Energy Efficient Arithmetic in AI Times
Tue 14 May
Seminars and Conferences

Energy Efficient Arithmetic in AI Times

Abstract
The number representation has a significant impact on the performance, the storage and the energy efficiency of computers. The scenario is ever changing in the Artificial Intelligence (AI) and the Machine Learning areas. In this talk, prof. Alberto Nannarelli will give an overview on the-state-of-the-art of the numerical formats used in AI, and the hardware to implement them, by outlining challenges and benefits in designing efficient floating-point units and accelerators.
About
Alberto Nannarelli Alberto Nannarelli is Professor Emeritus at the Technical University of Denmark (DTU). He graduated in electrical engineering from the University of Roma "La Sapienza", Italy, in 1988 and received the M.S. and the Ph.D. in electrical and computer engineering from the University of California at Irvine in 1995 and 1999, respectively. He worked for SGS-Thomson Microelectronics and for Ericsson Telecom as design engineer and for Rockwell Semiconductor Systems as a summer intern. From 1999, he was with the Department of Electrical Engineering, University of Roma "Tor Vergata", Italy, as a post-doc researcher, until he joined DTU in 2003. His research interests include computer arithmetic, computer architecture, and VLSI design