检索结果-南通市图书馆

Byzantine Robust Federated Learning Scheme Based on Backdoor Triggers

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computers, Materials & Continua 2024年第5期79卷 2813-2831页

作者： Zheng Yang Ke Gu Yiming Zuo school of computer and Communication engineering Changsha University of Science and TechnologyChangsha410114China

Federated learning is widely used to solve the problem of data decentralization and can provide privacy protectionfor data owners. However, since multiple participants are required in federated learning, this allows attackers tocompromise. Byzantine attacks pose great threats to federated learning. Byzantine attackers upload maliciouslycreated local models to the server to affect the prediction performance and training speed of the global model. Todefend against Byzantine attacks, we propose a Byzantine robust federated learning scheme based on backdoortriggers. In our scheme, backdoor triggers are embedded into benign data samples, and then malicious localmodels can be identified by the server according to its validation dataset. Furthermore, we calculate the adjustmentfactors of local models according to the parameters of their final layers, which are used to defend against datapoisoning-based Byzantine attacks. To further enhance the robustness of our scheme, each localmodel is weightedand aggregated according to the number of times it is identified as malicious. Relevant experimental data showthat our scheme is effective against Byzantine attacks in both independent identically distributed (IID) and nonindependentidentically distributed (non-IID) scenarios.

关键词： Federated learning Byzantine attacks backdoor triggers

Multi-scale persistent spatiotemporal transformer for long-term urban traffic flow prediction

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Electronic Science and Technology 2024年第1期22卷 53-69页

作者： Jia-Jun Zhong Yong Ma Xin-Zheng Niu Philippe Fournier-Viger Bing Wang Zu-kuan Wei school of computer Science and engineering University of Electronic Science and Technology of ChinaChengdu611731China College of computer Science&Software engineering Shenzhen UniversityShenzhen518060China school of computer Science Southwest Petroleum UniversityChengdu610500China

Long-term urban traffic flow prediction is an important task in the field of intelligent transportation,as it can help optimize traffic management and improve travel efficiency.To improve prediction accuracy,a crucial issue is how to model spatiotemporal dependency in urban traffic data.In recent years,many studies have adopted spatiotemporal neural networks to extract key information from traffic data.However,most models ignore the semantic spatial similarity between long-distance areas when mining spatial dependency.They also ignore the impact of predicted time steps on the next unpredicted time step for making long-term predictions.Moreover,these models lack a comprehensive data embedding process to represent complex spatiotemporal dependency.This paper proposes a multi-scale persistent spatiotemporal transformer(MSPSTT)model to perform accurate long-term traffic flow prediction in cities.MSPSTT adopts an encoder-decoder structure and incorporates temporal,periodic,and spatial features to fully embed urban traffic data to address these issues.The model consists of a spatiotemporal encoder and a spatiotemporal decoder,which rely on temporal,geospatial,and semantic space multi-head attention modules to dynamically extract temporal,geospatial,and semantic characteristics.The spatiotemporal decoder combines the context information provided by the encoder,integrates the predicted time step information,and is iteratively updated to learn the correlation between different time steps in the broader time range to improve the model’s accuracy for long-term prediction.Experiments on four public transportation datasets demonstrate that MSPSTT outperforms the existing models by up to 9.5%on three common metrics.

关键词： Graph neural network Multi-head attention mechanism Spatio-temporal dependency Traffic flow prediction

Lightweight Cross-Modal Multispectral Pedestrian Detection Based on Spatial Reweighted Attention Mechanism

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computers, Materials & Continua 2024年第3期78卷 4071-4089页

作者： Lujuan Deng Ruochong Fu Zuhe Li Boyi Liu Mengze Xue Yuhao Cui school of computer and Communication engineering Zhengzhou University of Light IndustryZhengzhou450002China

Multispectral pedestrian detection technology leverages infrared images to provide reliable information for visible light images, demonstrating significant advantages in low-light conditions and background occlusion scenarios. However, while continuously improving cross-modal feature extraction and fusion, ensuring the model’s detection speed is also a challenging issue. We have devised a deep learning network model for cross-modal pedestrian detection based on Resnet50, aiming to focus on more reliable features and enhance the model’s detection efficiency. This model employs a spatial attention mechanism to reweight the input visible light and infrared image data, enhancing the model’s focus on different spatial positions and sharing the weighted feature data across different modalities, thereby reducing the interference of multi-modal features. Subsequently, lightweight modules with depthwise separable convolution are incorporated to reduce the model’s parameter count and computational load through channel-wise and point-wise convolutions. The network model algorithm proposed in this paper was experimentally validated on the publicly available KAIST dataset and compared with other existing methods. The experimental results demonstrate that our approach achieves favorable performance in various complex environments, affirming the effectiveness of the multispectral pedestrian detection technology proposed in this paper.

