Giacomo Indiveri

What should neuromorphic computing architectures be like?

Download the presentation

Neural networks and deep learning algorithms are currently achieving impressive state-of-the-art results on computing tasks that operate on stored data sets. However, artificial computing systems are still vastly outperformed by biological neural processing ones for tasks that involve processing of sensory data acquired in real-time in complex and uncertain settings, and closed-loop interactions with the environment. This difference is remarkable especially when size and energy consumption are factored in.

One of the reasons for this gap is that neural computation in biological systems is very different from the way today's computers operate: it is tightly linked to the properties of their computational embodiment, to the physics of their computing elements and to their temporal dynamics. I will argue that hybrid analog/digital microelectronic circuits that use the physics of Silicon to directly emulate the biophysics of the neural processes they model represent a promising technology for building intelligent and energy-efficient autonomous cognitive agents. I will demonstrate examples of brain-inspired architectures that integrate massively parallel arrays of such circuits to implement on-chip on-line spike-based learning and computation, and will describe the advantages and disadvantages of these types of computing architectures compared to conventional computing systems, in this domain.

About the panel speaker:

Giacomo Indiveri is a Professor at the Faculty of Science of the University of Zurich, Switzerland. He obtained an M.Sc. degree in electrical engineering and a Ph.D. degree in computer science from the University of Genoa, Italy. Indiveri was a post-doctoral research fellow in the Division of Biology at Caltech and at the Institute of Neuroinformatics of the University of Zurich and ETH Zurich. In 2006 he attained the "habilitation" in Neuromorphic Engineering at the ETH Zurich Department of Information Technology and Electrical Engineering, and in 2011 he won an ERC Starting Grant on "Neuromorphic processors: event-based VLSI models of cortical circuits for brain-inspired computation".  His research interests lie in the study of neural computation, with particular interest in spike-based learning and selective attention mechanisms, and in the hardware implementation of real-time sensory-motor systems using neuromorphic circuits and VLSI technology.