The 2nd Workshop on Advances in IoT Architecture and Systems
Keynote speakers: Dr. Schuyler Eldridge and Dr. Pradip Bose
Title: System Architectural Support for AI at the Edge
Abstract: The emergence of the Internet-of-Things (IoT) revolution has prompted computer architects to re-think the notion of how a new generation computing system should be defined or viewed. Intelligent edge devices can work in isolation, or in cooperative "swarms" - usually (but not always) with the support of the back-end cloud. In this talk, we will examine the space of cloud-backed edge cognition, with a special focus on particular domains like autonomous vehicles, cyber-security and dynamic resource management. We will examine research challenges associated with increasing the energy efficiency at the edge, while preserving the cognitive resilience of the overall solution engine.
Dr. Schuyler Eldridge is a Research Scientist within the "Efficient and Resilient Systems" department at IBM T. J. Watson Research Center. He holds a Ph.D from Boston University. His research interests are in architectural support for machine learning, open-source hardware design and efficient resilience.
Dr. Pradip Bose is a Distinguished Research Staff Member and Manager of the "Efficient and Resilient Systems" department at IBM T. J. Watson Research Center. He holds a Ph.D from University of Illinois at Urbana-Champaign. He has 35 years of experience at IBM and is a Fellow of the IEEE.
Internet-of-Things (IoT) is poised for a disruptive growth in near future where a wide variety of resources will be connected to the Internet. An IoT system will be the biggest connected system the human race has ever built. The system has to be open to support any future devices which may wish to connect to the system, which will be huge in size, supporting trillions of devices, millions of human users and transporting data of several magnitudes more than that of today’s Internet.
The broader vision of IoT encompasses all types of domains. Domain or a use-case has profound effect on the architectural decisions. Applications which will be hosted on such a platform will have diverse requirements. The architecture has to support various data and computation requirements, because many of these application are time critical. There are already examples of extreme types of applications, applications which are periodic in nature with diverse periodicity (eg. Reminders for pills, bus/train movements, periodic health checkup), applications which are highly aperiodic (investment on a stock, evacuation for emergency) again with diverse aperiodic distributions. Moreover, data volume requirements for different applications are also diverse. It may be difficult, if not impossible, to design an architecture which is generic enough across domains.
The purpose of the workshop is to solidify this research area and community. The long term aim is to bring together academic researchers and industry practitioners in this field to discuss the current state of research, common problems, discover new opportunities for collaboration, exchange ideas, and envision new areas of research, applications, and approaches.
Paper Submission Deadline
April 1, 2018
Notification of Acceptance
April 12, 2018
May 7, 2018
June 3, 2018 (Morning Half)