Researching | Principles and Applications of Nonlinearity in Optical Neural Networks (Invited)

Acta Optica Sinica, Volume. 45, Issue 14, 1420011(2025)

Principles and Applications of Nonlinearity in Optical Neural Networks (Invited)

Yufei Wang^1,2, Yumeng Chen¹, Yongzheng Yang¹, Kun Liao^1、*, Xiaoyong Hu^1,2、**, and Qihuang Gong^1,2

Author Affiliations

¹State Key Laboratory of Artificial Microstructure and Mesoscopic Physics, School of Physics, Peking University, Beijing 100871, China

²Yangtze Delta Institute of Optoelectronics, Peking University, Nantong 226010, Jiangsu , China

show less

Abstract Get PDF(in Chinese)

Figures&Tables (10)

References (65)

Paper Information

References(65)

[1] LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 521, 436-444(2015).

[2] Yuan H Y, Du Z C, Qi H X et al. Microcomb-driven photonic chip for solving partial differential equations[J]. Advanced Photonics, 7, 016007(2025).

[3] Fu T Z, Sun R, Huang Y Y et al. Review of on‑chip integrated optical neural networks (invited)[J]. Chinese Journal of Lasers, 51, 0119002(2024).

[4] Chen B, Zhang Z Y, Dai T G et al. Photonic neural networks and its applications[J]. Laser & Optoelectronics Progress, 60, 0600001(2023).

[5] Waldrop M M. The chips are down for Moore’s law[J]. Nature, 530, 144-147(2016).

[6] Leshno M, Lin V Y, Pinkus A et al. Multilayer feedforward networks with a nonpolynomial activation function can approximate any function[J]. Neural Networks, 6, 861-867(1993).

[7] Yu W Z, Zheng S, Zhao Z Y et al. Reconfigurable low-threshold all-optical nonlinear activation functions based on an add-drop silicon microring resonator[J]. IEEE Photonics Journal, 14, 5559807(2022).

[8] Jha A, Huang C R, Prucnal P R. Reconfigurable all-optical nonlinear activation functions for neuromorphic photonics[J]. Optics Letters, 45, 4819-4822(2020).

[9] Zhang M, Wang C, Cheng R et al. Monolithic ultra-high-Q lithium niobate microring resonator[J]. Optica, 4, 1536-1537(2017).

[10] Yuan S, Wu Y K, Dang Z Z et al. Strongly enhanced second harmonic generation in a thin film lithium niobate heterostructure cavity[J]. Physical Review Letters, 127, 153901(2021).

[11] Li G H Y, Sekine R, Nehra R et al. All-optical ultrafast ReLU function for energy-efficient nanophotonic deep learning[J]. Nanophotonics, 12, 847-855(2022).

[12] Li G H Y, Leefmans C R, Williams J et al. Deep learning with photonic neural cellular automata[J]. Light: Science & Applications, 13, 283(2024).

[13] Inagaki T, Inaba K, Leleu T et al. Collective and synchronous dynamics of photonic spiking neurons[J]. Nature Communications, 12, 2325(2021).

[14] Wang H, Hu J Q, Morandi A et al. Large-scale photonic computing with nonlinear disordered media[J]. Nature Computational Science, 4, 429-439(2024).

[15] Guo X X, Barrett T D, Wang Z M et al. Backpropagation through nonlinear units for the all-optical training of neural networks[J]. Photonics Research, 9, B71-B80(2021).

[16] Yu J R, Yang X X, Gao G Y et al. Bioinspired mechano-photonic artificial synapse based on graphene/MoS2 heterostructure[J]. Science Advances, 7, eabd9117(2021).

[17] Yang Z, Tan W M, Zhang T J et al. MXene-based broadband ultrafast nonlinear activator for optical computing[J]. Advanced Optical Materials, 10, 2200714(2022).

[18] Chen C D, Yang Z, Wang T et al. Ultra-broadband all-optical nonlinear activation function enabled by MoTe2/optical waveguide integrated devices[J]. Nature Communications, 15, 9047(2024).

[19] Liao K, Li C T, Dai T X et al. Matrix eigenvalue solver based on reconfigurable photonic neural network[J]. Nanophotonics, 11, 4089-4099(2022).

[20] Liao K, Lian Y X, Yu M T et al. Hetero-integrated perovskite/Si3N4 on-chip photonic system[J]. Nature Photonics, 19, 358-368(2025).

[21] Basani J R, Heuck M, Englund D R et al. All-photonic artificial neural network processor via nonlinear optics[J]. Physical Review Applied, 22, 014009(2024).

