检索结果-内蒙古大学图书馆

Over-the-Air distributed neural network in Internet of Things with Threat Modeling for Replay Attacks 1

International Conference on Communications, Signal processing, and Systems, CSPS 2023

作者： Ren, Chao Zeng, Chuyue Li, Yingqi Zhang, Haijun University of Science and Technology Beijng China

ISBN: (数字)9789819975020

ISBN: (纸本)9789819975556

Large scale distributed neural networks have demonstrated promise for various inference tasks in Internet of Things (IoT) devices, including intelligent security monitoring and defense against network threats. However, the massive amounts of data generated by IoT applications and limited computational capabilities present significant challenges in implementing typical applications, such as secure protocols for data confidentiality. Over-The-Air (OTA) computation, a recently proposed physical layer computing architecture, has great potential to address these issues. In this paper, we propose an OTA distributed neural network with the mutual benefit of joint computing and communication. However, the open channel environment in which the network’s forward computation is implemented renders OTA-based joint computing and communication methods vulnerable to replay attacks, thereby compromising the accuracy of the network performance and wasting valuable bandwidth resources due to backpropagation of contaminated information during OTA computing. A threat model of network is established to investigate the impact of replay attacks during the iterative process. Our analysis and numerical results demonstrate that the replay attacks have a significantly impact on the network. Specifically, the test accuracy rate decreases from 85 to 35%, and the convergence rate decreases by an average of 40%. When the number of iterations is set to 500, the success probability of replay attacks is 0.378. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

distributed neural network and Particle Swarm Optimization for Micro-grid Adaptive Power Allocation

引用

neural processing LETTERS 2022年第4期54卷 3215-3233页

作者： Fu, Zao He, Xing Liu, Ping Palizban, Ali Liao, Wengjing Southwest Univ Chongqing Key Lab Nonlinear Circuits & Intelligen Coll Elect & Informat Engn Chongqing 400715 Peoples R China British Columbia Inst Technol Burnaby BC V5A IS6 Canada Southwest Univ Coll Comp & Informat Sci Chongqing 400715 Peoples R China

The hybrid algorithm strategy proposed in this paper aims to combine the optimal power flow with voltage-var optimization to meet the load demand, reduce the transmission line losses and maintain the voltage within a practicable range. A distributed neural network algorithm is used to seek an optimal solution of active power flow which minimizes the cost of active power. In order to ensure that the optimal power flow will not cause a serious impact to the stability of the power grid, voltage-var optimization engines which employ a multi-algorithm coordination are presented to optimize the losses of power grid and the bus voltage. The simulation of IEEE 30-bus shows that the proposed hybrid algorithm strategy can not only minimize the cost of active power generation, but also satisfy the load demand under the precondition that all the bus voltage is within the reference range. The percentages of power losses comparisons verify that the proposed hybrid algorithm strategy can decrease the transmission line losses of the power grid effectively, which will not bring a serious influence to the stability of the power grid.

关键词： distributed neural networks Volatage-var optimization Hybrid algorithm strategy Smart grids

来源：评论

学校读者我要写书评

暂无评论

A 16-channel Reconfigurable neural Spike Sorting System with distributed Tiles and neural network Engine Based on FPGA

A 16-channel Reconfigurable Neural Spike Sorting System with...

引用

2023 IEEE Biomedical Circuits and Systems Conference, BioCAS 2023

作者： Li, Tianhao Hu, Siyan Li, Yongfu Liu, Yan Shanghai Jiao Tong University Department of Micro-Nano Electronics China

