检索结果-内蒙古大学图书馆

Monolithically integrated asynchronous optical recurrent accelerator

ELIGHT 2025年第1期5卷 1-14页

作者： Wu, Bo Zhou, Haojun Cheng, Junwei Zhang, Wenkai Zhang, Shiji Huang, Chaoran Huang, Dongmei Zhou, Hailong Dong, Jianji Zhang, Xinliang Huazhong Univ Sci & Technol Sch Opt & Elect Informat Wuhan Natl Lab Optoelect Wuhan 430074 Peoples R China Chinese Univ Hong Kong Dept Elect Engn Shatin Hong Kong Peoples R China Hong Kong Polytech Univ Photon Res Inst Dept Elect & Elect Engn Hong Kong Peoples R China Xidian Univ Xian 710071 Peoples R China

computing with light is widely recognized as a promising paradigm for overcoming the energy and latency limitations of electronic computing. However, the energy consumption and latency in current optical computing hardware predominantly arise in the electrical domain rather than the optical domain, primarily due to frequent signal conversions between optical (analog) and electrical (digital) formats. Furthermore, as the operating frequency of optical computing surpasses the GHz range, the synchronization of parallel electrical signals and the management of optical delays become increasingly critical. These challenges exacerbate energy consumption and latency, particularly in recurrent optical operations. To address these limitations, we propose a novel asynchronous computing paradigm for on-chip optical recurrent accelerators based on wavelength encoding, effectively mitigating synchronization challenges. By leveraging the intrinsic causality of wavelength relay, our approach eliminates the need for rigorous temporal alignment. To demonstrate the flexibility and efficacy of this asynchronous paradigm, we present two advanced recurrent models-an optical hidden Markov model and an optical recurrent neural network-monolithically integrated for the first time. These models incorporate hundreds of linear and nonlinear computing units densely packed into a compact footprint of just 10 mm2. Experimental evaluations on various benchmark tasks underscore the superior energy efficiency and low latency of the proposed asynchronous optical accelerators. This innovation enables the efficient processing of large-scale parallel signals and positions optical processors as a pivotal technology for applications such as autonomous driving and intelligent robotics.

关键词： Asynchronous operation optical recurrent accelerator optical hidden Markov model optical recurrent neural network

来源：评论

学校读者我要写书评

暂无评论

Confluence of Photonics and Artificial Intelligence

引用

IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS 2025年第3期31卷

作者： Chen, Young-Kai Coherent Corp Santa Clara CA 95054 USA

Over the past decades, significant advances have been made in the fields of photonic technologies, artificial intelligence, and machine learning techniques. Recent AI progress in language models, perception and self-learning capability has had a significant impact on a wide range of communities. The new photonics enhances AI with faster data processing, reduced energy consumption, and improved interconnects, enabling more efficient and powerful AI systems. In this paper, we will highlight the synchroneity of these advances from performance enhancement of AI processing by photonics and advanced photonics with AI processing, leading to new photonic neural networks for future photonic AI processors.

关键词： optical switches Artificial intelligence High-speed optical techniques optical fiber communication optical distortion optical interconnections Photonics optical amplifiers optical polarization Training optical interconnects optical switching photonic inverse design photonic processing photonic neural network photonic computing artificial intelligence machine learning

来源：评论

学校读者我要写书评

暂无评论

optical network modelling-based data analytics for network monitoring and security analysis using hybrid computing models

引用

optical AND QUANTUM ELECTRONICS 2023年第14期55卷 1-16页

作者： Li, Fang Xie, Yalou Han, Yong Zhengzhou Univ Henan Prov Peoples Hosp Cent China Fuwai Hosp Dept InformatCtr Cent China Zhengzhou 450003 Henan Peoples R China

With the ever-increasing complexity and scale of optical networks, efficient network monitoring and robust security analysis have become paramount. In this study, we propose a novel approach that combines optical network modeling with data analytics, leveraging the power of hybrid computing models. Aim of this research is to propose novel technique in optical network modelling based on data analysis in network monitoring with security enhancement using hybrid computing model with deep learning techniques. Here the optical network monitoring and anomaly detection is carried out using fuzzy density clustering based markov K-means attention recurrent neural network (FDCM-KARNN). By continuously monitoring network performance and traffic patterns, our system can proactively detect and respond to network aberrations, minimizing downtime and mitigating potential risks. Then by hybrid computing based on edge and cloud network the vast amounts of network data and provide accurate predictions has been enabled. The experimental analysis is carried out for various monitored optical network dataset based on network security analysis and data analysis in terms of training accuracy, throughput, packet delivery ratio, data integrity, precision. The proposed technique attained training accuracy of 99%, precision of 98%, packet delivery ratio of 97%, data integrity of 85%, throughput of 96%.

