CPS-IoT Week 2020 Half Day Tutorial
- Radu Marculescu (firstname.lastname@example.org), University of Texas, Austin
- Umit Y. Ogras (email@example.com), Arizona State University
Tutorial Organization and Contents:
- 9:00AM – 10:15AM: The first part will cover the fundamentals for edge computing and intelligence at the edge. The speakers will introduce the need for edge AI, existing work in the area. Then, they will discuss model compression and inference, as well as novel neural architecture space (NAS) explorations using network science.
- 10:15AM – 10:30AM: Break
- 10:30AM – 11:45AM: The second part will move into the details of hardware architectures required to support the algorithms introduced in Part 1. After introducing the state-of-the-art architectures, the speakers will present novel hardware-software solutions targeting edge AI with applications from image and human activity recognition areas.
- 11:45AM – 12:00PM: Live Demo and Concluding Remarks – The speakers will conclude the tutorial with a demonstration of distributed edge inference and human activity recognition with wearable IoT devices. Then, the speakers will answer audience questions and finalize the tutorial.
This tutorial is intended for an audience relatively new to the IoT area, in general, with emphasis on edge computing, AI hardware, and energy-efficient cyber-physical systems. The tutorial assumes a minimal background in machine learning and system optimization techniques. The presentation will introduce the relevant background IoT material, give an overview of the current state-of-the-art in machine learning techniques relevant to edge computing, and discuss in detail run-time resource optimization and dynamic power management techniques for training and inference on edge devices. The material discussed in this tutorial is highly relevant to students and researchers from industry and academia interested in the future of IoT and edge computing.
Short Biography of the Presenters:
Radu Marculescu is a professor and the Laura Jennings Turner Chair in Engineering in the ECE Department at University of Texas at Austin. He received his Ph.D. from Univ. of Southern California in 1998 and has been on the ECE faculty at Carnegie Mellon University between 2000-2019. He has co-authored many papers in a range of topics covering machine learning and systems design, low-power design and optimization, manycore systems-on-chip, networks-on-chip, embedded and cyber-physical systems, edge computing. His work on networks-on-chip design and optimization is widely recognized, most recently with the 2019 IEEE Computer Society 2019 Edward J. McCluskey Technical Achievement Award. His current research projects include machine learning and optimization for manycore systems design, AI approaches for HW/SW co-design, adversarial activity in social networks, and distributed learning approaches for edge devices.
Umit Y. Ogras received his Ph.D. degree in Electrical and Computer Engineering from Carnegie Mellon University, Pittsburgh, PA, in 2007. From 2008 to 2013, he worked as a Research Scientist at the Strategic CAD Laboratories, Intel Corporation. He is currently an Associate Professor at the School of Electrical, Computer and Energy Engineering at Arizona State University. Dr. Ogras has received 2018 DARPA Young Faculty Award, 2017 NSF CAREER Award, 2012 Intel Strategic CAD Labs Research Award, and best paper awards at 2019 CASES, 2017 CODES-ISSS, 2012 IEEE Donald O. Pederson Transactions on CAD and 2011 IEEE VLSI Transactions. His research interests include energy-efficient embedded systems, wearable internet-of-things, flexible hybrid electronics (FHE), multicore architectures and mobile platforms.