检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Ullah, Rehmat Wu, Di Harvey, Paul Kilpatrick, Peter Spence, Ivor Varghese, Blesson The School of Computer Science University of St Andrews United Kingdom The School of Electronics Electrical Engineering and Computer Science Queen's University Belfast United Kingdom Autonomous Networking Research & Innovation Department Rakuten Mobile Japan

Federated learning (FL) is a privacy-preserving distributed machine learning technique that trains models while keeping all the original data generated on devices locally. Since devices may be resource constrained, offloading can be used to improve FL performance by transferring computational workload from devices to edge servers. However, due to mobility, devices participating in FL may leave the network during training and need to connect to a different edge server. This is challenging because the offloaded computations from edge server need to be migrated. In line with this assertion, we present FedFly, which is, to the best of our knowledge, the first work to migrate a deep neural network (DNN) when devices move between edge servers during FL training. Our empirical results on the CIFAR-10 dataset, with both balanced and imbalanced data distribution, support our claims that FedFly can reduce training time by up to 33% when a device moves after 50% of the training is completed, and by up to 45% when 90% of the training is completed when compared to state-of-the-art offloading approach in FL. FedFly has negligible overhead of up to two seconds and does not compromise accuracy. Finally, we highlight a number of open research issues for further investigation. FedFly can be downloaded from https://***/qub-blesson/FedFly. © 2021, CC BY.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Semi-Automatic Line-System Provisioning with Integrated Physical-Parameter-Aware Methodology: Field Verification and Operational Feasibility

arXiv

引用

arXiv 2024年

作者： Nishizawa, Hideki Borraccini, Giacomo Sasai, Takeo Huang, Yue-Kai Mano, Toru Anazawa, Kazuya Namiki, Masatoshi Usui, Soichiroh Matsumura, Tatsuya Sone, Yoshiaki Wang, Zehao Okamoto, Seiji Inoue, Takeru Ip, Ezra D’Amico, Andrea Chen, Tingjun Curri, Vittorio Wang, Ting Asahi, Koji Takasugi, Koichi NTT Network Innovation Labs 1-1 Hikarinooka Kanagawa Yokosuka Japan Optical Networking & Sensing Department NEC Laboratories America Inc. 4 Independence Way Suite 200 PrincetonNJ08540 United States NTT Network Innovation Center 3-9-11 Midori-choMusashino-shi Tokyo Japan Department of Electrical and Computer Engineering Duke University DurhamNC27708 United States Torino10129 Italy NEC Corporation Hinode Abiko Chiba1131 Japan

We propose methods and an architecture to conduct measurements and optimize newly installed optical fiber line systems semi-automatically using integrated physics-aware technologies in a data center interconnection (DCI) transmission scenario. We demonstrate, for the first time, digital longitudinal monitoring (DLM) and optical line system (OLS) physical parameter calibration working together in real-time to extract physical link parameters for transmission performance optimization. Our methodology has the following advantages over traditional design: a minimized footprint at user sites, accurate estimation of the necessary optical network characteristics via complementary telemetry technologies, and the capability to conduct all operation work remotely. The last feature is crucial, as it enables remote operation to implement network design settings for immediate response to quality of transmission (QoT) degradation and reversion in the case of unforeseen problems. We successfully performed semi-automatic line system provisioning over field fiber networks facilities at Duke University, Durham, NC. The tasks of parameter retrieval, equipment setting optimization, and system setup/provisioning were completed within 1 hour. The field operation was supervised by on-duty personnel who could access the system remotely from different time zones. By comparing Q-factor estimates calculated from the extracted link parameters with measured results from 400G transceivers, we confirmed that our methodology has a reduction in the QoT prediction errors (±0.3 dB) over existing design (±0.6 dB). Copyright © 2024, The Authors. All rights reserved.

