Scientists create artificial synapses that can learn independently

Developments and advances in artificial intelligence (AI) have been largely due to technologies that mimic the functioning of the human brain. In the world of information technology, such AI systems are called neural networks. These contain algorithms that can be trained, among other things, to mimic the way the brain recognizes speech and images. However, running an artificial neural network consumes a lot of time and energy.

Now, researchers at the Thales Center for Scientific Research (CNRS), the University of Bordeaux in Paris-Sud, and Evry have developed an artificial synapse called memristor directly on a chip. In this way they are closer to intelligent systems that will require less time and energy to learn, and that can learn independently. In the human brain, synapses function as the connections between neurons. In the human brain, synapses function as the connections between neurons. Connections are reinforced and learning is improved as more of these synapses are stimulated. It consists of a thin ferroelectric layer (which may be spontaneously polarized) that is enclosed between two electrodes. The use of voltage pulses, their resistance can be adjusted, like biological neurons. The synaptic connection will be strong when the resistance is low, and vice versa. The memristor’s ability for learning is based on this adjustable resistance. Artificial intelligence systems have been developed considerably in the last two years. Neural networks built with learning algorithms are now capable of performing tasks that synthetic systems previously could not do.

For example, intelligent systems can now compose music, play games and beat human players. Some may even identify suicidal behavior, or differentiate between what is legal and what is not. All this is thanks to the ability of AI to learn, the only limitation is the amount of time and effort needed to consume the data that serve as a springboard. With the memristor, this learning process can be greatly improved. It will continue to work on the memristor, especially in exploring ways to optimize its function. To begin with, researchers have successfully built a physical model to help predict how it works.

Add Comment