Nanowire Synapses are 30,000x faster than Nature’s.

Nanowire Synapses are 30,000 times faster than nature’s

Artificial-intelligence systems are increasingly limited by the hardware used to implement them. A new superconducting circuit mimics brain cell connections, and uses only 0.3 percent the energy as its human counterparts. It also operates 30,000 times faster.

Artificial neural networks are composed of components called neurons that receive data and work together to solve problems, like recognizing faces. The neural network adjusts its synapses, or the links between neurons, repeatedly to determine if the patterns of behavior it produces are more effective at solving a problem. The network learns over time which patterns produce the best results. The network then uses these patterns by default, simulating the learning process in the brain.