New Hardware for Neuromorphic Computing Developed by International Research Team

International research team develops new hardware for neuromorphic computing
Nanodevice architecture and working principle. Credit: Nature Communications (2023). DOI: 10.1038/s41467-023-43891-y

Later on, advanced machinery should not only adhere to algorithms quickly and exactly, but also function cleverly—in other words, in a manner that resembles the human brain. Dortmund, Loughborough, Kiev and Nottingham scientists have devised an idea that takes inspiration from vision and that could potentially make future artificial intelligence a lot more compact and efficient.

They established an on-chip phonon-magnon reservoir for neuromorphic computing which has recently been showcased as Editor’s Highlight by Nature Communications.

Human sensory organs alter information such as light or scent into a signal that the brain processes through countless neurons connected by even more synapses. The faculty of the brain to train, specifically transform synapses, combined with the neurons’ vast number, enables humans to process very complex external signals and rapidly produce a response to them.

Researchers are attempting to mirror the principle of signal transmission and training with complex, neuromorphic computer systems—systems that bear resemblance to the neurobiological structures of the human nervous system. However, modern technologies are still infinitely far from achieving comparable information density and efficiency.

One of the methods designed to enhance neuromorphic systems is the reservoir computing framework. Here, the input signals are translated into a multidimensional space known as a reservoir. The reservoir is not trained and only speeds up recognition by a simplified artificial neural network.

This results in a significant reduction of computational resources and training time. A typical example of natural reservoir computing is human vision: In the eye, the visual information is pre-processed by hundreds of millions of the retina’s photoreceptors and converted into electrical signals that are transmitted by the optic nerve to the brain. This process significantly reduces the amount of data processed in the brain by the visual cortex.

Modern computer systems can impersonate reservoir functions when handling digitized signals. However, the fundamental breakthrough will be achieved when reservoir computing can be conducted directly with analog signals by a natural physical system, as in human vision.

The international team with researchers from Dortmund, Loughborough, Kyiv, and Nottingham have developed a new concept that brings such breakthroughs much closer.

The concept suggests a reservoir based on acoustic waves (phonons) and spin waves (magnons) mixed in a chip of 25x100x1 cubic microns. The chip consists of a multimode acoustic waveguide through which multiple different acoustic waves can be transmitted and which is covered by a patterned 0.1-micron-thickness magnetic film.

The information delivered by the train of ultrashort laser pulses is pre-processed before the recognition by conversion to the propagating phonon-magnon wavepacket. Short wavelengths of the propagating waves result in high information density, which enables the confident recognition of visual shapes drawn by a laser on a remarkably small area of less than one photopixel.

Researcher Alexander Balanov from Loughborough University, one of the concept’s authors, states, “The potential of the physical system proposed as a reservoir was immediately obvious for us because of its amazing combination of variability and multidimensionality.”

His colleague Professor Sergey Savel’ev emphasizes the similarity of the demonstrated working principle with the functionality of the human brain: “The functionality of the developed reservoir is based on the interference and mixture of the optically generated waves, which is very similar to the recently suggested mechanism of the information processing in the biological cortex.”

Dr. Alexey Scherbakov, who led the project at TU Dortmund University, says, “Our concept is very promising because it is based on conversion of the income signal to high-frequency acoustic waves, just like in modern wireless communication devices.”

“Our acoustic frequency range above 10 GHz is a bit higher than available right now, but it is targeted by the next wireless communication standards. Thus, who knows, probably in a couple of years, your mobile phone will help you make very human decisions.”

More information:
Dmytro D. Yaremkevich et al, On-chip phonon-magnon reservoir for neuromorphic computing, Nature Communications (2023). DOI: 10.1038/s41467-023-43891-y

Provided by
TU Dortmund University

International research team develops new hardware for neuromorphic computing (2024, February 7)
retrieved 7 February 2024

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.

Leave a Reply

Your email address will not be published. Required fields are marked *