检索结果-内蒙古大学图书馆

8th International Conference on Signal and Information processing, network and Computers, ICSINC 2021

作者： An, Tingting Sun, Songlin Liu, Rui School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing China Trusted Distributed Computing and Services Ministry of Education Key Laboratory Beijing China San Jose State University San JoseCA95192 United States

ISBN: (纸本)9789811933868

The development of 5G technology and Ultra HD video provide the basis for panoramic video, namely virtual reality (VR). At present, the traditional VQA method is not effective on panoramic video. Therefore, it is crucial to design objective VQA models for the standardization of panoramic video industry. With the development of deep learning, excellent algorithms of VQA methods based on convolutional neural network have emerged. In this paper, we propose a full reference VQA model based on spatial-temporal 3D convolutional neural network, the feature extraction combined the time and spatial information. we verify and optimize the proposed VQA model based on VQA-ODV panoramic video database, its objective score has a higher correlation with subjective scores than that of traditional VQA methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Improving network Resource Utilization for distributed Wireless Sensing Applications 24

Improving Network Resource Utilization for Distributed Wirel...

引用

2024 International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile networks and Mobile Computing, MobiHoc 2024

作者： Gayen, Snehadeep Shankar, Yamini Chakraborty, Ayon SENSE Lab IIT Madras India

ISBN: (纸本)9798400705212

Edge-assisted wireless sensing is increasingly popular, where complex neural network models perform inference tasks on wireless channel state information (CSI) data streamed from IoT devices. However large volumes of CSI data sent across the network for inference can significantly impact network bandwidth and reduce the Quality of Experience. This paper tackles the challenge of optimizing network resource utilization in wireless sensing systems by compressing and subsampling CSI streams. We evaluate methods that quantize and selectively subsample CSI data before transmission to the edge server, which is then fed to the inference models. Such approach reduces bandwidth and computational load, improving data transmission and processing efficiency. Experiments conducted in two real testbeds (indoors as well as outdoors) show how CSI compression preserves sensing information integrity while enhancing system performance in terms of latency, energy efficiency, and throughput. By integrating quantization and subsampling with edge computing, this work enhances wireless sensing systems, making them more scalable and efficient in utilizing network resources. © 2024 Copyright held by the owner/author(s).

关键词： Data transfer

来源：评论

学校读者我要写书评

暂无评论

Scale-Train: A Scalable DNN Training Framework for a Heterogeneous GPU Cloud

引用

IEEE ACCESS 2022年 10卷 68468-68481页

作者： Kim, Kyeonglok Lee, Hyeonsu Oh, Seungmin Seo, Euiseong Sungkyunkwan Univ Dept Comp Sci & Engn Suwon 16419 South Korea

In order to cope with the growing scale of deep neural network (DNN) models and training data, the use of cloud computing for distributed DNN training is becoming increasingly popular. The amount of available resources in a cloud continuously changes according to users' demands. Although distributed DNN training has a long execution time ranging from several hours to several days, existing frameworks cannot provide a dynamic scale function or have high scale in/out overhead. Therefore, it is difficult to achieve higher performance by adding graphics processing unit (GPU) nodes to a running training cluster, even when surplus GPU resources become available. In addition, the inability to dynamically reconfigure the training cluster prohibits the reform of the cluster topology when it was sub-optimally created. This paper proposes a dynamic scaling technique with which the dynamic addition and removal of new workers can be performed without suspending the ongoing training job. In addition, we propose a heterogeneity-aware straggler-proof technique so that, even when the performance of the GPUs in the cloud are uneven, a performance benefit can be guaranteed through the addition of the surplus resources. The proposed scheme improved throughput by up to a factor of 17.52 during scaling out the existing cluster of five workers to ten compared to the existing checkpoint-based scheme. Furthermore, training was continued at 95.52% of the maximum performance while being stopped for 841 seconds in Elastic Horovod, which supports dynamic scaling. Finally, even when GPUs of different performances were mixed, the error between the determined batch size and the optimal batch size was 3.37% on average.

