Photonic chips curb AI training’s energy appetite
Optical AI chips have struggled to implement a crucial algorithm used to train neural networks-backpropagation. In a recent paper published in the journal Science by a Stanford University team, the first implementation of this training method on a photonic chips was described.
Source:
https://spectrum.ieee.org/backpropagation-optical-ai?share_id=7627273&socialux=facebook&utm_campaign=RebelMouse&utm_content=IEEE+Spectrum&utm_medium=social