Learning at the Speed of Light: A New Type of OpticalNeural Network

摘要

Most, if not all, optical hardware-based neural networks are slow during the neural learning phase. This limitation has been not only a speed bottleneck, but it has contributed to the lack of wide-spread use of optical neural systems. We present a novel solution - Optical Fixed-Weight Learning Neural Networks. Standard neural networks learn new function mappings by the changing of their synaptic weights. However, the Fixed-Weight Neural Networks learn new mappings by dynamically changing recurrent neural signals. The (fixed) synaptic weights of the FWL-NN implement a learning "algorithm" which adjusts the recurrent signals toward their proper values.