Researchers demonstrate the ‘nanomagnetic’ computing of AI


A team of researchers from Imperial College London has demonstrated how it is possible to achieve artificial intelligence (AI) with tiny nanomagnets that interact like neurons in the brain.

This new “nanomagnetic” calculation method could reduce the energy costs associated with AI. This is crucial given that AI energy costs are doubling globally every 3.5 months.

The research was published in the journal Nanotechnology.

Performing Type IA Treatment with Nanomagnets

In the research paper, the team demonstrated the first evidence that nanomagnet arrays can achieve type IA processing. They also showed how these nanomagnets can be used for “time series prediction” tasks, which include things like predicting insulin levels for diabetic patients.

Classical neural networks are based on the functioning of the human brain, the neurons communicating with each other for the processing of information. However, it has been difficult to use magnets directly in this process, with researchers not knowing how to insert data or extract information.

To simulate magnetic interactions, experts typically rely on software running on traditional silicon-based computers, which can simulate the brain. The current advancement has seen the team use magnets themselves to process and store data, eliminating the need for software simulation.

Nanomagnets are not all the same. Instead, they come in various “states” depending on their direction. By applying a magnetic field to an array of nanomagnets, the state of the magnets can change depending on the properties of the input field and the states of the surrounding magnets.

Design the new technique

The team was able to take this and devise a technique to count the number of magnets in each state after the field passed.

Dr. Jack Gartside is co-first author of the study.

“We’ve been trying to figure out how to enter data, ask a question, and get an answer from magnetic computing for a long time,” Dr Gartside said. “Now that we’ve proven it can be done, it paves the way for eliminating the computer software that does the power-intensive simulation.”

Killian Stenning is co-first author of the article.

“The way magnets interact gives us all the information we need; the laws of physics themselves become the computer,” Stenning said.

Dr. Will Branford is team leader.

“The long-term goal was to make hardware inspired by the software algorithms of Sherrington and Kirkpatrick,” Dr Branford said. “It was not possible to use the spins on the atoms in conventional magnets, but by increasing the spins in nano-patterned arrays we were able to achieve the necessary control and readout.”

Reduce energy waste

Much of the energy used for AI in conventional silicon-chip computers is wasted due to inefficient transport of electrons during processing and memory storage. On the other hand, nano-magnets do not require the physical transport of particles like electrons. They process and transfer information with a “magnon” wave, with each magnet affecting the state of others around it.

This process results in less wasted energy. The processing and storage of information is done together rather than separately, as is the case in traditional computers. With all these advances, nanomagnetic computing could be up to 100,000 times more efficient than conventional computing.


Comments are closed.