Journal Article10.1016/j.optcom.2023.130121
Microdisk modulator-assisted optical nonlinear activation functions for photonic neural networks
Bin Wang,Weizhen Yu,Jinming Duan,Shuwen Yang,Zhenyu Zhao,Shuang Zheng,Weifeng Zhang +6 more
3
TL;DR: Microdisk modulator-assisted optical nonlinear activation functions for photonic neural networks enable high-speed electro-optic and all-optical nonlinear activation functions in a single device.
read more
Abstract: On-chip implementation of optical nonlinear activation functions (NAFs) is essential for realizing large-scale photonic neural chips. To improve the reconfigurability and enrich the functions of photonic neural network (PNN), an NAF unit with a unique capability of performing multiple types of NAFs is highly preferred. In this work, we propose and experimentally demonstrate a microdisk modulator (MDM)-assisted NAF unit, in which multiple types of electro-optic NAFs with a high response speed and multiple types of all-optical NAFs with a low latency and reduced power consumption are performed in a single device. The fabricated MDM has an add-drop configuration, in which a lateral PN junction is incorporated to achieve high-speed resonance wavelength tuning. With the use of the high-speed nonlinear electro-optic effect, three different kinds of electro-optic NAFs including sigmoid function, radial basis function (RBF), and negative rectified linear unit (ReLU) function are realized by exploiting the free-carrier dispersion effect on silicon. Moreover, thanks to the strong optical confinement in the disk cavity, all-optical NAFs can be performed by exploiting the thermo-optic effect on silicon. In the experiment, four different kinds of all-optical NAFs including softplus function, RBF, clamped ReLU function, and leaky ReLU function are demonstrated. With the use of the realized clamped ReLU function, a convolutional neural network is simulated to perform a handwritten digit classification benchmark task, and an accuracy as high as 98 % is demonstrated. Thanks to its strong electro-optic and thermo-optic effects, the proposed MDM provides a unique capability of performing multiple types of electro-optic and all-optical NAFs, which is potential to be served as a flexible nonlinear unit in large-scale PNN chips.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
On-chip electro-optical spiking VO₂/Si device with an inhibitory leaky integrate-and-fire response
Juan Morcillo,Pablo Sanchis,Jorge O. Parra +2 more
- 26 Jul 2024
TL;DR: Researchers propose an electro-optical spiking device on a silicon photonics platform using VO₂/Si waveguides and microheaters, achieving a leaky integrate-and-fire response with inhibitory optical spiking, enabling scalable and energy-efficient photonic-based spiking neural networks.
On-chip electro-optical spiking VO₂/Si device with an inhibitory leaky integrate-and-fire response
Juan Morcillo,Pablo Sanchis,Jorge O. Parra +2 more
- 26 Jul 2024
TL;DR: Researchers propose an electro-optical spiking device for silicon photonics, leveraging VO₂'s temperature-driven insulator-metal transition to achieve a leaky integrate-and-fire response with inhibitory optical spiking, enabling scalable and energy-efficient photonic-based spiking neural networks.
Integrated Neuromorphic Photonic Computing for AI Acceleration: Emerging Devices, Network Architectures, and Future Paradigms
Gaofei Wang,Junyan Che,Chen Gao,Zhou Han,Jiabin Shen,Zengguang Cheng,Zhou Peng +6 more
Abstract: Abstract Deep learning stands as a cornerstone of modern artificial intelligence (AI), revolutionizing fields from computer vision to large language models (LLMs). However, as electronic hardware approaches fundamental physical limits—constrained by transistor scaling challenges, von Neuman architecture, and thermal dissipation—critical bottlenecks emerge in computational density and energy efficiency. To bridge the gap between algorithmic ambition and hardware limitations, photonic neuromorphic computing emerges as a transformative candidate, exploiting light's inherent parallelism, sub‐nanosecond latency, and near‐zero thermal losses to natively execute matrix operations—the computational backbone of neural networks. Photonic neural networks (PNNs) have achieved influential milestones in AI acceleration, demonstrating single‐chip integration of both inference and in situ training—a leap forward with profound implications for next‐generation computing. This review synthesizes a decade of progress in PNNs core components, critically analyzing advances in linear synaptic devices, nonlinear neuron devices, and network architectures, summarizing their respective strengths and persistent challenges. Furthermore, application‐specific requirements are systematically analyzed for PNN deployment across computational regimes: cloud‐scale and edge/client‐side AIs. Finally, actionable pathways are outlined for overcoming material‐ and system‐level barriers, emphasizing topology‐optimized active/passive devices and advanced packaging strategies. These multidisciplinary advances position PNNs as a paradigm‐shifting platform for post‐Moore AI hardware.
References
Deep learning
TL;DR: Deep learning is making major advances in solving problems that have resisted the best attempts of the artificial intelligence community for many years, and will have many more successes in the near future because it requires very little engineering by hand and can easily take advantage of increases in the amount of available computation and data.
67K
Neural computation of decisions in optimization problems
John J. Hopfield,David W. Tank +1 more
TL;DR: Results of computer simulations of a network designed to solve a difficult but well-defined optimization problem-the Traveling-Salesman Problem-are presented and used to illustrate the computational power of the networks.
6K
Artificial neural networks: fundamentals, computing, design, and application
I.A Basheer,M Hajmeer +1 more
TL;DR: A bird's eye review of the various types of ANNs and the related learning rules is presented, with special emphasis on backpropagation ANNs theory and design, and a generalized methodology for developing successful ANNs projects from conceptualization, to design, to implementation is described.
3.2K
Deep learning with coherent nanophotonic circuits
Yichen Shen,Nicholas C. Harris,Scott Skirlo,Dirk Englund,Marin Soljacic +4 more
- 01 Jul 2017
TL;DR: A new architecture for a fully optical neural network is demonstrated that enables a computational speed enhancement of at least two orders of magnitude and three order of magnitude in power efficiency over state-of-the-art electronics.
2.9K
Low-Dose CT With a Residual Encoder-Decoder Convolutional Neural Network
TL;DR: This work combines the autoencoder, deconvolution network, and shortcut connections into the residual encoder–decoder convolutional neural network (RED-CNN) for low-dose CT imaging and achieves a competitive performance relative to the-state-of-art methods in both simulated and clinical cases.