ISBN: (纸本)9798350300260

The increase in the amount of neural recording data will bring challenges for data communication, which has either limited bandwidth or power budget. Real-time neural spike detection and classification will greatly reduce the data bandwidth while keeping fidelity of recording, however, require manual tuning or complex algorithm for parameter optimization. In this work, we propose a customized machine-learning processing engine with distributed neural spike clustering tiles. A customized artificial neural network(ANN) module in the engine was trained offline and utilized online to optimize the initialization of existing sorting algorithm to achieve high sorting accuracy and minimized hardware complexity. By dynamically varying the data resolution of the sorting engine and distributed channels, the hardware resource occupied can be optimized. The algorithm was simulated on MATLAB and implemented on an FPGA platform. It shows that 94% sorting accuracy was achieved with 196 LUTs and 238 FFs on ZYNQ-7020 FPGA platform. The simulation results and FPGA implementation results prove the effectiveness of the sorting algorithm. © 2023 IEEE.

关键词： Improved K-means neural network neural spike sorting

来源：评论

学校读者我要写书评

暂无评论

Enhancing Security in distributed Computing Through Quantum neural network-Enabled Blockchain 7th

Enhancing Security in Distributed Computing Through Quantum ...

引用

7th International Conference on Intelligent Computing and Optimization, ICO 2023

作者： Xiuliang, Qiu Rajawat, Anand Singh Goyal, S.B. Solanki, Ram Kumar Chengyi University College Jimei University Xiamen361021 China Faculty of Information Technology City University Petaling Jaya46100 Malaysia School of Computers Science & Engineering Sandip University Nashik422213 India

ISBN: (纸本)9783031733239

It is more crucial than ever to maintain proper security in the constantly changing world of distributed computing due to the complexity and variety of cyber threats that are on the rise. In this study, we describe a novel approach that combines blockchain technology with quantum neural networks to increase the security of distributed systems. Quantum neural networks (QNNs) are based on the principles of quantum physics and are more robust and secure than classical systems. We take advantage of both the higher processing capability of quantum computing and the decentralized, unchangeable nature of blockchain by integrating QNN into a blockchain architecture. Our suggested methodology not only expedites transaction verifications but also greatly lowers exposure to common security issues like double spending and the Sybil attack. Initial tests reveal a considerable decline in fraudulent activity as well as an improvement in system speed and reduced latency. This study shows that a QNN-capable blockchain has the potential to herald in a new era of dependable, efficient, and secure distributed computing. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Quantum computers

来源：评论

学校读者我要写书评

暂无评论

Fully distributed Deep neural network: F2D2N 1

引用

9th International Conference on Mobile, Secure, and Programmable networking, MSPN 2023

作者： Leite, Ernesto Mourlin, Fabrice Paradinas, Pierre Conservatoire National des Arts et Métiers 292 rue Saint Martin Paris75141 France UPEC 61 Av. du Général de Gaulle Créteil94000 France

ISBN: (数字)9783031524264

ISBN: (纸本)9783031524257

Recent advances in Artificial Intelligence (AI) have accelerated the adoption of AI at a pace never seen before. Large Language Models (LLM) trained on tens of billions of parameters show the crucial importance of parallelizing models. Different techniques exist for distributing Deep neural networks but they are challenging to implement. The cost of training GPU-based architectures is also becoming prohibitive. In this document we present a distributed approach that is easier to implement where data and model are distributed in processing units hosted on a cluster of machines based on CPUs or GPUs. Communication is done by message passing. The model is distributed over the cluster and stored locally or on a datalake. We prototyped this approach using open sources libraries and we present the benefits this implementation can bring. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep neural network Training With distributed K-FAC

引用

IEEE TRANSACTIONS ON PARALLEL AND distributed SYSTEMS 2022年第12期33卷 3616-3627页

作者： Pauloski, J. Gregory Huang, Lei Xu, Weijia Chard, Kyle Foster, Ian T. Zhang, Zhao Univ Chicago Dept Comp Sci Chicago IL 60637 USA Texas Adv Comp Ctr Austin TX 78758 USA

