Title: Towards a Universal Analog Computing Paradigm

Abstract: In the digital era, is there a role for analog ? In this talk I will present a perspective in the defense of analog computing and argue that vast frontiers of unexplored research opportunities exist in this domain. Analog computers operate by exploiting computational primitives inherent in the physics of devices to efficiently implement complex mathematical functions and dynamical systems. I will first present some fundamental analog principles that are commonly used for synthesizing different mathematical functions using circuit elements like transistors. Then, using representative examples from my research and others, I will show how some of these principles could be scaled up to implement processors for machine learning, emulation, simulation and error-correcting decoders. In the last part of my talk, I will present some new theoretical results showing that choosing a unique analog representation could lead to a computing paradigm with fundamental limits of energy-efficiency superior than the existing paradigm.

Bio: Shantanu Chakrabartty is a Professor of Computer Science and Engineering at Washington University in St. Louis. Prior to joining Washington Universit, he was a faculty  at Michigan State University, where he served as Professor and Director of the Adaptive Integrated Microsystems Laboratory. He received his B.Tech in Electrical Engineering from the Indian Institute of Technology in Delhi, India before receiving his MS and PhD from Johns Hopkins University in 2002 and 2004. Professor Chakrabartty is an alumnus of the US National Academy Frontiers of Engineering and has received numerous awards including the National Science Foundation CAREER Award (2010), MSU Teacher-Scholar Award (2011), and the MSU Innovation of the Year Award (2012).