Engineering AI Edge Devices for Life

Printer-friendly versionSend by emailPDF version

Prof. Lih-Yih Chiou,
Program Director on integrated circuit Design for the Academy of Innovative Semiconductor and Sustainable Manufacturing (AISSM) and Center Director of Future Computing for Miin Wu School of Computing (SOC).
Department of Electrical Engineering, National Cheng Kung University.

Abstract

As impacts of artificial intelligent (AI) and internet of things (IoT) are increased, the amount of data transmitted to centralized computing hubs surmount rapidly. Under this situation, the concept of decentralized computation by using edge computing devices has been widely discussed because it can pre-process data and further reduce latency of transferring data to the cloud server. There are many things needed to consider when designing such an AI Edge device. In NCKU, we have been working on several aspects. (1) Energy-efficient computation engines, either digital, memory-centric based, or analog-based; (2) Light-weight custom compiler for special tailed hardware; (3) Model compression techniques to increase efficiency of AI engines; (4) Hardware security kernels against illegitimate attacks; (5) Design for biomedical applications.

Contact information

http://www.ee.ncku.edu.tw/subpage_div/teacher_new_2/index2.php?teacher_id=152

Email: lihyihatmail [dot] ncku [dot] edu [dot] tw