixed-signal integrated circuit design, specifically pioneering time-domain processing for hardware acceleration of reinforcement learning (RL) and deep neural networks. Dr. Yoon’s major contributions include the development of the first fully integrated, time-domain mixed-signal neuromorphic accelerator for RL, demonstrated in a 55nm CMOS process. This groundbreaking work, detailed in his most-cited paper (72 citations), introduced stochastic synapses to efficiently implement the exploration-exploitation trade-off critical to RL, enabling autonomous micro-robots to learn from real-time environmental feedback without cloud connectivity. His subsequent paper (44 citations) refined this architecture, achieving a remarkable 1.25-pJ/MAC energy efficiency across a wide voltage range (1.0–0.4V), proving that complex learning algorithms can run on severely power-constrained platforms. By merging analog-inspired computation with digital robustness, Dr. Yoon has laid the essential foundation for truly intelligent, self-learning agents in applications ranging from autonomous mobile robots to implantable medical devices, pushing the boundaries of what is possible at the sensor edge.
Research Focus
Key Achievements
Top Papers
- 1
- 2