Major nanowire breakthrough could help us copy brain memory networks

13 Aug 2018

Illustration of a human brain. Image: Anders Sandberg/Flickr (CC BY-NC 2.0)

Biological systems currently far exceed the abilities of artificial ones, but a new nanowire breakthrough helps to close the gap.

Despite us all having one and it being fundamental to who we are, the human brain remains the envy of computer scientists across the globe.

While not set in stone, estimates put the average memory storage capacity of the brain at as much as one petabyte, or the equivalent of 1,000TB.

This is why computer scientists and engineers are so eager to learn from it in the development of human-made artificial intelligence (AI).

Now, a team of researchers from the Science Foundation Ireland-funded AMBER centre and Duke University in the US has revealed a major discovery that helps us move one step closer to mimicking biology.

Finding the most efficient route

In a paper published to Nature Communications, the researchers described the discovery of the emergence of ‘winner takes all’ connectivity pathways in random networks.

This is where memory is distributed across the network but encoded in specific connectivity pathways, similar to those found in biological systems.

In the same way that humans will instinctively search for the most efficient route somewhere – such as cutting across a grass field rather than taking a longer pathway – finding new connectivity pathways helps AI make decisions better and faster.

Through experiment and simulation, the team was able to figure out the properties of nanowire networks that give rise to singular or multiple connectivity pathways.

With a single wire being 1,000 times thinner than a human hair, nanowires can be made from a number of different materials designed to prevent them from clumping together.

Why it is important

By changing the nanowire material, or the coating on the nanowire, the team found that networks can develop different types of connectivity pathways. Crucially, this identified the conditions required for the emergence of a single lowest-energy, most-efficient pathway.

“Even more surprising was that for silver nanowires, which prefer to self-select a single lowest-energy pathway across the random network. Once the pathway is established, it forms a series of discrete memory levels,” said Prof John Boland of the research team.

“These results point to the possibility of developing and independently addressing memory levels in complex systems, which we expect to have important implications for computers that operate in a more brain-like fashion.”

The next goal of the research is to understand how to engineer this single or multi-path behaviour, and to develop logic systems based on these nanowire network materials for cognitive signal processing, decision-making systems and, ultimately, neuromorphic computing applications.

Illustration of a human brain. Image: Anders Sandberg/Flickr (CC BY-NC 2.0)

Colm Gorey was a senior journalist with Silicon Republic

editorial@siliconrepublic.com