Researchers at the University of California, Davis have developed a multi-wavelength, Spiking, Nanophotonic, Neural Reservoir Computing (SNNRC) system with high-dimensional (HD) computing capability.
To continue to increase the processing speeds of next-generation computers, researchers are investigating artificial neural networks (ANN). These neural networks attempt to replicate some aspects of biological neural networks in order to mimic the structure and processing capabilities of neurons. Hardware-based ANNs can be electronic or photonic. However, ANNs constructed with electronics consume considerable energy and perform calculations relatively slowly. On the other hand, photonics - which rely on the behavior of light and light-particle interactions, allow for faster computational speeds and lower overall energy consumption, but have other limitations.
Researchers at the University of California, Davis have addressed the various limitations mentioned above by developing a SNNRC that delivers faster computational processing speeds with a reduced hardware footprint and significantly lower energy consumption. It targets new artificial intelligence engines that provide cluster-scale capabilities at a chip scale – using silicon photonics. This computing framework mimics a biological neural network by spiking, which means that is only fires its artificial neurons when a membrane potential reaches a critical value. Since this technique uses nanophotonics, it also requires significantly less hardware than current systems.
Neural Networks, Nanophotonics, Spiking, Reservoir Computing, SNNRC, next-generation computing capabilities, high-dimensional computing