关键词： Training Graphics processing units Computational modeling Cloud computing Synchronization Data models Computer architecture distributed training neural networks dynamic scaling heterogeneous cloud cluster management ring-allreduce

来源：评论

学校读者我要写书评

暂无评论

Effective distributed convolutional neural network architecture for remote sensing images target classification with a pre-training approach

引用

Journal of Systems Engineering and Electronics 2019年第2期30卷 238-244页

作者： LI Binquan HU Xiaohui School of Automation Science and Electrical Engineering Beihang University Institute of Software Chinese Academy of Sciences

How to recognize targets with similar appearances from remote sensing images(RSIs) effectively and efficiently has become a big challenge. Recently, convolutional neural network(CNN) is preferred in the target classification due to the powerful feature representation ability and better performance. However,the training and testing of CNN mainly rely on single *** machine has its natural limitation and bottleneck in processing RSIs due to limited hardware resources and huge time consuming. Besides, overfitting is a challenge for the CNN model due to the unbalance between RSIs data and the model *** a model is complex or the training data is relatively small,overfitting occurs and leads to a poor predictive performance. To address these problems, a distributed CNN architecture for RSIs target classification is proposed, which dramatically increases the training speed of CNN and system scalability. It improves the storage ability and processing efficiency of RSIs. Furthermore,Bayesian regularization approach is utilized in order to initialize the weights of the CNN extractor, which increases the robustness and flexibility of the CNN model. It helps prevent the overfitting and avoid the local optima caused by limited RSI training images or the inappropriate CNN structure. In addition, considering the efficiency of the Na¨?ve Bayes classifier, a distributed Na¨?ve Bayes classifier is designed to reduce the training cost. Compared with other algorithms, the proposed system and method perform the best and increase the recognition accuracy. The results show that the distributed system framework and the proposed algorithms are suitable for RSIs target classification tasks.

关键词： convolutional neural network (CNN) distributed architecture remote sensing images (RSIs) target classification pre-training

来源：评论

学校读者我要写书评

暂无评论

Parallel implementation of distributed acoustic sensor acquired signals: detection, processing, and classification

引用

JOURNAL OF APPLIED REMOTE SENSING 2022年第2期16卷

作者： Bencharif, Billel Alla Eddine Bayar, Salih Ozkan, Erkan Marmara Univ Elect & Comp Engn Istanbul Turkey SAMM Teknol Kocaeli Turkey

We aim to classify acoustic events recorded by a fiber optic distributed acoustic sensor (DAS). We derived the information from probing the fiber with light pulses and analyzing the Rayleigh backscatter. Then, we processed this data by a pipeline of processing algorithms to form the input for our machine learning classification model. We put random matrix theory to the test to distinguish the acoustic event of interest from the noise. We conditioned the raw trace using moving average and wavelet-based filtering algorithms to improve the signal-to-noise ratio. For raw, low pass, and wavelet-based filtered signals that we inject into a convolutional neural network, we rely on the magnitude of their complex coefficients to categorize the nature of the event. We also investigate Mel-Frequency Cepstral coefficients specific to the event as an input for the classifier and compare their performance to other signal representations. We run the experiments on the CNN for two-class and three-class classification using datasets from a DAS that is deployed for perimeter security and pipeline monitoring. We obtained the best results when using the MFCCs paired with wavelet denoising, achieving accuracies of 96.4% for the "event" class and 99.7% for the "no event" class when it comes to the two-class process. The three-class process yielded optimal accuracies of 83.3%, 81.3%, and 96.7% for the "digging," "walking," and "excavation" classes, respectively. Finally, the training execution time is exceptionally long because the extensive dataset and the model's architecture are complex. As a result, we make efficient use of the CPU and GPU to maximize our machine's power using the Keras API's sequence data generator. Compared with the serial implementation, we report an improvement of up to 4.87 times. (c) 2022 Society of Photo-Optical Instrumentation Engineers (SPIE)

关键词： classification convolutional neural networks distributed acoustic sensing parallel programming mel frequency cepstrum coefficient

来源：评论

学校读者我要写书评

暂无评论

Reconfigurable and Green FPGA Accelerator Design for Deep neural networks on IIoT Devices

Reconfigurable and Green FPGA Accelerator Design for Deep Ne...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Pourghasemian, Mohsen Lastovka, Martin Saghezchi, Firooz B. Gacanin, Haris Rwth Aachen University Distributed Signal Processing Aachen Germany