[22] Zuo Y, Li B H, Zhao Y J et al. All-optical neural network with nonlinear activation functions[J]. Optica, 6, 1132-1137(2019).

[23] Opala A, Panico R, Ardizzone V et al. Training a neural network with exciton-polariton optical nonlinearity[J]. Physical Review Applied, 18, 024028(2022).

[24] Cheng Z G, Ríos C, Pernice W H P et al. On-chip photonic synapse[J]. Science Advances, 3, e1700160(2017).

[25] Chakraborty I, Saha G, Roy K. Photonic in-memory computing primitive for spiking neural networks using phase-change materials[J]. Physical Review Applied, 11, 014063(2019).

[26] Feldmann J, Youngblood N, Wright C D et al. All-optical spiking neurosynaptic networks with self-learning capabilities[J]. Nature, 569, 208-214(2019).

[27] Brückerhoff-Plückelmann F, Bente I, Becker M et al. Event-driven adaptive optical neural network[J]. Science Advances, 9, eadi9127(2023).

[28] Teo T Y, Ma X X, Pastor E et al. Programmable chalcogenide-based all-optical deep neural networks[J]. Nanophotonics, 11, 4073-4088(2022).

[29] Pour Fard M M, Williamson I A D, Edwards M et al. Experimental realization of arbitrary activation functions for optical neural networks[J]. Optics Express, 28, 12138-12148(2020).

[30] Williamson I A D, Hughes T W, Minkov M et al. Reprogrammable electro-optic nonlinear activation functions for optical neural networks[J]. IEEE Journal of Selected Topics in Quantum Electronics, 26, 7700412(2020).

[31] Tait A N, Ferreira de Lima T, Nahmias M A et al. Silicon photonic modulator neuron[J]. Physical Review Applied, 11, 064043(2019).

[32] Ashtiani F, Geers A J, Aflatouni F. An on-chip photonic deep neural network for image classification[J]. Nature, 606, 501-506(2022).

[33] Bandyopadhyay S, Sludds A, Krastanov S et al. Single-chip photonic deep neural network with forward-only training[J]. Nature Photonics, 18, 1335-1343(2024).

[34] Nozaki K, Matsuo S, Fujii T et al. Femtofarad optoelectronic integration demonstrating energy-saving signal conversion and nonlinear functions[J]. Nature Photonics, 13, 454-459(2019).

[35] Pappas C, Kovaios S, Moralis-Pegios M et al. Programmable tanh-, ELU-, sigmoid-, and sin-based nonlinear activation functions for neuromorphic photonics[J]. IEEE Journal of Selected Topics in Quantum Electronics(2023).

[36] Xu Z F, Tang B S, Zhang X Y et al. Reconfigurable nonlinear photonic activation function for photonic neural network based on non-volatile opto-resistive RAM switch[J]. Light: Science & Applications, 11, 288(2022).

[37] Zhong C Y, Liao K, Dai T X et al. Graphene/silicon heterojunction for reconfigurable phase-relevant activation function in coherent optical neural networks[J]. Nature Communications, 14, 6939(2023).

[38] Liu C, Ma Q, Luo Z J et al. A programmable diffractive deep neural network based on a digital-coding metasurface array[J]. Nature Electronics, 5, 113-122(2022).

[39] Chen Y T, Nazhamaiti M, Xu H et al. All-analog photoelectronic chip for high-speed vision tasks[J]. Nature, 623, 48-57(2023).

[40] Li X K, Ma J X, Li X Y et al. High-efficiency reinforcement learning with hybrid architecture photonic integrated circuit[J]. Nature Communications, 15, 1044(2024).

[41] Bernstein L, Sludds A, Panuski C et al. Single-shot optical neural network[J]. Science Advances, 9, eadg7904(2023).

[42] Zhu H H, Zou J, Zhang H et al. Space-efficient optical computing with an integrated chip diffractive neural network[J]. Nature Communications, 13, 1044(2022).

[43] Xue Z W, Zhou T K, Xu Z H et al. Fully forward mode training for optical neural networks[J]. Nature, 632, 280-286(2024).

[44] Antonik P, Marsal N, Brunner D et al. Human action recognition with a large-scale brain-inspired photonic computer[J]. Nature Machine Intelligence, 1, 530-537(2019).

[45] Zhan Z Y, Wang H, Liu Q et al. Photonic diffractive generators through sampling noises from scattering media[J]. Nature Communications, 15, 10643(2024).

[46] Wang D L, Nie Y K, Hu G L et al. Ultrafast silicon photonic reservoir computing engine delivering over 200 TOPS[J]. Nature Communications, 15, 10841(2024).

