Nonlinear unitary circuits for photonic neural networks

Abstract

Photonics has unlocked the potential for energy-efficient acceleration of deep learning. Most approaches toward photonic deep learning have diligently reproduced traditional deep learning architectures using photonic platforms, separately implementing linear-optical matrix calculations and nonlinear activations via electro-optical conversion, optical nonlinearities, and signal-encoded materials. Here we propose a concept of nonlinear unitary photonic circuits to achieve the integration of linear and nonlinear expressivity essential for deep neural networks. We devise a building block for two-dimensional nonlinear unitary operations, featuring norm-preserving mappings with nonconservative inner products, which enables the construction of high-dimensional nonlinear unitary circuits. Using deep nonlinear unitary circuits, we demonstrate exponential growth in trajectory length and near-complete coverage of the output space, both of which are essential for deep learning. Along with neuroevolutionary learning examples for the regression of a nonconvex function, our results pave the way to photonic neural networks with highly expressive inference and stable training.

33 pages, 7 figures

References in corpus (19)

• Deep Learning with Coherent Nanophotonic Circuits
• All-Optical Machine Learning Using Diffractive Deep Neural Networks
• Neuromorphic Silicon Photonic Networks
• Synthetic Dimension in Photonics
• All Optical Neural Network with Nonlinear Activation Functions
• Efficient, Compact and Low Loss Thermo-Optic Phase Shifter in Silicon
• The physics of optical computing
• Experimentally realized in situ backpropagation for deep learning in nanophotonic neural networks
• The Expressive Power of Parameterized Quantum Circuits
• A single photonic cavity with two independent physical synthetic dimensions
• 422 Million Q Planar Integrated All-Waveguide Resonator with a 3.4 Billion Absorption Limited Q and Sub-MHz Linewidth
• Universal spectra of random Lindblad operators
• Arbitrary linear transformations for photons in the frequency synthetic dimension
• Universal Nonlinear Disordered Wave Packet Subdiffusion: 12 Decades
• Temporal coupled-mode theory in nonlinear resonant photonics: From basic principles to contemporary systems with 2D materials, dispersion, loss, and gain
• Heavy tails and pruning in programmable photonic circuits
• Dynamical phase diagram of parity-time symmetry with competing saturable channels
• Programmable photonic time circuits for highly scalable universal unitaries
• Partially Unitary Learning