关键词： Q factor measurement

来源：评论

学校读者我要写书评

暂无评论

Comprehensive Utility Function for Resource Allocation in Mobile Edge Computing

引用

computers, Materials & Continua 2021年第2期 1461-1477页

作者： Zaiwar Ali Sadia Khaf Ziaul Haq Abbas Ghulam Abbas Lei Jiao Amna Irshad Kyung Sup Kwak Muhammad Bilal Telecommunications and Networking(TeleCoN)Research Laboratory GIK Institute of Engineering Sciences and TechnologyTopi23640Pakistan Faculty of Electrical Engineering GIK Institute of Engineering Sciences and TechnologyTopi23640Pakistan Faculty of Computer Sciences and Engineering GIK Institute of Engineering Sciences and TechnologyTopi23640Pakistan Department of Information and Communication Technology University of Agder(UiA)Grimstad4898Norway Department of Information and Communication Engineering Inha UniversityIncheon22212Korea Department of Computer and Electronics Systems Engineering Hankuk University of Foreign StudiesGyeonggi-do17035Korea

In mobile edge computing(MEC),one of the important challenges is how much resources of which mobile edge server(MES)should be allocated to which user equipment(UE).The existing resource allocation schemes only consider CPU as the requested resource and assume utility for MESs as either a random variable or dependent on the requested CPU *** paper presents a novel comprehensive utility function for resource allocation in *** utility function considers the heterogeneous nature of applications that a UE offloads to *** proposed utility function considers all important parameters,including CPU,RAM,hard disk space,required time,and distance,to calculate a more realistic utility value for ***,we improve upon some general algorithms,used for resource allocation in MEC and cloud computing,by considering our proposed utility *** name the improved versions of these resource allocation schemes as comprehensive resource allocation *** UE requests are modeled to represent the amount of resources requested by the UE as well as the time for which the UE has requested these *** utility function depends upon the UE requests and the distance between UEs and MES,and serves as a realistic means of comparison between different types of UE ***(or selecting)an optimal MES with the optimal amount of resources to be allocated to each UE request is a challenging *** show that MES resource allocation is sub-optimal if CPU is the only resource *** taking into account the other resources,i.e.,RAM,disk space,request time,and distance in the utility function,we demonstrate improvement in the resource allocation algorithms in terms of service rate,utility,and MES energy consumption.

关键词： Cloud computing energy efficient resource allocation mobile edge computing service rate user equipment utility function

来源：评论

学校读者我要写书评

暂无评论

Generative AI for Advanced UAV networking

引用

IEEE Network 2024年

作者： Sun, Geng Xie, Wenwen Niyato, Dusit Du, Hongyang Kang, Jiawen Wu, Jing Sun, Sumei Zhang, Ping Jilin University College of Computer Science and Technology Changchun130012 China Jilin University Key Laboratory of Symbolic Computation and Knowledge Engineering Ministry of Education Changchun130012 China Nanyang Technological University College of Computing and Data Science Singapore639798 Singapore The University of Hong Kong Department of Electrical and Electronic Engineering Hong Kong999077 Hong Kong Guangdong University of Technology School of Automation Guangzhou510006 China Agency for Science Technology and Research Institute for Infocomm Research Singapore639798 Singapore Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing100876 China

With the impressive achievements of chatGPT and Sora, generative artificial intelligence (GAI) has received increasing attention. Not limited to the field of content generation, GAI is also widely used to solve the problems in wireless communication scenarios due to its powerful learning and generalization capabilities. Therefore, we discuss key applications of GAI in improving unmanned aerial vehicle (UAV) communication and networking performance in this article. Specifically, we first review the key technologies of GAI and the important roles of UAV networking. Then, we show how GAI can improve the communication, networking, and security performances of UAV systems. Subsequently, we propose a novel framework of GAI for advanced UAV networking, and then present a case study of UAV-enabled spectrum map estimation and transmission rate optimization based on the proposed framework to verify the effectiveness of GAI-enabled UAV systems. Finally, we discuss some important open directions. © 2024 IEEE.

关键词： Aircraft communication

来源：评论

学校读者我要写书评

暂无评论

Convolution Filter Pruning for Transfer Learning on Small Dataset

Convolution Filter Pruning for Transfer Learning on Small Da...

引用

2020 International computer Symposium, ICS 2020

作者： Liao, Ching Ya Liu, Pangfeng Wu, Jan-Jan National Taiwan University Department of Computer Science and Information Engineering Taipei Taiwan Graduate Institute of Networking and Multimedia National Taiwan University Taipei Taiwan Academia Sinica Institute of Information Science Taipei Taiwan

