Apparatus and methods for optical neural networks

Abstract

The present invention provides an optical neural network constructed based on photonic integrated circuits to perform neuromorphic computing. In optical neural networks, matrix multiplication is implemented using one or more optical interferometric units that can apply arbitrary weighted matrix multiplications to an array of input optical signals. Nonlinear excitation is achieved through optical nonlinear units, which can be based on nonlinear optical effects such as saturable absorption. These calculations are implemented optically, resulting in high computational speed with low power consumption in optical neural networks.

Description

[0001] Related Application Cross Reference [0002] This application claims priority to U.S. Application No. 62,344,621, entitled "METHODS AND DESIGN OF OPTICALNEURAL NETWORK," filed June 2, 2016, which is hereby incorporated by reference in its entirety . [0003] governmental support [0004] This invention was made with Government support under Contract No. W911NF-13-D-0001 awarded by the Army Research Office. The government has certain rights in this invention. Background technique [0005] For a wide range of tasks, such as perception, communication, learning, and decision making, existing computers based on von Neumann architectures are generally more power-hungry and less efficient than their biological counterparts, the central nervous system. With the ever-increasing volume of data associated with big data processing, it would be beneficial to develop computers that can learn, combine, and analyze large amounts of information quickly and efficiently. For exampl...

AI Technical Summary

Problems solved by technology

However, the computational speed and power efficiency achieved with these hardware architectures are still limited by the electrical clock rate and ohmic losses

Images