ISBN: (纸本)9798350351255

We propose an extremely energy efficient and reconfigurable accelerator for performing Deep neural network (DNN) inferences on a Field-Programmable Gate Array (FPGA). Our design allows on the fly reconfigurability of the model to adapt it to any arbitrary DNN inference task and perform it with an extremely low latency, on the scale of tens of micro seconds. Our design can be adopted by resource constrained Industrial Internet of Things (IIoT) devices (e.g., mobile robots) to fulfill different DNN inference tasks. By parallelization of the matrix multiplications in each layer of the DNN and adopting an adjustable sliding window at the synthesize stage, our accelerator strikes a balance between the computing latency and the scarce resource utilization on the FPGA. It also uses a low-overhead control scheme for the accelerator's internal interactions to update the DNN model on-the-fly. The results show that for a DNN model with 64 input features, 9 hidden layers (each with 64 neurons), and 64 output variables, our proposed inference accelerator speeds up the inference by 28.8 times and decreases the energy consumption by 31 % compared to an AMD Deep Learning processing Unit. © 2024 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Interference Management in Space-Air-Ground Integrated networks With Fully distributed Rate-Splitting Multiple Access

引用

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS 2025年第1期24卷 149-164页

作者： Zhang, Shengyu Mao, Yijie Clerckx, Bruno Quek, Tony Q. S. Singapore Univ Technol & Design Informat Syst Technol & Design Pillar Singapore 487372 Singapore Shanghai Tech Univ Sch Informat Sci & Technol Shanghai 201210 Peoples R China Imperial Coll London Dept Elect & Elect Engn London SW7 2AZ England Yonsei Univ Yonsei Frontier Lab Seoul 03722 South Korea

Despite the allure of ubiquitous, high-speed, and low-latency connectivity offered by Space-Air-Ground Integrated networks (SAGINs), the co-existence of Low Earth Orbit (LEO) satellites and Unmanned Aerial Vehicles (UAVs) within the same frequency band poses significant challenges in interference management. Traditional optimization approaches, requiring seconds or even minutes for beamforming design, simply cannot keep pace with this dynamic environment. This work addresses these challenges by proposing a Fully-distributed Rate-Splitting Multiple Access (FD-RSMA), which enables efficient cross-system interference management in SAGINs with statistical Channel State Information (CSI) at the Transmitter (CSIT). Building upon FD-RSMA, we study the precoder design of LEO satellites and UAVs along with common rate allocations of RSMA to maximize Weighted Ergodic Sum Rate (WESR). To handle channel randomness, we employ a Sample Average Approximation (SAA) approach. Furthermore, a Deep Learning (DL)-based precoder design algorithm, called GruCN, which marries the advantages of Gate Recurrent Unit (GRU) and Convolutional neural network (CNN), is proposed to efficiently tackle the non-convex optimization problem. Numerical results demonstrate the effectiveness and efficiency of our proposed DL-assisted FD-RSMA. Compared to conventional RSMA approaches, FD-RSMA improves up to 20% of WESR performance, while the GruCN achieves around 50% higher WESR performance and up to four orders of magnitude lower processing time than the conventional optimization approaches.

关键词： Interference Low earth orbit satellites Satellites Satellite broadcasting Optimization NOMA Resource management Quality of service Autonomous aerial vehicles Multicast communication Deep learning interference management rate-splitting multiple access (RSMA) space-air-ground integrated networks (SAGIN)

来源：评论

学校读者我要写书评

暂无评论

DRaNN_PSO: A deep random neural network with particle swarm optimization for intrusion detection in the industrial internet of things

引用

JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES 2022年第10期34卷 8112-8121页

作者： Ahmad, Jawad Shah, Syed Aziz Latif, Shahid Ahmed, Fawad Zou, Zhuo Pitropakis, Nikolaos Edinburgh Napier Univ Sch Comp Edinburgh EH10 5DT Scotland Coventry Univ Res Ctr Intelligent Healthcare Coventry England Fudan Univ Sch Informat Sci & Engn Shanghai Peoples R China NUST Pakistan Navy Engn Coll Dept Cyber Secur Karachi 75350 Pakistan

