CS SEMINAR

Machine Learning Accelerators for AI and Big Data Analysis: GPU, TPU, and NPU

Speaker
Professor Won Woo Ro
School of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea

Chaired by
Dr PEH Li Shiuan, Provost's Chair Professor, School of Computing
peh@comp.nus.edu.sg

13 Aug 2018 Monday, 01:00 PM to 03:00 PM

Executive Classroom, COM2-04-02

Abstract:

In this talk, Professor Won Woo Ro will present recent trends in hardware systems for AI applications and big data analysis. He will mainly discuss the processing logic part such as NVIDIA Graphics Processing Units (GPU), Google TPU, and various machine learning accelerators (Intel, Microsoft, Facebook, etc.). Then, he will also present some other system design issues and techniques to develop memory hierarchy and supporting software frameworks for the accelerators. The topics will cover both academia and industrial trends. Towards the end, he will also present his experience especially when developing a neural processing unit for deep learning applications.


Biodata:

Education:
Ph.D. in Electrical Engineering, University of Southern California, USA, 2004
M.S. in Electrical Engineering, University of Southern California, USA, 1999
B.S. in Electrical Engineering, Yonsei University, Korea, 1996

Employment:
2007 ~ Present, Yonsei University, School of Electrical and Electronic Engineering, Yonsei University, Professor
2004 ~ 2007, California State University, Northridge, Department of Electrical and Computer Engineering, Assistant Professor
2006 ~ 2007, ARM Inc., Irvine, California, Contract Software Engineer, Software Development and Consulting for Design Verification - Development of RealView SoC Designer
2004, University of California Irvine, Assistant Specialist, Post-Doc Researcher
2003 ~ 2004, Apple Inc., Cupertino, California, College Intern - System level power measurement and analysis for each component of G5 based Macintosh desktop systems and servers - Development of an automated tool and test environment that enable the automatic power measurement
1999 ~ 2000, University of Southern California, Teaching Assistant, For graduate-level computer systems architecture course and broadband network architectures
1996 ~ 1997, IBM Korea, Seoul, Korea, System Engineer - Technical support for the IBM workstations - Consulting and introducing CAD systems and downstream to workstations