[47] Chen Z J, Sludds A, Davis R et al. Deep learning with coherent VCSEL neural networks[J]. Nature Photonics, 17, 723-730(2023).

[48] Shi Y, Ren J Y, Chen G Y et al. Nonlinear germanium-silicon photodiode for activation and monitoring in photonic neuromorphic networks[J]. Nature Communications, 13, 6048(2022).

[49] Vandoorne K, Mechet P, Van Vaerenbergh T et al. Experimental demonstration of reservoir computing on a silicon photonics chip[J]. Nature Communications, 5, 3541(2014).

[50] Sunada S, Uchida A. Photonic neural field on a silicon chip: large-scale, high-speed neuro-inspired computing and sensing[J]. Optica, 8, 1388-1396(2021).

[51] Brunner D, Soriano M C, Mirasso C R et al. Parallel photonic information processing at gigabyte per second data rates using transient states[J]. Nature Communications, 4, 1364(2013).

[52] Duport F, Schneider B, Smerieri A et al. All-optical reservoir computing[J]. Optics Express, 20, 22783-22795(2012).

[53] Larger L, Soriano M C, Brunner D et al. Photonic information processing beyond Turing: an optoelectronic implementation of reservoir computing[J]. Optics Express, 20, 3241-3249(2012).

[54] Vinckier Q, Duport F, Smerieri A et al. High-performance photonic reservoir computer based on a coherently driven passive cavity[J]. Optica, 2, 438-446(2015).

[55] Nie Y K, Yang B, Wang D L et al. Integrated laser graded neuron enabling high-speed reservoir computing without a feedback loop[J]. Optica, 11, 1690-1699(2024).

[56] Shen Y W, Li R Q, Liu G T et al. Deep photonic reservoir computing recurrent network[J]. Optica, 10, 1745-1751(2023).

[57] Lupo A, Picco E, Zajnulina M et al. Deep photonic reservoir computer based on frequency multiplexing with fully analog connection between layers[J]. Optica, 10, 1478-1485(2023).

[58] Maass W. Networks of spiking neurons: the third generation of neural network models[J]. Neural Networks, 10, 1659-1671(1997).

[59] Yao M, Richter O, Zhao G S et al. Spike-based dynamic computing with asynchronous sensing-computing neuromorphic chip[J]. Nature Communications, 15, 4464(2024).

[60] Xiang S Y, Shi Y C, Guo X X et al. Hardware-algorithm collaborative computing with photonic spiking neuron chip based on an integrated Fabry-Perot laser with a saturable absorber[J]. Optica, 10, 162-171(2023).

[61] Zhang Y H, Robertson J, Xiang S Y et al. All-optical neuromorphic binary convolution with a spiking VCSEL neuron for image gradient magnitudes[J]. Photonics Research, 9, B201-B209(2021).

[62] Zhang Y H, Xiang S Y, Cao X Y et al. Experimental demonstration of pyramidal neuron-like dynamics dominated by dendritic action potentials based on a VCSEL for all-optical XOR classification task[J]. Photonics Research, 9, 1055-1061(2021).

[63] Xiang S Y, Shi Y C, Zhang Y H et al. Photonic integrated neuro-synaptic core for convolutional spiking neural network[J]. Opto-Electronic Advances, 6, 230140(2023).

[64] Zhang Y H, Xiang S Y, Guo X X et al. Spiking information processing in a single photonic spiking neuron chip with double integrated electronic dendrites[J]. Photonics Research, 11, 2033-2041(2023).

[65] Gao S, Xiang S Y, Song Z W et al. Hardware implementation of ultra-fast obstacle avoidance based on a single photonic spiking neuron[J]. Laser & Photonics Reviews, 17, 2300424(2023).

Tools

Get Citation

Copy Citation Text

Yufei Wang, Yumeng Chen, Yongzheng Yang, Kun Liao, Xiaoyong Hu, Qihuang Gong. Principles and Applications of Nonlinearity in Optical Neural Networks (Invited)[J]. Acta Optica Sinica, 2025, 45(14): 1420011

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Optics in Computing

Received: Apr. 15, 2025

Accepted: Jun. 23, 2025

Published Online: Jul. 18, 2025

The Author Email: Kun Liao (kunliao@pku.edu.cn), Xiaoyong Hu (xiaoyonghu@pku.edu.cn)

DOI:10.3788/AOS250924

CSTR:32393.14.AOS250924

Topics

laser devices and laser physics

Lasers and Laser Optics

laser manufacturing

Instrumentation, Measurement and Metrology