Energy-efficient neural network inference with microcavity exciton-polaritons
arXiv:2108.12648 · doi:10.1103/PhysRevApplied.16.024045
Abstract
We propose all-optical neural networks characterized by very high energy efficiency and performance density of inference. We argue that the use of microcavity exciton-polaritons allows to take advantage of the properties of both photons and electrons in a seamless manner. This results in strong optical nonlinearity without the use of optoelectronic conversion. We propose a design of a realistic neural network and estimate energy cost to be at the level of attojoules per bit, also when including the optoelectronic conversion at the input and output of the network, several orders of magnitude below state-of-the-art hardware implementations. We propose two kinds of nonlinear binarized nodes based either on optical phase shifts and interferometry or on polariton spin rotations.
References in corpus (45)
- Deep Learning with Coherent Nanophotonic Circuits
- All-Optical Machine Learning Using Diffractive Deep Neural Networks
- Quantum fluids of light
- Parallel convolution processing using an integrated photonic tensor core
- Photonics for artificial intelligence and neuromorphic computing
- 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
- Observation of Superfluidity of Polaritons in Semiconductor Microcavities
- An Analysis of Deep Neural Network Models for Practical Applications
- Lasing in topological edge states of a 1D lattice
- Binarized Neural Networks
- Neuromorphic Silicon Photonic Networks
- Attojoule Optoelectronics for Low-Energy Information Processing and Communications: a Tutorial Review
- Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unit
- Exciton-polariton topological insulator
- All-optical polariton transistor
- Exciton-polaritons in van der Waals heterostructures embedded in tunable microcavities
- All Optical Neural Network with Nonlinear Activation Functions
- A one femtojoule athermal silicon modulator
- Reinforcement Learning in a large scale photonic Recurrent Neural Network
- Wave Physics as an Analog Recurrent Neural Network
- Realizing the Hamiltonian in polariton simulators
- A silicon photonic modulator neuron
- Optical circuits based on Polariton Neurons in Semiconductor Microcavities
- Polariton Condensate Transistor Switch
- Bose-Einstein Condensation of Long-Lifetime Polaritons in Thermal Equilibrium
- Quantum correlations of confined exciton-polaritons
- Two-Dimensional hybrid perovskites sustaining strong polariton interactions at room temperature
- Quantum-correlated photons from semiconductor cavity polaritons
- Human action recognition with a large-scale brain-inspired photonic computer
- Propagation and amplification dynamics of 1D polariton condensates
- Probing the Dynamics of Spontaneous Quantum Vortices in Polariton Superfluids
- Polaritonic neuromorphic computing outperforms linear classifiers
- Topological order and equilibrium in a condensate of exciton-polaritons
- Direct measurement of polariton-polariton interaction strength in the Thomas-Fermi regime of exciton-polariton condensation
- Control and ultrafast dynamics of a two-fluid polariton switch
- Room temperature Organic Exciton-Polariton Condensate in a Lattice
- Dark Solitons in Waveguide Polariton Fluids Shed Light on Interaction Constants
- Fully reconfigurable coherent optical vector-matrix multiplication
- Interactions and scattering of quantum vortices in a polariton fluid
- Large-scale spatiotemporal photonic reservoir computer for image classification
- Neuromorphic computing in Ginzburg-Landau polariton lattice systems
- Perceptrons with Hebbian learning based on wave ensembles in plastic potentials
- Optically and electrically controlled polariton spin transistor
- Coherently driven microcavity-polaritons and the question of superfluidity
- All-optical cascadable universal logic gate with sub-picosecond operation
Cited by in corpus (6)
- Polariton lattices as binarized neuromorphic networks
- Neural networks based on ultrafast time-delayed effects in exciton-polaritons
- The Roles of Kerr nonlinearity in a Bosonic Quantum Neural Network
- Modeling Nonlinear Optics with the Transfer Matrix Method
- Quantization of Polaritons Confined in Dielectric Structures
- Polaritonic Machine Learning for Graph-based Data Analysis