Scaling deep neural network training to more processors and larger batch sizes is key to reducing end-to-end training time;yet, maintaining comparable convergence and hardware utilization at larger scales is challenging. Increases in training scales have enabled natural gradient optimization methods as a reasonable alternative to stochastic gradient descent and variants thereof. Kronecker-factored Approximate Curvature (K-FAC), a natural gradient method, preconditions gradients with an efficient approximation of the Fisher Information Matrix to improve per-iteration progress when optimizing an objective function. Here we propose a scalable K-FAC algorithm and investigate K-FAC's applicability in large-scale deep neural network training. Specifically, we explore layer-wise distribution strategies, inverse-free second-order gradient evaluation, and dynamic K-FAC update decoupling, with the goal of preserving convergence while minimizing training time. We evaluate the convergence and scaling properties of our K-FAC gradient preconditioner, for image classification, object detection, and language modeling applications. In all applications, our implementation converges to baseline performance targets in 9-25% less time than the standard first-order optimizers on GPU clusters across a variety of scales.

关键词： Training Parallel processing Program processors Convergence Computational modeling Data models Deep learning Optimization methods neural networks scalability high-performance computing

来源：评论

学校读者我要写书评

暂无评论

DistrEdge: Speeding up Convolutional neural network Inference on distributed Edge Devices 36

DistrEdge: Speeding up Convolutional Neural Network Inferenc...

引用

36th IEEE International Parallel and distributed processing Symposium (IEEE IPDPS)

作者： Hou, Xueyu Guan, Yongjie Han, Tao Zhang, Ning New Jersey Inst Technol Newark NJ 07102 USA Windsor Univ Windsor ON Canada

ISBN: (纸本)9781665481069

As the number of edge devices with computing resources (e.g., embedded GPUs, mobile phones, and laptops) increases, recent studies demonstrate that it can be beneficial to collaboratively run convolutional neural network (CNN) inference on more than one edge device. However, these studies make strong assumptions on the devices' conditions, and their application is far from practical. In this work, we propose a general method, called DistrEdge, to provide CNN inference distribution strategies in environments with multiple IoT edge devices. By addressing heterogeneity in devices, network conditions, and nonlinear characters of CNN computation, DistrEdge is adaptive to a wide range of cases (e.g., with different network conditions, various device types) using deep reinforcement learning technology. We utilize the latest embedded AI computing devices (e.g., NVIDIA Jetson products) to construct cases of heterogeneous devices' types in the experiment. Based on our evaluations, DistrEdge can properly adjust the distribution strategy according to the devices' computing characters and the network conditions. It achieves 1.1 to 3x speedup compared to state-of-the-art methods.

关键词： distributed computing convolutional neural network edge computing deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

On Generalized Zeroing neural network Under Discrete and distributed Time Delays and Its Application to Dynamic Lyapunov Equation

引用

IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS 2022年第8期52卷 5114-5126页

作者： Zuo, Qiuyue Li, Kenli Xiao, Lin Wang, Yaonan Li, Keqin Hunan Univ Coll Comp Sci & Elect Engn Changsha 410082 Peoples R China Hunan Normal Univ Hunan Prov Key Lab Intelligent Comp & Language In Changsha 410081 Peoples R China Hunan Univ Coll Elect & Informat Engn Changsha 410082 Peoples R China SUNY Coll New Paltz Dept Comp Sci New Paltz NY 12561 USA

