Unusual magnets helped create an AI that works without electronics: how it is possible

Researcher uses nanomagnets to create artificial intelligence that would not consume a lot of energy, but at the same time would be very effective.

Enjamin Jungfleisch, associate professor at the University of Delaware in the USA, is inspired by the human brain's ability to process and store information. He wants AI to work without electronics — on nanomagnets, writes interestingengineering.com.

The researcher is confident that his work can help lead to the future of energy-efficient artificial intelligence (AI) systems.

Just as neurons are the basic unit of information processing in the human brain, magnons are the fundamental quantum exciters that make up the spin waves, or magnetic waves, in a magnetic system. Magnon connections are similar to synapses between neurons that send signals along specific routes. Jungfleisch works with nanomagnetic arrays and intends to use them as a neural network in the brain to process and transmit data.

A scientist adjusts microwave components to measure the dynamics of spin waves, which depends on the magnetic field. He makes lattices out of them so that they interact. The similarity between neurons and nanomagnets is striking because they can store information. “We put these nanomagnets in a network, and they sense each other,” — explained the researcher.

Unlike a traditional computing device, which uses separate blocks of memory and processor, nanomagnets become twice as efficient at performing both tasks. It can help create energy-efficient processes and then develop a chatbot or even create an image based on them. The operation of such a system does not depend on electrons, since magnetic excitations are used for data storage and processing. Interestingly, nanomagnets can store a history of their states and can also be taught. Currently, training cycles take two to three hours. However, in the future, the scientist expects the process to take a few minutes.

In a recent publication, the researcher described a three-dimensional nanomagnetic structure that can be easily fabricated and read using existing techniques.