Paper published in AIP Applied Physics Letters and featured on EurekAlert!

The paper "A recipe for creating ideal hybrid memristive-CMOS neuromorphic processing systems" by Elisabetta Chicca (Neuromorphic Behaving Systems group) and Giacomo Indiveri (Institute of Neuroinformatics, University of Zurich and ETH Zurich) has been published in the AIP Applied Physics Letters.

In our paper we present the fundamental difference of biological computation in the nervous systems and conventional artificial intelligence. We highlight how neuromorphic engineering focusses on principles of biological neural systems and how these features can be implemented in electronic circuits. With the recent developments in fabricaiton of memristive devices we emphasize on their use in neuromorphic application and circuits. These devices could pave the way towards ultra-low power brain-inspired processing systems and extreme-edge computing relevant for Internet-of-Things applications. 

The paper continues to suggest a recipe for the use of memristive devices and how they could implement the computing principles apparent in mammalian brains. Lastly use-cases of neuromorphic systems with learning abilities which can complement conventional processing systems are highlighted.
 

The paper has also been commented on by EurekAlert!. They quote "The electronic neural processing systems that we build are not intended to compete with the powerful and accurate artificial intelligence systems that run on power-hungry large computer clusters for natural language processing or high-resolution image recognition and classification" and assess the work as "[...] significant, because it can lead to a better understanding of how to implement sophisticated signal processing using extremely low-power and compact devices".