mlsanityData to 7 Oct 2026

A compact butterfly-style silicon photonic-electronic neural chip for hardware-efficient deep learning

arXiv 2111.06705 ↗First posted 11 Nov 2021cs.ET cs.LG physics.app-phcited by 13

Not on the map. Why?

Abstract (arXiv)

The optical neural network (ONN) is a promising hardware platform for next-generation neurocomputing due to its high parallelism, low latency, and low energy consumption. Previous ONN architectures are mainly designed for general matrix multiplication (GEMM), leading to unnecessarily large area cost and high control complexity. Here, we move beyond classical GEMM-based ONNs and propose an optical subspace neural network (OSNN) architecture, which trades the universality of weight representation for lower optical component usage, area cost, and energy consumption. We devise a butterfly-style photonic-electronic neural chip to implement our OSNN with up to 7x fewer trainable optical components compared to GEMM-based ONNs. Additionally, a hardware-aware training framework is provided to minimize the required device programming precision, lessen the chip area, and boost the noise robustness. We experimentally demonstrate the utility of our neural chip in practical image recognition tasks, showing that a measured accuracy of 94.16% can be achieved in hand-written digit recognition tasks with 3-bit weight programming precision.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 22 of its 75 references that are in the map. First posted on arXiv 11 Nov 2021.

Builds on

22 of 75 references.

References · 22

  1. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
  2. 2014Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluationcited by 304
  3. 2016EIE: Efficient Inference Engine on Compressed Deep Neural Networkcited by 291
  4. 2014cuDNN: Efficient Primitives for Deep Learningcited by 240
  5. 2014Speeding-up Convolutional Neural Networks Using Fine-tuned CP-Decompositioncited by 199
  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. 2020Photonics for artificial intelligence and neuromorphic computingcited by 111
  9. 2015Convolutional neural networks with low-rank regularizationcited by 102
  10. 2016Neuromorphic Silicon Photonic Networkscited by 66
  11. 2020Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unitcited by 64
  12. 2017CirCNN: Accelerating and Compressing Deep Neural Networks Using Block-CirculantWeight Matricescited by 53
10 more
  1. 2018Training of photonic neural networks through in situ backpropagationcited by 45
  2. 2020Efficient On-Chip Learning for Optical Neural Networks Through Power-Aware Sparse Zeroth-Order Optimizationcited by 24
  3. 2020Massively Parallel Amplitude-Only Fourier Neural Networkcited by 21
  4. 2019Reprogrammable Electro-Optic Nonlinear Activation Functions for Optical Neural Networkscited by 19
  5. 2020Photonic tensor cores for machine learningcited by 15
  6. 2018Building Efficient Deep Neural Networks with Unitary Group Convolutionscited by 9
  7. 2021Silicon photonic-electronic neural network for fibre nonlinearity compensationcited by 8
  8. 2019Efficient training and design of photonic neural network through neuroevolutioncited by 7
  9. 2020Modeling Silicon-Photonic Neural Networks under Uncertaintiescited by 7
  10. 2018Efficient Recurrent Neural Networks using Structured Matrices in FPGAscited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 1; 2023: 3; 2024: 5; 2025: 3; 2026: 1.

13 papers.

Sort:

  1. 2022NeurOLight: A Physics-Agnostic Neural Operator Enabling Parametric Photonic Device Simulationcited by 17
  2. 2024SimPhony: A Device-Circuit-Architecture Cross-Layer Modeling and Simulation Framework for Heterogeneous Electronic-Photonic AI Systemcited by 6
  3. 2023M3ICRO: Machine Learning-Enabled Compact Photonic Tensor Core based on PRogrammable Multi-Operand Multimode Interferencecited by 5
  4. 2024TeMPO: Efficient Time-Multiplexed Dynamic Photonic Tensor Core for Edge AI with Compact Slow-Light Electro-Optic Modulatorcited by 5
  5. 2024SCATTER: Algorithm-Circuit Co-Sparse Photonic Accelerator with Thermal-Tolerant, Power-Efficient In-situ Light Redistributioncited by 5
  6. 2024PACE: Pacing Operator Learning to Accurate Optical Field Simulation for Complicated Photonic Devicescited by 5
  7. 2023Integrated multi-operand optical neurons for scalable and hardware-efficient deep learningcited by 4
  8. 2023Neural Network Methods for Radiation Detectors and Imagingcited by 2
  9. 2025Hardware-Efficient Photonic Tensor Core: Accelerating Deep Neural Networks with Structured Compressioncited by 2
  10. 2024DOCTOR: Dynamic On-Chip Temporal Variation Remediation Toward Self-Corrected Photonic Tensor Acceleratorscited by 1
  11. 2025Efficient training for large-scale optical neural network using an evolutionary strategy and attention pruningcited by 0
  12. 2025Toward Large-Scale Photonics-Empowered AI Systems: From Physical Design Automation to System-Algorithm Co-Explorationcited by 0
1 more
  1. 2026Harnessing Photonics for Machine Intelligencecited by 0