关键词： optical network modelling Data analytics network monitoring Security analysis Hybrid computing models Machine learning neural networks

来源：评论

学校读者我要写书评

暂无评论

Quantum neural network Boosting Identifying Orbital Angular Momentum Modes

引用

ADVANCED QUANTUM TECHNOLOGIES 2025年第0期

作者： Merabet, Badreddine Li, Liulin Shah, Syed Afaq Ali Guo, Zhongyi Hefei Univ Technol Sch Comp & Informat Engn Hefei 230000 Peoples R China

Quantum machine learning (QML) offers a breakthrough by producing atypical patterns that classical systems cannot efficiently produce. This study explores quantum neural networks (QNN) for classifying orbital angular momentum (OAM) modes. A fully quantum model achieving 99.8% accuracy across 10 OAM classes and a hybrid ConvQuantum model reaching accuracy levels of 99.8% under atmospheric turbulence strength (AT) of Cn2=1x10-14m-2/3$C_n<^>2 = 1 \times {{10}<^>{-14}}{{m}<^>{-2/3}}$ is developed. The ConvQuantum model significantly reduces prediction time, averaging 0.00025 s per sample, which is essential for real-time free space optical (FSO) communication. To demonstrate quantum technology's potential in OAM-FSO communication, a 100 x 100 grayscale image is successfully transmitted, and compared the results with several known models in the field, including feedforward neural networks (FNN), convolutional neural network (CNN), and a vision transformer (ViT) model. The findings indicate that the QNN models outperform classical approaches in calculation time, showcasing the efficiency and enhanced processing power of quantum computing in OAM-FSO communication.

关键词： free-space optical (FSO) communication orbital angular momentum (OAM) quantum machine learning (QML) quantum neural networks (QNN)

来源：评论

学校读者我要写书评

暂无评论

Data-Driven Modeling of Mach-Zehnder Interferometer-Based optical Matrix Multipliers

引用

JOURNAL OF LIGHTWAVE TECHNOLOGY 2023年第16期41卷 5425-5436页

作者： Cem, Ali Yan, Siqi Ding, Yunhong Zibar, Darko Da Ros, Francesco Tech Univ Denmark Dept Elect & Photon Engn DK-2800 Lyngby Denmark Huazhong Univ Sci & Technol Sch Opt & Elect Informat Wuhan 430074 Peoples R China

Photonic integrated circuits are facilitating the development of optical neural networks, which have the potential to be both faster and more energy efficient than their electronic counterparts since optical signals are especially well-suited for implementing matrix multiplications. However, accurate programming of photonic chips for optical matrix multiplication remains a difficult challenge. Here, we describe both simple analytical models and data-driven models for offline training of optical matrix multipliers. We train and evaluate the models using experimental data obtained from a fabricated chip featuring a Mach-Zehnder interferometer mesh implementing 3-by-3 matrix multiplication. The neural network-based models outperform the simple physics-based models in terms of prediction error. Furthermore, the neural network models are also able to predict the spectral variations in the matrix weights for up to 100 frequency channels covering the C-band. The use of neural network models for programming the chip for optical matrix multiplication yields increased performance on multiple machine learning tasks.

关键词： Machine learning neuromorphic computing optical matrix multiplication

来源：评论

学校读者我要写书评

暂无评论

Fundamentals and recent developments of free-space optical neural networks

引用

JOURNAL OF APPLIED PHYSICS 2024年第3期136卷 030701-030701页

作者： McNeil, Alexander Montes Li, Yuxiao Zhang, Allen Moebius, Michael Liu, Yongmin Northeastern Univ Dept Elect & Comp Engn Boston MA 02115 USA Charles Stark Draper Lab Cambridge MA 02139 USA Charles Stark Draper Lab Cambridge MA 02139 USA Northeastern Univ Dept Mech & Ind Engn Boston MA 02115 USA

