检索结果-内蒙古大学图书馆

photonics approaches to the implementation of neuromorphic computing

PHYSICS-USPEKHI 2023年第12期66卷 1284-1297页

作者： Musorin, A., I Shorokhov, A. S. Chezhegov, A. A. Baluyan, T. G. Safronov, K. R. Chetvertukhin, A., v Grunin, A. A. Fedyanin, A. A. Lomonosov Moscow State Univ Fac Phys Leninskie Gory 1Str 2 Moscow 119991 Russia

Physical limitations on the operation speed of electronic devices has motivated the search for alternative ways to process information. The past few years have seen the development of neuromorphic photonicsia branch of photonics where the physics of optical and optoelectronic devices is combined with mathematical algorithms of artiecial neural networks. Such a symbiosis allows certain classes of computation problems, including some involving artiecial intelligence, to be solved with greater speed and higher energy efeciency than can be reached with electronic devices based on the von Neumann architecture. We review optical analog computing, photonic neural networks, and methods of matrix multiplication by optical means, and discuss the advantages and disadvantages of existing approaches.

关键词： neuromorphic photonics artificial intelligence machine learning reservoir computing matrix +/- vector multiplication photonic computing neural networks optical coprocessor photonic tensor computing optical Fourier transform integrated photonics Mach +/- Zehnder interferometer ring resonators waveguides

来源：评论

学校读者我要写书评

暂无评论

Magneto-optical devices for photonic computing 5

Magneto-optical devices for photonic computing

引用

conference on AI and Optical Data Sciences v

作者： Shoji, Y. Tokyo Inst Technol Lab Future Interdisciplinary Res Sci & Technol Dept Elect & Elect Engn 2-12-1 OokayamaMeguro Ku Tokyo 1528550 Japan

ISBN: (纸本)9781510670679;9781510670662

A magneto-optical device provides a nonvolatile operation by integrating ferromagnetic materials, which enables reconfigurable photonic integrated circuits. In addition, an artificial neural network with all-optical signal processing based on magneto-optical memory can be expected. We present integrated magneto-optical devices and discuss applications to photonic computing.

关键词： photonic computing Magneto-optical effect Silicon photonics Neural network

来源：评论

学校读者我要写书评

暂无评论

Monolithically integrated optical computing component for photonic matrix multiplication based on electro-optically modulated vertical-cavity surface-emitting lasers 11

Monolithically integrated optical computing component for ph...

引用

conference on Semiconductor Lasers and Laser Dynamics XI

作者： Ledentsov, N., Jr. Dijkstra, J. Makarov, Yu Shchukin, v. A. Chorchos, L. Kropp, J-R Pernice, W. H. P. Ledentsov, N. N. VI Syst GmbH Hardenbergstr 7 D-10623 Berlin Germany Heidelberg Univ Kirchhoff Inst Phys Grabengasse 1 D-69117 Heidelberg Germany

ISBN: (纸本)9781510673236;9781510673229

Leading-edge machine learning algorithms require large amounts of matrix multiplications, which absorb significant computational resources in modern digital electronic systems. Analogue optical computing systems consisting of light sources, modulators and receivers may perform these operations with much higher efficiency. We introduce an integrated device based on electro-optically modulated (EOM) vertical-cavity surface-emitting lasers (vCSELs) that can perform analog multiplication at >28 GHz and at <20 mW power consumption. Due to its monolithic integration, the EOM vCSEL may be used as a building block for integrated optical computing devices. Development of such devices can help create 3-dimensionally integrated computing and communication systems.