The Industrial Internet of Things (IIoT) is a rapidly emerging technology that increases the efficiency and productivity of industrial environments by integrating smart sensors and devices with the internet. The advancements in communication technologies have introduced stable connectivity and a higher data transfer rate in the IIoT. The IIoT devices generate a massive amount of information that requires intel-ligent data processing techniques for the development of cybersecurity mechanisms. In this regard, deep learning (DL) can be an appropriate choice. This paper proposes a Deep Random neural network (DRaNN) based fast and reliable attack detection scheme for IIoT environments. The RaNN is an advanced variant of the traditional Artificial neural network (ANN) with a highly distributed nature and better generalization capabilities. To attain a higher attack detection accuracy, the proposed RaNN is optimally trained by incorporating hybrid particle swarm optimization (PSO) with sequential quadratic programming (SQP). The SQP-enabled PSO facilitates the neural network to select optimal hyperparameters. The efficacy of the suggested scheme is analyzed in both binary and multiclass configurations by conducting extensive experiments on three new IIoT datasets. The experimental outcomes demonstrates the promising perfor-mance of the proposed design for all datasets. (c) 2022 The Author(s). Published by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY license (http://***/licenses/by/4.0/).

关键词： Cybersecurity Deep learning IIoT Intrusion detection Random neural network

来源：评论

学校读者我要写书评

暂无评论

Employ AI to Improve AI Services : Q-Learning Based Holistic Traffic Control for distributed Co-Inference in Deep Learning

引用

IEEE TRANSACTIONS ON SERVICES COMPUTING 2022年第2期15卷 627-639页

作者： Zhang, Chaofeng Dong, Mianxiong Ota, Kaoru Adv Inst Ind Technol Dept Informat & Elect Engn Tokyo 1400011 Japan Muroran Inst Technol Dept Informat & Elect Engn Muroran Hokkaido 0508585 Japan

As the inevitable part of intelligent service in the new era, the services for AI tasks themselves have received significant attention, which due to the urgency of energy and computing resources, is difficult to implement in a stable and widely distributed system and coordinately utilize remote edge devices and cloud. In this article, we introduce an AI-based holistic network optimization solution to schedule AI services. Our proposed deep Q-learning algorithm optimizes the overall throughput of AI co-inference tasks themselves by balancing the uneven computation resources and traffic conditions. We use a multi-hop DAG (Directed Acyclic Graph) to describe a deep neural network (DNN) based co-inference network structure and introduce the virtual queue to analyze the Lyapunov stability for the system. Then, a priority-based data forwarding strategy is proposed to maximize the bandwidth efficiency, and we develop a Real-time Deep Q-learning based Edge Forwarding Scheme Optimization Algorithm (RDFO) to maximize the overall task processing rate. Finally, we conduct the platform simulation for the distributed co-inference system. Through the comparison with other benchmarks, we testify to the optimality of our proposal.

关键词： Task analysis Cloud computing Artificial intelligence Internet of Things Optimization Feature extraction Deep learning Machine learning network management Internet of Things service-oriented systems resource management

来源：评论

学校读者我要写书评

暂无评论

EFFICIENT FEDERATED LEARNING WITH SMOOTH AGGREGATION FOR NON-IID DATA FROM MULTIPLE EDGES 49

EFFICIENT FEDERATED LEARNING WITH SMOOTH AGGREGATION FOR NON...

引用

49th IEEE International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Wang, Qianru Li, Qingyang Guo, Bin Cui, Jiangtao Xidian Univ Xian Shaanxi Peoples R China Northwestern Polytech Univ Xian Shaanxi Peoples R China

ISBN: (纸本)9798350344868;9798350344851

Federated learning (FL) learns an optimal global model by aggregating local models trained on distributed data from different devices. Due to heterogeneous data distributions across devices, local models will be divergent, resulting in the global model's performance degradation. Recent studies attempt to balance local models to obtain a global model that can adapt to each device. But they ignore a more challenging problem that redundant local models from devices will break the balance, resulting in the global model overfitting redundant local models. Therefore, we propose FedSmooth, a novel global aggregation algorithm. FedSmooth first identifies the redundant local models without sensitive local information (e.g., label distribution), then designs a smooth global aggregation to strengthen the effect of local models that can accelerate finding the optimal global model. Experimental results show that our method outperforms 4 SOTA baseline methods even if there is more redundancy.

关键词： Federated learning edge-collaborative computing redundant data deep neural network

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：