Augmented reality (AR) is a set of technologies that will fundamentally change the way we interact with our environment. It represents a merging of the physical and the digital worlds into a rich, context aware and accessible user interface delivered through a socially acceptable form factor such as eyeglasses. One of the biggest challenges in realizing a comprehensive AR experience is managing power consumption to ensure both adequate battery life and a physically comfortable thermal envelope. This presentation reviews AR glasses silicon challenges and advanced concepts in minimizing power in data transfer across components, leveraging highly efficient accelerators while maintaining programmability, and the potential of emerging technologies for low power computing.
Edith Beigné joined Facebook Inc. in Menlo Park in November 2018 to lead the AR/VR Silicon Research team. Before that, she was with CEA-LETI, Grenoble, France, from 1998 to 2018 where she was the Research Director of Integrated Circuits and System Division. Since 2009, she has been a senior scientist in the digital and mixed-signal design lab where she focused on low power and adaptive circuit techniques, exploiting asynchronous design and advanced technology nodes like FDSOI 28nm and 14nm for many different applications from high-performance MPSoC to ultra-low power IoT applications. Her main research interests today are low power digital and mixed-signal circuits and design with emerging technologies. She is part of ISSCC TPC since 2014 and part of VLSI’s symposium since 2015. Distinguished Lecturer for the SSCS in 2016/2017, Women-in-Circuits Committee chair and JSSC Associate Editor since 2018. She visited Stanford University in 2018.
Artificial intelligence is being used in a variety of edge-computing devices such as biomedical sensors, wearables and autonomous systems. Processing these sensor-level machine learning tasks come at the cost of high computational complexity and memory storage which is overwhelming for these light weight and battery constrained devices. Equally important is the need for designing smarter AI systems that can reason over in the face of a highly variable and unpredictable world. This talk overviews some research solutions that enable performing data analytics from variety of multimodal sensors in real time while consuming low power. I will also talk about adding reasoning in these systems to improve acting and learning performance. Combining these solutions will bring exciting opportunities for future micro AI processors
Tinoosh Mohsenin is an Associate Professor in the Department of Computer Science and Electrical Engineering at UMBC and Director of the Energy Efficient High Performance Computing Lab. Prof. Mohsenin’s research focus is on designing low power processors for high computational machine learning and knowledge extraction techniques used in wearables, Interment of Things and Autonomous systems. She has over 100 peer-reviewed journal and conference publications and is the recipient of NSF CAREER award in 2017, the best paper award in the ACM Great Lakes VLSI conference 2016, and the best paper honorable award in the IEEE Circuits and Systems Symposium 2017 for developing processors in biomedical and deep learning. She has previously served as Associate Editor in IEEE Transactions on Circuits and Systems-I (TCAS-I) and IEEE Transactions on Biomedical Circuits and Systems (TBioCAS). She is the General Chair of 30th ACM GLSVLSI conference in 2020 and was the Program Chair and Local Arrangement Chair for the 29th ACM GLSVLSI and 50th IEEE ISCAS conferences, in 2019 and 2017 respectively.
Edge Artificial Intelligence (AI) is the new mega-trend, as privacy concerns and networks bandwidth/latency bottlenecks prevent cloud offloading of AI functions in many application domains, from autonomous driving to advanced prosthetics. Hence we need to push AI toward sensors and actuators. I will give an overview of recent efforts in developing systems of-on-chips based on open source hardware and capable of performing significant analytics and AI functions "at the extreme edge", i.e. within the limited power budget of traditional microcontrollers that can be co-located and integrated with the sensors/actuators themselves. These open, extreme edge AI platforms create a exciting playground for research and innovation.
Luca Benini is the chair of digital Circuits and Systems at D-ITET ETHZ. He received a Ph.D. degree in electrical engineering from Stanford University in 1997. He has served as Chief Architect for the Platform2012/STHORM project in STmicroelectronics, Grenoble in the period 2009-2013. He is also a professor at University of Bologna and has held visiting and consulting researcher positions at EPFL, IMEC, Hewlett-Packard Laboratories, Stanford University.
Dr. Benini's research interests are in the design of energy-efficient digital systems with special emphasis on ultra-low-power System-on-Chip and green HPC systems. He is also active in the area of smart sensors and sensor networks for consumer, biomedical and Internet-of-Things applications. In these areas he has coordinated tens of funded projects, including an ERC Advanced Grant on Multi-scale thermal management of Computing Systems.
He has been general chair of the Design Automation and Test in Europe Conference, of the Network on Chip Symposium and of the International Symposium on Low Power Electronics and Design. He is Associate Editor of the IEEE Transactions on Computer-Aided Design of Circuits and Systems and the ACM Transactions on Embedded Computing Systems.
He has published more than 700 papers in peer-reviewed international journals and conferences, four books and several book chapters. He is a Fellow of the IEEE and the ACM and a member of the Academia Europaea. He is the recipient of the 2016 IEEE CAS Mac Van Valkenburg award.