关键词： vertical cavity surface-emitting laser optical computing analogue computing photonic integrated circuits

来源：评论

学校读者我要写书评

暂无评论

Adiabatic tapers based on photonic inverse design 32

Adiabatic tapers based on photonic inverse design

引用

conference on Physics and Simulation of Optoelectronic devices XXXII

作者： Dhingra, N. Chen, Y. H. Shenoy, v. Ansys Software Pvt Ltd Prestige Tech Pk2nd Floor Bengaluru 560103 India Ansys Taiwan Unit D 29F-17Sec 5 Xinyi Rd Taipei 110615 Taiwan

ISBN: (纸本)9781510670211;9781510670204

Adiabatic mode evolution is an important concept in photonic devices and is utilized in several applications including spot size conversion, mode coupling between the two waveguides, and broadband power splitters/couplers. In such devices, the linear adiabatic tapers are not always the best choice in terms of the device footprint and hence the taper profile needs to be optimized to provide small footprint while ensuring adiabatic mode evolution. In many instances, the best taper profile can be quite complex, and a parametric design might require a lot of parameters to accurately define the taper profile. In such cases, the traditional optimization using parameter sweep can be challenging. Several techniques such as Fast Quasi Adiabatic (FAQUAD) dynamics and constant loss approach have been proposed in the literature for such optimization. We implement these techniques and conduct a brief comparison using two different integrated photonic components. We also briefly discuss the challenges associated with relying solely on the optimization algorithms (inverse design) without utilizing constraints associated with these techniques.

关键词： photonic inverse design adiabatic tapers modal analysis FDTD

来源：评论

学校读者我要写书评

暂无评论

Cryogenic integrated optical modulators for classical and quantum computing

Cryogenic integrated optical modulators for classical and qu...

引用

conference on photonic computing - From Materials and devices to Systems and Applications

作者： Pintusa, Paolo Singh, Anshuman Ranzani, Leonardo Pinna, Sergio Xie, Weigiang Huang, Duanni Gustafsson, Martin, v Casula, Giovanni Andrea Shoji, Yuya Takamura, Yota Mizumoto, Tetsuya D. Moody, Galan Soltani, Mo Bowers, John E. Univ Calif Santa Barbara Santa Barbara CA 93106 USA Univ Cagliari Cagliari Italy Raytheon BBN Technol Cambridge MA USA Tokyo Inst Technol Tokyo Japan

ISBN: (纸本)9781510678866;9781510678873

In recent years, standard CMOS microprocessors have approached their maximum power dissipation per unit area, effectively placing a limit on computational power. This highlights the urgent need to explore alternative technologies. One promising avenue is the use of superconductors, which demonstrate zero resistivity below a critical temperature. However, circuits based on superconductors necessitate the use of cryostats to maintain low temperatures, presenting challenges in data transfer with the room temperature environment. While coaxial cables are often employed for this purpose, they suffer from limited data transfer rates and contribute significantly to heat load. On the contrary, photonics integrated circuits (PICs) coupled with optical fibers present a viable solution. They enable scalable, cost-effective, and power-efficient optical interconnections capable of supporting high data transfer rates while minimizing heat transfer. In this presentation, We will discuss the latest advancements in cryogenic PICs, focusing on their application in interfacing with cryogenic computing systems such as single-flux-quantum logic circuits and superconducting qubits.

关键词： photonic integrated circuits cryogenic devices electro-optic modulator magneto-optic modulator

来源：评论

学校读者我要写书评

暂无评论

Dynamic Edge/Cloud Resource Allocation for Distributed Computation under Semi-Static Demands 59

Dynamic Edge/Cloud Resource Allocation for Distributed Compu...

引用

59th Annual IEEE International conference on Communications (IEEE ICC)

作者： Sartzetakis, Ippokratis Pantazopoulos, Panagiotis Katsaros, Konstantinos v. Sourlas, vasilis varvarigos, Emmanouel

