Scientists have created a laser-based artificial neuron that replicates the functions, dynamics, and data-processing capabilities of a biological graded neuron. Operating at a signal processing speed of 10 GBaud — about a billion times faster than its natural counterpart — this groundbreaking technology could drive major advances in artificial intelligence and advanced computing systems.
Biological neurons come in various types, including graded neurons, which process information through continuous changes in membrane potential for precise and nuanced signal transmission. In contrast, spiking neurons use all-or-none action potentials, sending signals in a binary, on-or-off manner.
“Our laser graded neuron overcomes the speed limitations of current photonic versions of spiking neurons and has the potential for even faster operation,” said research team leader Chaoran Huang from the Chinese University of Hong Kong. “By leveraging its neuron-like nonlinear dynamics and fast processing, we built a reservoir computing system that demonstrates exceptional performance in AI tasks such as pattern recognition and sequence prediction.”
To read more, click here.