Abstract:

The rapid emergence of Physical AI—where artificial intelligence directly interacts with and controls the physical world—has placed unprecedented demands on edge computing systems. Unlike cloud-centric AI, Physical AI applications such as robotics, autonomous systems, and smart industrial platforms require real-time responsiveness, high reliability, and strict energy efficiency under constrained resources. This keynote explores energy-performance aware edge computing architectures designed to meet these challenges. We discuss the fundamental trade-offs between energy consumption, latency, and computational accuracy, and introduce techniques that jointly optimize hardware, system software, and AI workloads at the edge.

Biography: 

Hyung Gyu Lee received the Ph.D. degree in Computer Science and Engineering from Seoul National University, Seoul, Korea, in 2007. He worked as a senior engineer at Samsung Electronics from 2007 to 2010 and as a research faculty member at the Georgia Institute of Technology from 2010 to 2012. He is currently an associate professor in the School of Digital Software at Duksung Women’s University, Korea. His research interests include embedded system design, low-power systems, and memory system architecture with a focus on emerging non-volatile memory and storage technologies. His current research focuses on edge AI computing for Physical AI systems.