Research has long sought to develop computers to work as energy efficient as our brains. A study, led by researchers at the University of Gothenburg, has succeeded for the first time in combining memory function and computational function in the same component. This discovery opens the way to more efficient technologies, everything from cell phones to self-driving cars.
In recent years, computers have been able to tackle advanced cognitive tasks, such as language and image recognition or display superior chess skills, thanks in large part to artificial intelligence (AI). At the same time, the human brain remains unparalleled in its ability to perform tasks effectively and efficiently.
“Finding new ways to perform computations that are similar to energy-efficient brain operations has been a major goal of research for decades. Cognitive tasks, such as image and voice recognition, require significant computer power and mobile applications, in particular, such as cell phones, require drones and satellites. Energy-efficient solutions,” says Johann Uckermann, Professor of Applied Lower Electronics at the University of Gothenburg.
Working with a research team at Tohoku University, Aukermann led a study that has now taken an important step forward in achieving this goal. In the study now published in the journal nature materialsResearchers have succeeded for the first time in connecting the two main tools of advanced computation: oscillator networks and memristors.
Aukermann describes oscillators as oscillating circuits that can perform calculations that are comparable to human neurons. Memristors are programmable resistors that can also perform arithmetic operations and have built-in memory. This makes it comparable to memory cells. Merging the two is a major advance by researchers.
“This is an important achievement because we are showing that it is possible to combine a memory function with a computational function in the same component. These components work more like energy-saving neural networks in the brain, allowing them to become important building blocks in the future, more like brain-like computers.” .
Enables energy-saving technologies
According to Johann Aukermann, this discovery will enable faster, easier-to-use and lower energy-intensive technologies to be provided in many areas. He feels it’s a huge advantage that the research team has managed to produce the ingredients in an extremely small space: hundreds of ingredients fit into an area equivalent to a single bacteria. This can be especially important in smaller applications such as mobile phones.
“More energy-efficient arithmetic could lead to new functions in mobile phones. Examples are digital assistants like Siri or Google. Today, all processing is done by servers because computations require a lot of power for a small phone size. If Instead, calculations can be done locally, on a physical phone, and can be done faster and easier without having to connect to servers.”
He points to self-driving cars and planes as other examples of areas where more energy-efficient calculations could drive developments.
“The more energy efficient it is to perform cognitive computations, the more applications it becomes. That’s why our study really has the potential to advance this area.”
Energy-saving tuning of sigmoid neurons
Johan Aukermann, Memristive Control of Mutual Spindle Nano-Hall Oscillator Synchronization for Neural Computing, nature materials (2021). DOI: 10.1038/s41563-021-01153-6. www.nature.com/articles/s41563-021-01153-6
Presented by the University of Gothenburg
the quote: New discovery opens the way for brain-like computers (2021, November 29) Retrieved on November 29, 2021 from https://techxplore.com/news/2021-11-discovery-brain-like.html
This document is subject to copyright. Notwithstanding any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.