Machine learning with artificial neural networks has recently transformed many scientific fields by introducing new data analysis and information processing techniques. Despite these advancements, efficient implementation of machine learning on conventional computers remains challenging due to speed and power constraints. optical computing schemes have quickly emerged as the leading candidate for replacing their electronic counterparts as the backbone for artificial neural networks. Some early integrated photonic neural network (IPNN) techniques have already been fast-tracked to industrial technologies. This review article focuses on the next generation of optical neural networks (ONNs), which can perform machine learning algorithms directly in free space. We have aptly named this class of neural network model the free space optical neural network (FSONN). We systematically compare FSONNs, IPNNs, and the traditional machine learning models with regard to their fundamental principles, forward propagation model, and training process. We survey several broad classes of FSONNs and categorize them based on the technology used in their hidden layers. These technologies include 3D printed layers, dielectric and plasmonic metasurface layers, and spatial light modulators. Finally, we summarize the current state of FSONN research and provide a roadmap for its future development. (c) 2024 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license

关键词： optical data processing

来源：评论

学校读者我要写书评

暂无评论

Efficient on-chip training of large-scale optical neural network through block adjoint training algorithm

引用

OPTICS EXPRESS 2024年第26期32卷 46633-46648页

作者： Yang, Zhiwei Zhang, Tian Dai, Jian Xu, Kun Beijing Univ Posts & Telecommun State Key Lab Informat Photon & Opt Commun Beijing 100876 Peoples R China Beijing Univ Posts & Telecommun Sch Elect Engn Beijing 100876 Peoples R China

MZI-based block optical neural networks (BONNs), which utilize block matrix multiplication to achieve large-scale network models, have garnered significant attention but still lack efficient training algorithms. In this article, by calculating the original field and adjoint field for the block matrices in BONNs and directly updating the phase values of all phase shifters within the optical mesh, we propose an on-chip block adjoint training (BAT) algorithm for large-scale BONNs. To demonstrate the effectiveness of our proposed algorithm, the trained BONNs are applied in image classification tasks for MNIST and SVHN datasets. The calculated results demonstrate that the performance of the BAT algorithm (95.915% for the MNIST dataset and 82.64% for the SVHN dataset) is competitive with the traditional gradient algorithm based on artificial neural networks (96.238% and 84.182%), but the BONNs can infer 1.5 times and 1.3 times faster than artificial neural networks, respectively. By studying the influence of the block size and the inputted position of the padded zero signals, we demonstrate that the BAT algorithm based on the BONNs with 12 block sizes can achieve higher performance by adding the padded zero signals to the same side beside the normal inputted signals. Additionally, we demonstrate that substituting the complete weight matrices with unitary matrices to construct BONNs is an efficient way to reduce both the system area and the required trainable parameters. Finally, we demonstrate the relatively good robustness of the BAT algorithm and the imprecision alleviation method by using on-chip retraining. Notably, our proposed BAT algorithm shows excellent potential for more complex tasks and network models.

关键词： Genetic algorithms Machine vision neural networks optical neural systems Phase shift Stochastic gradient descent

来源：评论

学校读者我要写书评

暂无评论

OPIMA: optical Processing-in-Memory for Convolutional neural network Acceleration

引用

IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS 2024年第11期43卷 3888-3899页

作者： Sunny, Febin Shafiee, Amin Balasubramaniam, Abhishek Nikdast, Mahdi Pasricha, Sudeep Colorado State Univ Elect & Comp Engn Dept Ft Collins CO 80523 USA

Recent advances in machine learning (ML) have spotlighted the pressing need for computing architectures that bridge the gap between memory bandwidth and processing power. The advent of deep neural networks has pushed traditional Von Neumann architectures to their limits due to the high latency and energy consumption costs associated with data movement between the processor and memory for these workloads. One of the solutions to overcome this bottleneck is to perform computation within the main memory through processing-in-memory (PIM), thereby limiting data movement and the costs associated with it. However, dynamic random-access memory-based PIM struggles to achieve high throughput and energy efficiency due to internal data movement bottlenecks and the need for frequent refresh operations. In this work, we introduce OPIMA, a PIM-based ML accelerator, architected within an optical main memory. OPIMA has been designed to leverage the inherent massive parallelism within main memory while performing high-speed, low-energy optical computation to accelerate ML models based on convolutional neural networks. We present a comprehensive analysis of OPIMA to guide design choices and operational mechanisms. In addition, we evaluate the performance and energy consumption of OPIMA, comparing it with conventional electronic computing systems and emerging photonic PIM architectures. The experimental results show that OPIMA can achieve 2.98x higher throughput and 137x better energy efficiency than the best known prior work.

