Optical Soliton Neural Networks

The chapter describes the realization of photonic integrated circuits based on photorefractive solitonic waveguides. In particular, it has been shown that X-junctions formed by soliton waveguides can learn information switching their state. X junctions perform both supervised and unsupervised learning. doing so, complex networks interconnected behave like a biological neural network, where is stored as preferred trajectories within network. this way, possible to create “episodic” psycho-memories, able memorize bit-by-bit, subsequently use recognize unknown data. Using optical systems, also more advanced dense networks, capable recognizing keywords packets (procedural psycho-memory) possibly comparing them with data (semantic psycho-memory). chapter, we shall describe how Solitonic Neural Networks work, showing close parallel between systems.