"Approximate Computing for Low Power and Information Hiding"

Location: LTS Auditorium, 8080 Greenmead Drive

Speaker:
Gang Qu
Professor, Department of Electrical and Computer Engineering

Abstract:
There are many interesting advances in approximate computing recently targeting the energy efficiency in system design and execution, in particular for applications of the Internet of Things and machine learning. The basic idea is to trade computation accuracy for power and energy during all phases of the computation, from data to algorithm and hardware implementation.

In this talk, we will report some of our recent efforts in approximate computing for the low power design as well as security based information hiding. More specifically, we will first introduce a new data format and an estimate-and-recompute paradigm that can significantly facilitate approximate computation for energy efficient execution.

Then we will demonstrate with examples the potential of embedding information in approximate hardware and approximate data, as well as during approximate computation. We will briefly discuss the security applications based on such information hiding.

Speaker Bio:
Gang Qu is a professor in the Department of Electrical and Computer Engineering at UMD, where he leads the Maryland Embedded Systems and Hardware Security Lab (MeshSec) and the Wireless Sensor Laboratory.

His research activities are on trusted integrated circuit design, nano-scale hardware security primitives, energy efficient system design and wireless sensor networks.

Qu has focused recently on applications in the Internet of Things, cyber-physical systems, and machine learning. He has published approximately 200 conference papers and journal articles on these topics with several best paper awards.

Qu is an enthusiastic teacher. He has taught and co-taught various security courses, including a popular MOOC on Hardware Security through Coursera.

Qu received his doctorate in computer science from the University of California, Los Angeles (UCLA).