关键词： Energy consumption Costs Computer architecture optical computing optical fiber networks Throughput Energy efficiency Convolutional neural networks Photonics High-speed optical techniques machine learning (ML) acceleration photonic memory processing-in-memory (PIM) silicon photonics

来源：评论

学校读者我要写书评

暂无评论

Reliability analysis of optical neural networks with non-ideal signal transmission

引用

optical FIBER TECHNOLOGY 2024年 87卷

作者： Su, Ye Fu, Pengju Ye, Yichen Chai, Junxiong Jiang, Xiao Yang, Hongyu Xie, Yiyuan Southwest Univ Coll Elect & Informat Engn Chongqing Peoples R China Key Lab Nonlinear Circuits & Intelligent Informat Chongqing Peoples R China Key Lab Networks & Cloud Comp Secur Univ Chongqing Peoples R China

Machine learning plays a significant role in various fields. As a fundamental part of machine learning, matrix operations are one of the most important computational parts of neural networks. However, the advancement of Moore's law is hindered by limitations in the number and size of transistors on electronic chips. Therefore, implementing an optical neural network (ONN) using photonic devices to accomplish matrix-vector multiplication (MVM) has attracted significant attention. Compared with traditional electronic neural networks, ONN exhibits faster computational speed, higher parallel input capability and greater energy efficiency. Nevertheless, crosstalk and signal loss are inevitable characteristics in optical signal transmission, which can degrade the performance of ONN. In this paper, we present analysis models and simulations to assess the effects of crosstalk and loss on optical computing systems. We have observed that the accuracy of ONN degrades when crosstalk and loss are present. In severe cases, ONN is unable to complete the classification task. To mitigate the interference of crosstalk, we have designed a novel double-layer structure to reduce the number of optical devices required for signal transmission and optimize the impact of crosstalk on ONN performance. For the identical classification task, the ONN based on a double-layer structure exhibits better interference immunity than the generic ONN, with a 10% performance improvement. In particular, performance is improved by 50% when only crosstalk is considered.

关键词： Loss and crosstalk optical neural networks optical matrix computing

来源：评论

学校读者我要写书评

暂无评论

Human emotion recognition with a microcomb-enabled integrated optical neural network

引用

NANOPHOTONICS 2023年第20期12卷 3883-3894页

作者： Cheng, Junwei Xie, Yanzhao Liu, Yu Song, Junjie Liu, Xinyu He, Zhenming Zhang, Wenkai Han, Xinjie Zhou, Hailong Zhou, Ke Zhou, Heng Dong, Jianji Zhang, Xinliang Huazhong Univ Sci & Technol Wuhan Natl Lab Optoelect Wuhan 430074 Peoples R China Opt Valley Lab Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Sch Comp Sci & Technol Wuhan 430074 Peoples R China Univ Elect Sci & Technol China Key Lab Opt Fiber Sensing & Commun Networks Chengdu 611731 Peoples R China

State-of-the-art deep learning models can converse and interact with humans by understanding their emotions, but the exponential increase in model parameters has triggered an unprecedented demand for fast and low-power computing. Here, we propose a microcomb-enabled integrated optical neural network (MIONN) to perform the intelligent task of human emotion recognition at the speed of light and with low power consumption. Large-scale tensor data can be independently encoded in dozens of frequency channels generated by the on-chip microcomb and computed in parallel when flowing through the microring weight bank. To validate the proposed MIONN, we fabricated proof-of-concept chips and a prototype photonic-electronic artificial intelligence (AI) computing engine with a potential throughput up to 51.2 TOPS (tera-operations per second). We developed automatic feedback control procedures to ensure the stability and 8 bits weighting precision of the MIONN. The MIONN has successfully recognized six basic human emotions, and achieved 78.5 % accuracy on the blind test set. The proposed MIONN provides a high-speed and energy-efficient neuromorphic computing hardware for deep learning models with emotional interaction capabilities.

关键词： integrated optics optical computing optical neural network optical frequency comb human emotion recognition

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：