ISBN: (纸本)9781728190549

Edge computing is a recent paradigm where the processing takes place close to the data sources. It therefore reduces latency and saves bandwidth compared to traditional cloud computing. The latter can continue to play a supportive role. Edge-cloud computing provides benefits in many use cases including distributed computation algorithms, where the processing is divided into a number of tasks that are executed in parallel on different equipment. An important relevant challenge is to allocate the appropriate resources to process the data that are continuously generated from user devices. The issue becomes more complicated when we take into account the variations in the volume of the generated data as a function of time. In this paper we present a resource allocation algorithm for distributed computation with emphasis on machine learning algorithms. We consider that the resource requirements vary with time in a semi-static way that exhibits some daily pattern. We distinguish between periodic (expected) variations that occur during the day, and sporadic variations due to unexpected events. We propose an Integer Linear Programming algorithm to allocate the periodic resource requirements. To handle the non-periodic requirements, we consider a suitable prediction algorithm coupled with a reconfiguration algorithm that allocates the predicted required resources. Our results indicate that our proposal outperforms traditional allocation algorithms in terms of resource utilization, monetary cost and achieved accuracy.

关键词： cloud and edge computing distributed computing distributed machine learning resource allocation prediction dynamic

来源：评论

学校读者我要写书评

暂无评论

Physical Aware Clustering Training Method for Integrated photonic Convolution Neural Network with Nonlinear Distributed Weights

Physical Aware Clustering Training Method for Integrated Pho...

引用

Asia Communications and photonics conference (ACP) / International conference on Information photonics and Optical Communications (IPOC)

作者： Jiang, Yue Zhang, Wenjia Liu, Xuying Zhu, Wenyu He, Zuyuan Shanghai Jiao Tong Univ State Key Lab Adv Opt Commun Syst & Networks Shanghai 200240 Peoples R China

ISBN: (纸本)9781665481557

The computing power of the photonic Convolution Neural Network (CNN) have achieved Tera-level operations per second (OPS) for supporting machine vision algorithms. However, restrained by the nonlinear characteristics of weighting devices, it is hard to realize convergent training of photonic CNN under poor controlling precision. In this paper we proposed a Physical Aware Clustering training method where the Physical Aware Cluster Quantizer is embed with the straight through estimator (STE) algorithm for integrated photonic CNN with quantized and nonlinear distributed weights. Experiment shows that, by employing the upgraded STE methods (namely, STE plus), the photonic CNN using micro ring weighting bank and PAM4 controlling modules achieves the nearly 99.3% accuracy ratio for Fashion MNIST recognition task, whereas only 41.5% of that by using baseline STE training algorithms with the same physical devices.

关键词： component photonic integrated devices Optical computing photonic convolution neural network

来源：评论

学校读者我要写书评

暂无评论

Benchmarking Deep Learning Models for Object Detection on Edge computing devices 22nd

Benchmarking Deep Learning Models for Object Detection on Ed...

引用

22nd International conference on Service Oriented computing

作者： Alqahtani, Daghash K. Cheema, Muhammad Aamir Toosi, Adel N. Univ Melbourne Melbourne Vic Australia Monash Univ Melbourne Vic Australia

ISBN: (纸本)9789819608041;9789819608058

Modern applications, such as autonomous vehicles, require deploying deep learning algorithms on resource-constrained edge devices for real-time image and video processing. However, there is limited understanding of the efficiency and performance of various object detection models on these devices. In this paper, we evaluate the performance of several state-of-the-art object detection models, including YOLOv8 (Nano, Small, Medium), EfficientDet Lite (Lite0, Lite1, Lite2), and SSD (SSD MobileNet v1, SSDLite MobileDet), on popular edge devices such as the Raspberry Pi 3, 4, and 5 (with and without TPU accelerators), as well as the Jetson Orin Nano. We collect key performance metrics, including energy consumption, inference time, and Mean Average Precision (mAP). Our findings highlight models with lower mAP such as SSD MobileNet v1 are more energy-efficient and faster in inference, whereas higher mAP models like YOLOv8 Medium generally consume more energy and have slower inference, though with exceptions when accelerators like TPUs are used. Among the edge devices, Jetson Orin Nano stands out as the fastest and most energy-efficient option for request handling, despite having the highest idle energy consumption.

