IEEE Spectrum August 29, 2020
To perform neural-network calculations faster and more efficiently based on optical computing researchers in the US (industry) have built Mars device which has at its heart a chip that includes an analog optical processor, designed specifically to perform the mathematical operations that are fundamental to neural networks. The key optical processor is shrunk to nanometers and integrated them together on one chip for the purpose of speeding up neural-network calculations. The interferometer splits incoming light into two beams, which then take two different paths. The resulting two beams are then recombined. If the two paths are identical, the output looks just like the input. If the beams fall out of phase the two beams interfere and the output is entirely nulled. They expect to go into production with its Mars device in late in 2021…read more.
Credit: iStockphoto