Brain-inspired computing could be 100,000 times faster with superconducting circuits

Brain-inspired computing could be scaled up with superconducting hardware

Scientists have been looking to the brain for inspiration in designing computer systems. Researchers have gone further and created computer hardware that has a brain-like design. The \”neuromorphic\” chips have shown promise but their speed and complexity are limited by the conventional digital electronics. The signals between the individual components of the chip become backlogged, like traffic on a gridlocked road. This slows down computation.

A team from the National Institute of Standards and Technology has now demonstrated a solution for these communication challenges, which may one day allow artificial neural systems operate 100,000 times faster that the human brain.

The brain is made up of 86 billion neurons. Each neuron can have thousands (known as synapses), of connections with its neighbor. The neurons use short electrical pulses, called spikes, to communicate with one another. This creates rich and time-varying patterns of activity that are the basis for cognition. In neuromorphic chip, electronic components are artificial neurons that route spiking signals in a brainlike network.