Fiber-based Photonic-FPGA Architecture and In-Fiber Computing

Hardware implementation of artificial neural networks facilitates real-time parallel processing of massive data sets. Optical neural networks offer low-volume 3D connectivity together with large bandwidth and minimal heat production in contrast to electronic implementation. In this presentation we will present a conceptual design for in-fiber optical neural network, i.e. a fiber-based realization of a photonic-FPGA. Neurons and synapses are realized in two ways: first as individual silica cores in a multi-core fiber and then within a multi-mode fiber. In the first realization optical signals are transferred transversely between cores by means of optical coupling. Pump driven amplification in erbium-doped cores mimics synaptic interactions. Simulations and experimental validation show classification and learning capabilities. Therefore, devices similar to our proposed multi-core fiber could potentially serve as building blocks for future large-scale small-volume optical artificial neural networks. In the second type of realization we propose the design of an optical artificial neural network-based imaging system that has the ability to self-study image signals from an incoherent light source in different colors. Our design consists of a multi-mode fiber realizing a stochastic neural network. We show that the signals, transmitted through the multi-mode fiber, can be used for image identification purposes and can also be reconstructed using artificial neural networks with a low number of nodes.