Keynote: AI Automation to Democratize AI for a better Society – an Inspiration from Electronic Design Automation

Jinjun Xiong (University at Buffalo, US)

Abstract

AI has become an increasingly powerful technological force that promises to impact all aspects of our society, from transportation to healthcare and education to sustainability.  But the diverse layers of software abstractions, hardware heterogeneity, and data privacy concerns have made the development of optimized AI solutions extremely challenging and costly. This results in the business world’s expensive investment only in a handful of selective and “profitable” AI solutions, leaving many critical societal needs, such as equitable education and sustainability, much less addressed than deserved. To truly democratize the power of AI for the benefit of the society and humanity, I argue that AI automation holds the key to drastically simplify the development of AI solutions and greatly improve AI productivity at a lower cost, just the same as the electronic design automation (EDA) does for improving the chip design productivity.

This talk attempts to draw a parallel analogy between the disciplined EDA design flow and the proposed AI automation flow. I will use some of my related research efforts in the past few years to illustrate both their similarities and differences, and more importantly, the research gaps. I will contextualize these efforts in an ultimate call for research actions and collaborations so that we can joint achieve this ambitious research goal of transforming the current computing paradigm with AI systems innovation to a truly democratized AI for a better society.

Curriculum Vitae

Dr. Jinjun Xiong is currently Empire Innovation Professor with the Department of Computer Science and Engineering at University at Buffalo (UB). Prior to that, he was a Senior Researcher and Program Director for AI and Hybrid Clouds Systems at the IBM Thomas J. Watson Research Center. He co-founded and co-directed the IBM-Illinois Center for Cognitive Computing Systems Research. His research interests are on across-stack AI systems research, which include AI applications, algorithms, tooling and computer architectures.  Many of his research results have been adopted in IBM’s products and tools. He published more than 150 peer-reviewed papers in top AI conferences and systems conferences. His publication won seven Best Paper Awards and eight Nominations for Best Paper Awards. He also won top awards from various international competitions, including the recent champion for the IEEE GraphChallenge on accelerating sparse neural networks, and champions for the DAC'19 Systems Design Contest on designing an object detection neural network for edge FPGA and GPU devices.