Slot 1

Neuromorphic Electronic Circuits for Compact Low-Power Real-Time Neural Processing Systems
Giacomo Indiveri, Institute of Neuroinformatics, Switzerland


Artificial neural networks and deep learning algorithms are currently achieving impressive state-of-the-art results in a wide range of applications that involve processing of digital data. However, despite the remarkable progress made in recent years, today’s artificial systems are still not able to compete with biological ones in tasks that involve real-time processing of sensory data acquired by interacting with the environment in complex and uncertain settings. One of the reasons 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. Conventional computers on the other hand operate with mainly serial and synchronous logic gates, with functions that are decoupled from their hardware implementation, and with discretized and virtual time.

In this tutorial I will describe examples of neuromorphic circuits that exploit the physics of Silicon to carry out neural computation in compact and low power VLSI devices. I will present neural network architectures with spike-based plasticity mechanisms that can be used to perform signal processing and pattern recognition tasks, and describe mixed-signal analog/digital neural processing modules that can be used to build large-scale multi-core neuromorphic processors. I will show examples of multi-chip multi-core neuromorphic computing platforms and demonstrate application examples that exploit their on-chip on-line learning properties.

The detailed list of topics that will be discussed in this tutorial is the following:

1. Principles and origins of neuromorphic engineering
2. Artificial neural networks and spiking neurons
3. Implementing neural dynamics with subthreshold analog circuits
4. Learning in VLSI spiking neural network architectures
5. Requirements and solutions for large-scale neuromorphic computing platforms
6. Application areas best suited for neuromorphic computing technologies


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 the California Institute of Technology (Caltech) and at the Institute of Neuroinformatics of the University of Zurich and ETH Zurich, where he attained the Habilitation in Neuromorphic Engineering in 2006. He is an ERC fellow and an IEEE Senior member. His research interests lie in the study of real and artificial neural processing systems, and in the hardware implementation of neuromorphic cognitive systems, using full custom analog and digital VLSI technology.

  Back to course info