[12p-N404-1~14] CS.2 3.3 情報フォトニクス・画像工学、4.4 Information Photonicsのコードシェアセッション

▲ [12p-N404-9] Progress in neuro-inspired photonic computing

〇Satoshi Sunada¹ (1.Kanazawa Univ.)

キーワード：ニューラルネットワーク、光コンピューティング、リザバーコンピューティング

Photonic neural networks have significant potential for high-speed neural processing with low latency and ultra-low energy consumption. In this talk, I briefly review recent progresses and challenges in on-chip photonic neural networks. Then, I will introduce our photonic neural processing chip, which allows for highly scalable and high-speed neuro-inspired computing. The principle of the computing is high-dimensional neural encoding based on a photonic neural field. The proposed photonic chip is potentially capable of the processing of more than one peta multiply-accumulate (MAC) operation per second, which is a dominant factor in neural processing, for a single input wavelength on a small footprint.