mlsanityData to 7 Oct 2026

Quantum-limited stochastic optical neural networks operating at a few quanta per activation

arXiv 2307.15712 ↗First posted 28 Jul 2023physics.optics cs.ET cs.LGcited by 8

Not on the map. Why?

Abstract (arXiv)

Energy efficiency in computation is ultimately limited by noise, with quantum limits setting the fundamental noise floor. Analog physical neural networks hold promise for improved energy efficiency compared to digital electronic neural networks. However, they are typically operated in a relatively high-power regime so that the signal-to-noise ratio (SNR) is large, and the noise can be treated as a perturbation. We study optical neural networks where all layers except the last are operated in the limit that each neuron can be activated by just a single photon, and as a result the noise on neuron activations is no longer merely perturbative. We show that by using a physics-based probabilistic model of the neuron activations in training, it is possible to perform accurate machine-learning inference in spite of the extremely high shot noise (SNR ~ 1). We experimentally demonstrated MNIST handwritten-digit classification with a test accuracy of 98% using an optical neural network with a hidden layer operating in the single-photon regime; the optical energy used to perform the classification corresponds to just 0.038 photons per multiply-accumulate (MAC) operation. Our physics-aware stochastic training approach might also prove useful with non-optical ultra-low-power hardware.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 25 of its 84 references that are in the map. First posted on arXiv 28 Jul 2023.

Builds on

25 of 84 references.

References · 25

  1. 2013Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computationcited by 2,031
  2. 2016DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradientscited by 557
  3. 2019Understanding Straight-Through Estimator in Training Activation Quantized Neural Netscited by 225
  4. 2023Challenges and Opportunities in Quantum Machine Learningcited by 202
  5. 2020Binary Neural Networks: A Surveycited by 184
  6. 2018All-Optical Machine Learning Using Diffractive Deep Neural Networkscited by 175
  7. 2021Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systemscited by 125
  8. 2016An Analysis of Deep Neural Network Models for Practical Applicationscited by 114
  9. 2016Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computingcited by 68
  10. 2020Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unitcited by 64
  11. 2019Wave Physics as an Analog Recurrent Neural Networkcited by 42
  12. 2020Surrogate gradients for analog neuromorphic computingcited by 38
13 more
  1. 2015Generalizing Pooling Functions in Convolutional Neural Networks: Mixed, Gated, and Treecited by 36
  2. 2021An optical neural network using less than 1 photon per multiplicationcited by 30
  3. 2017Vowel recognition with four coupled spin-torque nano-oscillatorscited by 22
  4. 2020Massively Parallel Amplitude-Only Fourier Neural Networkcited by 21
  5. 2020Noisy Machines: Understanding Noisy Neural Networks and Enhancing Robustness to Analog Hardware Errors Using Distillationcited by 17
  6. 2022Single-Shot Optical Neural Networkcited by 12
  7. 2021Understanding and mitigating noise in trained deep neural networkscited by 9
  8. 2019Improving Noise Tolerance of Mixed-Signal Neural Networkscited by 8
  9. 2016ASP Vision: Optically Computing the First Layer of Convolutional Neural Networks using Angle Sensitive Pixelscited by 4
  10. 2022Noise mitigation strategies in physical feedforward neural networkscited by 4
  11. 2017Energy-Efficient Hybrid Stochastic-Binary Neural Networks for Near-Sensor Computingcited by 3
  12. 2022RF-Photonic Deep Learning Processor with Shannon-Limited Data Movementcited by 2
  13. 2020How to do Physics-based Learningcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 4; 2025: 2; 2026: 2.

8 papers.

  1. 2024Training of Physical Neural Networkscited by 47
  2. 2024Stochastic Reservoir Computerscited by 4
  3. 2025Photonics-Enhanced Graph Convolutional Networkscited by 2
  4. 2024Quantum-secure multiparty deep learningcited by 1
  5. 2026Physical Foundation Models: Fixed hardware implementations of large-scale neural networkscited by 1
  6. 2024Impact of white Gaussian internal noise on analog echo-state neural networkscited by 0
  7. 2025Low-rank surrogate modeling and stochastic zero-order optimization for training of neural networks with black-box layerscited by 0
  8. 2026Measurement-Adapted Eigentask Representations for Photon-Limited Optical Readoutcited by 0