关键词： Deep Learning Object Detection Models Performance evaluation Inference Time Energy Efficiency Accuracy Edge

来源：评论

学校读者我要写书评

暂无评论

Leveraging v2P communications to improve the vRU's safety 20

Leveraging V2P communications to improve the VRU's safety

引用

20th IEEE International Wireless Communications and Mobile computing (IEEE IWCMC) - Green and Intelligent Communications

作者： Puga, Joel Fontainhas, Armando Andre Costa, Antonio Santos, Alexandre Nicolau, Maria Joao Meneses, Filipe Univ Minho Ctr Comp Graf Braga Portugal Univ Minho Algoritmi Res Ctr Braga Portugal

ISBN: (纸本)9798350361261;9798350361278

The advent of connected and autonomous vehicles (CAvs) holds immense promise for revolutionizing road safety and efficiency. Communications between the vehicles and other road users or elements (v2X) opens various possibilities that can prevent accidents and fatalities. vulnerable Road Users (vRUs), like pedestrians and cyclists, are at particular risk since a collision with a vehicle has a high likelihood of inflicting severe injuries. On the other hand, vRUs usually have one or more connected devices, and these can be used to increase their safety by allowing the prediction and prevention of collisions with CAvs. However, for this to be possible, it will be necessary to ensure not only reliable and low latency communication between the vRUs, CAvs and, eventually, Road Side Units (RSUs), but also the development of robust and reliable collision prediction algorithms capable of anticipating potential collisions and enabling timely interventions to prevent accidents. This paper presents a system that can be used to track a vehicle and a vRU, predict possible collisions and alert the vRU about the incoming danger. Additionally, it delves into the intricate aspects of collision prediction algorithms for CAvs, addressing the fundamental requirements for effective collision prediction and presenting a balanced approach that meets the demands of speed, accuracy, robustness, adaptability, and scalability.

关键词： v2P Collision detection vRU Safety v2X Wireless Communication

来源：评论

学校读者我要写书评

暂无评论

Energy-Efficiency Optimization in IoT Networks: algorithms, Techniques, and Case Studies 15

Energy-Efficiency Optimization in IoT Networks: Algorithms, ...

引用

15th International conference on Advances in computing, Control, and Telecommunication Technologies, ACT 2024

作者： Nagavelli, Umarani Dey, Niladri Sekhar Reddy, S. Pavan Kumar Department of AI & DS B V Raju Institute of Technology TS Narsapur India

ISBN: (纸本)9798331300579

The exponential growth of Internet of Things (IoT) devices has resulted in an unparalleled surge in the need for energy-efficient strategies to guarantee the sustainability and durability of IoT networks. This study article provides a thorough examination of energy-efficiency optimization in Internet of Things (IoT) networks, with a specific emphasis on the creation of algorithms, methodologies, and empirical investigations. The article commences by providing an overview of the significant significance of energy efficiency in Internet of Things (IoT) networks and its direct influence on the lifetime of devices, scalability of networks, and environmental sustainability. This emphasizes the urgent need for inventive approaches to tackle the difficulties presented by resource-limited Internet of Things (IoT) devices. Subsequently, the present study undertakes an in-depth examination of existing methodologies and algorithms for optimizing energy efficiency in Internet of Things (IoT) networks. This paper presents a comprehensive examination of each category, including insights into their respective strengths, limits, and suitability for various Internet of Things (IoT) applications. This study presents innovative algorithms and strategies that are especially developed to improve energy efficiency in Internet of Things (IoT) networks. These advancements use cutting-edge technology such as machine learning, edge computing, and low-power device design. The paper provides comprehensive explanations of these methodologies, accompanied by simulations and performance assessments, to showcase their efficacy in attaining energy conservation while maintaining network dependability and service excellence. The case studies presented provide valuable perspectives on the practical use of energy-efficient technologies, demonstrating their effectiveness in reducing operating expenses, carbon emissions, and system downtime. Moreover, this study aims to discuss the many problems and unre

关键词： Scalability

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：