ISBN: (纸本)9781728192550

In this paper, we propose a scheme to reduce the size of a pre-trained full-scale model with a domain-specific dataset. This scheme combines model compression and transfer learning. First, it identifies the sensitive parts of a full model using the target dataset. Then it applies transfer learning on the identified part of the network to construct a reduced and customized model. Our scheme can correct structure and parameters to prune for a target dataset, which makes the following transfer learning more *** apply our scheme on image classification applications using convolutional neural networks. We observe that different image categories activate different filters, so we can identify sensitive parts of the model for different categories of *** use VGG-16 (pre-trained by ImageNet dataset) as the full model and we apply transfer learning on two different datasets (Flowers-102 and Cats vs. Dogs). We apply three pruning criteria in our scheme. The accuracy of Flowers-102 dataset by the best criterion drops less than 2% while pruning 60% of the network. We also observe the effects of different settings of our scheme and examine their feasibility by experiments. © 2020 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

FedAdapt: Adaptive Offloading for IoT Devices in Federated Learning

arXiv

引用

arXiv 2021年

作者： Wu, Di Ullah, Rehmat Harvey, Paul Kilpatrick, Peter Spence, Ivor Varghese, Blesson School of Computer Science University of St Andrews United Kingdom School of Electronics Electrical Engineering and Computer Science Queen’s University Belfast United Kingdom Autonomous Networking Research & Innovation Department Rakuten Mobile Japan

Applying Federated Learning (FL) on Internet-of-Things devices is necessitated by the large volumes of data they produce and growing concerns of data privacy. However, there are three challenges that need to be addressed to make FL efficient: (i) execution on devices with limited computational capabilities, (ii) accounting for stragglers due to computational heterogeneity of devices, and (iii) adaptation to the changing network bandwidths. This paper presents FedAdapt, an adaptive offloading FL framework to mitigate the aforementioned challenges. FedAdapt accelerates local training in computationally constrained devices by leveraging layer offloading of deep neural networks (DNNs) to servers. Further, FedAdapt adopts reinforcement learning based optimization and clustering to adaptively identify which layers of the DNN should be offloaded for each individual device on to a server to tackle the challenges of computational heterogeneity and changing network bandwidth. Experimental studies are carried out on a lab-based testbed and it is demonstrated that by offloading a DNN from the device to the server FedAdapt reduces the training time of a typical IoT device by over half compared to classic FL. The training time of extreme stragglers and the overall training time can be reduced by up to 57%. Furthermore, with changing network bandwidth, FedAdapt is demonstrated to reduce the training time by up to 40% when compared to classic FL, without sacrificing accuracy. © 2021, CC BY.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Comparison of Random Forest and Extreme Gradient Boosting Fingerprints to Enhance an indoor Wifi Localization System

Comparison of Random Forest and Extreme Gradient Boosting Fi...

引用

2021 International Mobile, Intelligent, and Ubiquitous Computing Conference, MIUCC 2021

作者： Niang, Mariame Ndong, Massa Dioum, Ibra Diop, Idy Mashaly, Maggie Ghany, Mohamed A. Abd El Cheikh Anta Diop University of Dakar Faculty of Science and Technology Department of Physics Dakar Senegal Virtual University of Senegal Science and Technology Dakar Senegal Cheikh Anta Diop University of Dakar Polytechnic High School of Dakar Researcher of Electronic and Telecommunications Computer Science Department Dakar Senegal German University in Cairo Networking Department Cairo Egypt German University in Cairo Cairo Faculty of Information Engineering and Technology Dept. of Electronics Egypt

ISBN: (纸本)9781665412438

Machine Learning framework adds a new dimension to the localization estimation problem. It tries to find the most likely position using processed features in a radio map. This paper compares the performance of two machine learning tools, Random Forest (RF) and Extreme Gradient Boosting (XGBoost), in exploiting the multipath information for Indoor localization problem. The investigation was carried out in a noisy indoor scenario, where Non-Line-Of-Sight (NLOS) between target and sensors may affect the location of Wi-Fi Access point strongly. It is possible to improve the position system performance by using fingerprints techniques that employ multipath information in a Machine Learning framework, which operates a dataset generated by real time. Usually, real measurements produce the fingerprints localization features, and there is mismatching with the simulated data. Another drawback of NLOS features extraction is the noise level that occurs in position processing. Random Forest algorithm uses fully grown decision trees to classify possible emitter position, trying to achieve error mitigation by reducing variance. On the other hand, XGBoost approach uses weak learners, defined by high bias and low variance. The results of the simulation show that XGBoost reaches a Mean Square Error (MSE) of 1.77m while RF has 1.83m, as shown in Fig. (4). In real time analysis, 40 points were used for testing. RF has an MSE 1.85m and XGBoost has 1.82m. The results are compared to the state of the art and recently published papers. © 2021 IEEE.

