Three dimensional waveguide-interconnects for scalable integration of photonic neural networks

Abstract

Photonic waveguides are prime candidates for integrated and parallel photonic interconnects. Such interconnects correspond to large-scale vector matrix products, which are at the heart of neural network computation. However, parallel interconnect circuits realized in two dimensions, for example by lithography, are strongly limited in size due to disadvantageous scaling. We use three dimensional (3D) printed photonic waveguides to overcome this limitation. 3D optical-couplers with fractal topology efficiently connect large numbers of input and output channels, and we show that the substrate's footprint area scales linearly. Going beyond simple couplers, we introduce functional circuits for discrete spatial filters identical to those used in deep convolutional neural networks.

arXiv comment7 pages, 6 figures

References in corpus (10)

• Double-slit photoelectron interference in strong-field ionization of the neon dimer
• Deep Learning with Coherent Nanophotonic Circuits
• Synthesis of clathrate cerium superhydride CeH9 below 100 GPa with atomic hydrogen sublattice
• Attojoule Optoelectronics for Low-Energy Information Processing and Communications: a Tutorial Review
• All-optical Reservoir Computing
• Training of photonic neural networks through in situ backpropagation
• Photonic Wire Bonds for Terabit/s Chip-to-Chip Interconnects
• Large-scale photonic Ising machine by spatial light modulation
• Hybrid 2D/3D photonic integration for non-planar circuit topologies
• In-Datacenter Performance Analysis of a Tensor Processing Unit

Cited by in corpus (20)

• Deep learning in nano-photonics: inverse design and beyond
• The physics of optical computing
• Prospects and applications of photonic neural networks
• Meta-optics for spatial optical analog computing
• Direct (3+1)D laser writing of graded-index optical elements
• Computational inverse design for ultra-compact single-piece metalenses free of chromatic and angular aberration
• Understanding and mitigating noise in trained deep neural networks
• Programmable metasurfaces for future photonic artificial intelligence
• Integrated-photonic characterization of single-photon detectors for use in neuromorphic synapses
• Noise mitigation strategies in physical feedforward neural networks
• Optical pumping of quantum dot micropillar lasers
• Additive 3D photonic integration that is CMOS compatible
• Transformation-optics modeling of 3D-printed freeform waveguides
• Beating the classical phase precision limit using a quantum neuromorphic platform
• Observation of light propagation through a three-dimensional cavity superlattice in a 3D photonic band gap
• Multipartite entanglement generation in coupled microcavity arrays
• Impact of internal noise on convolutional neural networks
• Diagnostics and characterization of photonic circuits by wide-field spatio-temporal imaging
• Noise impact on recurrent neural network with linear activation function
• Ultrafast neural sampling with spiking nanolasers