Zeroing neural network (ZNN), an effective method for tracking solutions of dynamic equations, has been developed and improved by various strategies, typically the application of nonlinear activation functions (AFs) and varying parameters (VPs). Unlike VPs, AFs applied in ZNN models act directly on real-time error. The processing unit of v needs to obtain neural state in real time. In the implementation process, highly nonlinear AFs become an important cause of time delays, which eventually leads to instability and oscillation. However, most studies focus on exploring new theoretically valid AFs to improve performance of ZNNs, while ignoring the adverse effects of highly nonlinear AFs. The nonlinearity of AFs requires us fully consider time-delay tolerance of ZNNs using nonlinear AFs, so as to ensure that the model is not unstable even when disturbed by time delays. In this work, delay-perturbed generalized ZNN (DP-GZNN) is proposed to investigate time-delay tolerance of generalized ZNN (G-ZNN) in solving dynamic Lyapunov equation. Considering the nonlinearity of AFs, two delay terms are elegantly added to G-ZNN and DP-GZNN is then derived. After rigorous mathematical derivations, sufficient conditions in a linear matrix inequality (LMI) manner are presented for global convergence of DP-GZNN. Through rich numerical experiments, hyperparameters involved in the analysis process are discussed in detail. Comparative simulations are also conducted to compare the ability of different ZNN models to resist time delays. It is worth to mention that this is the first time to consider the ability of G-ZNN to resist discrete and distributed time delays.

关键词： Mathematical models Convergence Delay effects neural networks Delays Robustness Resists Discrete and distributed time delays dynamic Lyapunov equation (DLE) exponential convergence varying parameter (VP) zeroing neural network (ZNN)

来源：评论

学校读者我要写书评

暂无评论

DistrEE: distributed Early Exit of Deep neural network Inference on Edge Devices

DistrEE: Distributed Early Exit of Deep Neural Network Infer...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Peng, Xian Wu, Xin Xu, Lianming Wang, Li Fei, Aiguo Beijing China Beijing University of Posts and Telecommunications School of Electronic Engineering Beijing China

ISBN: (纸本)9798350351255

distributed DNN inference is becoming increasingly important as the demand for intelligent services at the network edge grows. By leveraging the power of distributed computing, edge devices can perform complicated and resource-hungry inference tasks previously only possible on powerful servers, enabling new applications in areas such as autonomous vehicles, industrial automation, and smart homes. However, it is challenging to achieve accurate and efficient distributed edge inference due to the fluctuating nature of the actual resources of the devices and the processing difficulty of the input data. In this work, we propose DistrEE, a distributed DNN inference framework that can exit model inference early to meet specific quality of service requirements. In particular, the framework firstly integrates model early exit and distributed inference for multi-node collaborative inferencing scenarios. Furthermore, it designs an early exit policy to control when the model inference terminates. Extensive simulation results demonstrate that DistrEE can efficiently realize efficient collaborative inference, achieving an effective trade-off between inference latency and accuracy. © 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

SMEGA²: distributed Asynchronous Deep neural network Training With a Single Momentum Buffer 51

SMEGA<SUP>2</SUP>: Distributed Asynchronous Deep Neural Netw...

引用

51st International Conference on Parallel processing (ICPP)

作者： Cohen, Rafi Hakimi, Ido Schuster, Assaf Technion Israel Inst Technol Dept Comp Sci Haifa Israel

ISBN: (纸本)9781450397339

As the field of deep learning progresses, and neural networks become larger, training them has become a demanding and time consuming task. To tackle this problem, distributed deep learning must be used to scale the training of deep neural networks to many workers. Synchronous algorithms, commonly used for distributing the training, are susceptible to faulty or straggling workers. Asynchronous algorithms do not suffer from the problems of synchronization, but introduce a new problem known as staleness. Staleness is caused by applying out-of-date gradients, and it can greatly hinder the convergence process. Furthermore, asynchronous algorithms that incorporate momentum often require keeping a separate momentum buffer for each worker, which cost additional memory proportional to the number of workers. We introduce a new asynchronous method, SMEGA(2), which requires a single momentum buffer regardless of the number of workers. Our method works in a way that lets us estimate the future position of the parameters, thereby minimizing the staleness effect. We evaluate our method on the CIFAR and ImageNet datasets, and show that SMEGA(2) outperforms existing methods in terms of final test accuracy while scaling up to as much as 64 asynchronous workers. Open-Source Code: https://***/rafi-cohen/SMEGA2

关键词： neural networks distributed deep learning asynchronous gradient descent

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：