关键词： Mean square error

来源：评论

学校读者我要写书评

暂无评论

Execution delay-aware task assignment in mobile edge cloud and internet cloud

Execution delay-aware task assignment in mobile edge cloud a...

引用

2019 International Conference on Sustainable Technologies for Industry 4.0, STI 2019

作者： Ahmed, Jargis Ashhab, Md. Mahir Razzaque, Md. Abdur Rahman, Md. Mustafizur University of Dhaka Green Networking Research Group Department of Computer Science and Engineering Bangladesh Green University of Bangladesh Department of Computer Science and Engineering Bangladesh

ISBN: (纸本)9781728160979

Mobile Edge Cloud (MEC) provides computation ability close proximate of user compared to Internet Cloud (IC), which ensures flexibility and reliability in executing codes on mobile devices. Offloading resource hungry mobile application to MEC by mobile devices for computation is challenging. On the other hand, maintaining task provisioning within delay constraint is important too. This assignment policy still requires improvement in decreasing completion time of task and thus providing better Quality of Experience (QoE) for users. In this paper, we formulated a delay-aware task assignment as an optimization problem. The optimal assignment of task is Integer Linear Problem (ILP) and NP-Complete. As optimal solution is time costly, hence we proposed a heuristic algorithm, which is done by cooperating the MEC and IC servers to provide feasible task assignment within delay constraint. Extensive performance evaluation of the proposed algorithm indicates significant improvement in successful tasks execution and better QoE can be achieved compared to existing state-of-the-art works. © 2019 IEEE.

关键词： Quality of service

来源：评论

学校读者我要写书评

暂无评论

An energy-efficient scheduling of heterogeneous network cells in 5G

An energy-efficient scheduling of heterogeneous network cell...

引用

2019 International Conference on Sustainable Technologies for Industry 4.0, STI 2019

作者： Alom Shuvo, Md. Shahin Munna, Azad Rahaman Adhikary, Tamal Razzaque, Md. Abdur University of Dhaka Green Networking Research Group Department of Computer Science and Engineering Bangladesh Green University of Bangladesh Department of Computer Science and Engineering Bangladesh

ISBN: (数字)9781728160993

ISBN: (纸本)9781728160979

In fifth-generation (5G) wireless communication system, small cell networks (SCNs) are the key technology to provide high Quality of Service (QoS) to the mobile users. However, meeting exponentially increasing data rate of the users demands deploying large number of SCNs which is energy-hungry if they are kept active always. Existing energy-saving approaches offer limited improvement due to considering mere historical data-driven two states system model. In this paper, we assume four states for network cells and design an optimal scheduling of small cells to satisfy user's QoS using minimum number of activated small cells in a network. It optimally schedules a small cell in one of the four states at a given time. Our analysis depicts that proposed system offers significant performance improvement compared to the state-of-the-art work. © 2019 IEEE.

关键词： Heterogeneous networks

来源：评论

学校读者我要写书评

暂无评论

Minimizing execution cost of user application codes in mobile device cloud

Minimizing execution cost of user application codes in mobil...

引用

2019 International Conference on Sustainable Technologies for Industry 4.0, STI 2019

作者： Saha, Sajeeb Habib, Md. Ahsan Sarkar, Sujan Razzaque, Md. Abdur Rahman, Md. Mustafizur University of Dhaka Green Networking Research Group Department of Computer Science and Engineering Bangladesh Department of Computer Science and Engineering Green University of Bangladesh Bangladesh

ISBN: (纸本)9781728160979

Mobile Device Cloud (MDC) has become a promising collaborative cloud computing environment that exploits idle resources on nearby mobile devices to execute compute-intensive applications. In the existing MDC system, user tasks are executed by the worker devices either voluntarily or with an amount that leads to overpayment resulting dissatisfaction for both user and worker devices. Therefore, a major challenge is to minimize the execution cost of user codes while supporting sufficient payments to the workers for their used resources. Furthermore, selection of reputed workers becomes more complicated when the workers are having heterogeneous resources. In this paper, we formulate a greedy algorithm to minimize the task execution cost with reputed worker devices, while maintaining the desired quality of execution. The simulation results exhibit that our proposed algorithm can minimize the task execution cost up to 30 % compared to a state-of-the-art work. © 2019 IEEE.

关键词： Mobile cloud computing

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：