In-Memory Neuromorphic Computing – From Device to Architecture

Release Time:2019-12-04Number of visits:155

Speak:     Prof.Kejie Huang

Time:       14:00-15:00, Dec. 5

Location:  SIST 1D 402

Host:       Prof. Yajun Ha

Abstract:

Memory access has been one of the most critical issues in modern computing systems, especially when they are used for artificial intelligence. Resistive non-volatile memories have been considered as one of the best candidates to deal with this issue. This topic will introduce the recent progress in resistive non-volatile memory devices and their strength and challenges in the applications of neuromorphic computing. After that, a bit-stream in-memory computing neural network processor will be introduced for robust neuromorphic computing..

Bio:

Kejie Huang received his Ph.D degree from the Department of Electrical Engineering, National University of Singapore (NUS), Singapore, in 2014. He has been a principal investigator at College of Information Science Electronic Engineering, Zhejiang University (ZJU) since 2016. Prior to joining ZJU, he spent ve years in the IC design industry, including Samsung and Xilinx, two years in the Data Storage Institute, Agency for Science, Technology and Research (A*STAR), and another three years in Singapore University of Technology and Design (SUTD), Singapore. He has authored or coauthored more than 30 scientic papers in international peer-reviewed journals and conference proceedings. He holds four granted international patents, and another eight pending ones. His research interests include low power circuits and systems design using emerging non-volatile memories, architecture and circuit optimization for recongurable neuromorphic systems, machine learning, and deep learning chip design.  He is the Associate Editor of the IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMSPART II: EXPRESS BRIEFS.

Sist seminar 18228