Superconducting optoelectronic circuits for neuromorphic computing

Abstract

Neural networks have proven effective for solving many difficult computational problems. Implementing complex neural networks in software is very computationally expensive. To explore the limits of information processing, it will be necessary to implement new hardware platforms with large numbers of neurons, each with a large number of connections to other neurons. Here we propose a hybrid semiconductor-superconductor hardware platform for the implementation of neural networks and large-scale neuromorphic computing. The platform combines semiconducting few-photon light-emitting diodes with superconducting-nanowire single-photon detectors to behave as spiking neurons. These processing units are connected via a network of optical waveguides, and variable weights of connection can be implemented using several approaches. The use of light as a signaling mechanism overcomes fanout and parasitic constraints on electrical signals while simultaneously introducing physical degrees of freedom which can be employed for computation. The use of supercurrents achieves the low power density necessary to scale to systems with enormous entropy. The proposed processing units can operate at speeds of at least $20$ MHz with fully asynchronous activity, light-speed-limited latency, and power densities on the order of 1 mW/cm${}^2$ for neurons with 700 connections operating at full speed at 2 K. The processing units achieve an energy efficiency of $\approx 20$ aJ per synapse event. By leveraging multilayer photonics with deposited waveguides and superconductors with feature sizes $>$ 100 nm, this approach could scale to systems with massive interconnectivity and complexity for advanced computing as well as explorations of information processing capacity in systems with an enormous number of information-bearing microstates.

34 pages, 22 figures

References in corpus (7)

• Deep Learning in Neural Networks: An Overview
• On-Chip Detection of Entangled Photons by Scalable Integration of Single-Photon Detectors
• Exponential expressivity in deep neural networks through transient chaos
• A superconducting-nanowire 3-terminal electronic device
• A Near-Infrared 64-pixel Superconducting Nanowire Single Photon Detector Array with Integrated Multiplexed Readout
• All-silicon sub-Gb/s telecom detector with low dark current and high quantum efficiency on chip
• Detection Mechanism in SNSPD: Numerical Results of a Conceptually Simple, Yet Powerful Detection Model

Cited by in corpus (46)

• Photonics for artificial intelligence and neuromorphic computing
• Neuromorphic Silicon Photonic Networks
• Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unit
• Training of photonic neural networks through in situ backpropagation
• The physics of optical computing
• Wave Physics as an Analog Recurrent Neural Network
• Reprogrammable Electro-Optic Nonlinear Activation Functions for Optical Neural Networks
• A silicon photonic modulator neuron
• Challenges in materials and devices for Resistive-Switching-based Neuromorphic Computing
• Prospects and applications of photonic neural networks
• Photonics for Neuromorphic Computing: Fundamentals, Devices, and Opportunities
• All-silicon light-emitting diodes waveguide-integrated with superconducting single-photon detectors
• Adiabatic Superconducting Artificial Neural Network: Basic Cells
• Circuit designs for superconducting optoelectronic loop neurons
• Quantifying power use in silicon photonic neural networks
• Optimization of photoluminescence from W centers in silicon-on-insulator
• A versatile, inexpensive integrated photonics platform
• Cryogenic Neuromorphic Hardware
• Principles of Neuromorphic Photonics
• Optoelectronic Intelligence
• Self-learning Machines based on Hamiltonian Echo Backpropagation
• Controlled integration of selected detectors and emitters in photonic integrated circuits
• Low-temperature emergent neuromorphic networks with correlated oxide devices
• High-fidelity two-qubit gates of hybrid superconducting-semiconducting singlet-triplet qubits
• Implementing an insect brain computational circuit using III-V nanowire components in a single shared waveguide optical network
• Microring resonator-coupled photoluminescence from silicon W~centers
• Shedding light on topological superconductors
• Optical Neural Network Based on Synthetic Nonlinear Photonic Lattices
• Integrated-photonic characterization of single-photon detectors for use in neuromorphic synapses
• Training Coupled Phase Oscillators as a Neuromorphic Platform using Equilibrium Propagation
• An active dendritic tree can mitigate fan-in limitations in superconducting neurons
• Design and resonator-assisted characterization of high performance lithium niobate waveguide crossings
• Choose your tools carefully: A Comparative Evaluation of Deterministic vs. Stochastic and Binary vs. Analog Neuron models for Implementing Emerging Computing Paradigms
• Artificial Nanophotonic Neuron with Internal Memory for Biologically Inspired and Reservoir Network Computing
• Optical Neural Ordinary Differential Equations
• Interfacing spiking VCSEL-neurons with silicon photonics weight banks towards integrated neuromorphic photonic systems
• Superconducting Optoelectronic Neurons I: General Principles
• Roadmap on Atto-Joule per Bit Modulators
• Superconducting Optoelectronic Neurons II: Receiver Circuits
• Superconducting Optoelectronic Neurons V: Networks and Scaling
• Holographic photonic neuron
• An ultrahigh-impedance superconducting thermal switch for interfacing superconductors to semiconductors and optoelectronics
• Digital-analog concept for superconducting perceptron-like neural networks
• Hierarchies among genuine multipartite entangling capabilities of quantum gates
• Nanoscale High Transition Temperature Superconducting Quantum Interference Device Transimpedance Amplifier
• The largest cognitive systems will be optoelectronic