CS Colloquium: Veljko Pejovic
Speaker: Veljko Pejovic
Date: Wednesday, November 16th, 2022
Time: 3:30 - 4:30 pm
Location: HFH 1132
Host: Elizabeth Belding
Title: No 100% accuracy - no problem! Make Mobiles do More with Approximation
Abstract:
Our reliance on smartphones demands continual advancement of mobile computing. Yet, our computing appetites grow much faster than the current battery technology does, producing a critical strain on the mobile’s limited resources. To address this issue, we propose Approximate Mobile Computing (AMC) and take a radical stance that computation does not need to be 100% precise. We first examine situations, such as mobile video playback and mobile deep learning for human activity recognition, where the properties of the input and the limitations of human perception open space for AMC. We then develop methods that bring AMC to consumer devices, including an Android compiler framework that enables dynamic tuning of the level of approximation according to the context of usage. Finally, we discuss how AMC can be harnessed for efficient sensing and model training as well.
Bio:
Veljko Pejovic received his PhD in computer science from the University of California Santa Barbara. From 2015 he is an assistant professor at the Faculty of Computer and Information Science, University of Ljubljana, Slovenia. Prior to this, he was a Research Fellow at the Computer Science Department, University of Birmingham, UK. In Ljubljana, he is leading research on mobile computing, focusing on resource-efficient mobile systems, human-computer interaction, and cybersecurity in ubiquitous systems. His awards include the best paper nomination at ACM UbiComp and the first prize at Orange D4D challenge for his work on epidemics modeling. More about his research can be found at http://lrss.fri.uni-lj.si/