关键词： Multispectral pedestrian detection convolutional neural networks depth separable convolution spatially reweighted attention mechanism

VKFQ:A Verifiable Keyword Frequency Query Framework with Local Differential Privacy in Blockchain

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computers, Materials & Continua 2024年第3期78卷 4205-4223页

作者： Youlin Ji Bo Yin Ke Gu school of computer and Communication engineering Changsha University of Science and TechnologyChangsha410114China

With its untameable and traceable properties,blockchain technology has been widely used in the field of data sharing.How to preserve individual privacy while enabling efficient data queries is one of the primary issues with secure data sharing.In this paper,we study verifiable keyword frequency(KF)queries with local differential privacy in blockchain.Both the numerical and the keyword attributes are present in data objects;the latter are sensitive and require privacy protection.However,prior studies in blockchain have the problem of trilemma in privacy protection and are unable to handle KF queries.We propose an efficient framework that protects data owners’privacy on keyword attributes while enabling quick and verifiable query processing for KF queries.The framework computes an estimate of a keyword’s frequency and is efficient in query time and verification object(VO)size.A utility-optimized local differential privacy technique is used for privacy protection.The data owner adds noise locally into data based on local differential privacy so that the attacker cannot infer the owner of the keywords while keeping the difference in the probability distribution of the KF within the privacy budget.We propose the VB-cm tree as the authenticated data structure(ADS).The VB-cm tree combines the Verkle tree and the Count-Min sketch(CM-sketch)to lower the VO size and query time.The VB-cm tree uses the vector commitment to verify the query results.The fixed-size CM-sketch,which summarizes the frequency of multiple keywords,is used to estimate the KF via hashing operations.We conduct an extensive evaluation of the proposed framework.The experimental results show that compared to theMerkle B+tree,the query time is reduced by 52.38%,and the VO size is reduced by more than one order of magnitude.

关键词： Security data sharing blockchain data query privacy protection

Multimodal Sentiment Analysis Based on a Cross-Modal Multihead Attention Mechanism

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computers, Materials & Continua 2024年第1期78卷 1157-1170页

作者： Lujuan Deng Boyi Liu Zuhe Li school of computer and Communication engineering Zhengzhou University of Light IndustryZhengzhou450002China

Multimodal sentiment analysis aims to understand people’s emotions and opinions from diverse data.Concate-nating or multiplying various modalities is a traditional multi-modal sentiment analysis fusion method.This fusion method does not utilize the correlation information between modalities.To solve this problem,this paper proposes amodel based on amulti-head attention mechanism.First,after preprocessing the original data.Then,the feature representation is converted into a sequence of word vectors and positional encoding is introduced to better understand the semantic and sequential information in the input sequence.Next,the input coding sequence is fed into the transformer model for further processing and learning.At the transformer layer,a cross-modal attention consisting of a pair of multi-head attention modules is employed to reflect the correlation between modalities.Finally,the processed results are input into the feedforward neural network to obtain the emotional output through the classification layer.Through the above processing flow,the model can capture semantic information and contextual relationships and achieve good results in various natural language processing tasks.Our model was tested on the CMU Multimodal Opinion Sentiment and Emotion Intensity(CMU-MOSEI)and Multimodal EmotionLines Dataset(MELD),achieving an accuracy of 82.04% and F1 parameters reached 80.59% on the former dataset.

关键词： Emotion analysis deep learning cross-modal attention mechanism

Prediction of Bandwidth of Metamaterial Antenna Using Pearson Kernel-Based Techniques

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computers, Materials & Continua 2024年第3期78卷 3449-3467页

作者： Sherly Alphonse S.Abinaya Sourabh Paul school of computer Science and engineering Vellore Institute of TechnologyChennai600127India school of Electronics engineering Vellore Institute of TechnologyChennai600127India

The use of metamaterial enhances the performance of a specific class of antennas known as metamaterial antennas.The radiation cost and quality factor of the antenna are influenced by the size of the antenna.Metamaterial antennas allow for the circumvention of the bandwidth restriction for small antennas.Antenna parameters have recently been predicted using machine learning algorithms in existing literature.Machine learning can take the place of the manual process of experimenting to find the ideal simulated antenna parameters.The accuracy of the prediction will be primarily dependent on the model that is used.In this paper,a novel method for forecasting the bandwidth of the metamaterial antenna is proposed,based on using the Pearson Kernel as a standard kernel.Along with these new approaches,this paper suggests a unique hypersphere-based normalization to normalize the values of the dataset attributes and a dimensionality reduction method based on the Pearson kernel to reduce the dimension.A novel algorithm for optimizing the parameters of Convolutional Neural Network(CNN)based on improved Bat Algorithm-based Optimization with Pearson Mutation(BAO-PM)is also presented in this work.The prediction results of the proposed work are better when compared to the existing models in the literature.

关键词： Antenna pearson optimization bandwidth metamaterial

Linear Quadratic Optimal control for Systems Governed by First-Order Hyperbolic Partial Differential Equations

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Systems Science & Complexity 2024年第1期37卷 230-252页

作者： XUE Xiaomin XU Juanjuan ZHANG Huanshui school of control Science and engineering Shandong UniversityJinan 250061China College of Electrical engineering and Automation Shandong University of Science and TechnologyQingdao 266000China

This paper focuses on linear-quadratic(LQ)optimal control for a class of systems governed by first-order hyperbolic partial differential equations(PDEs).Different from most of the previous works,an approach of discretization-then-continuousization is proposed in this paper to cope with the infinite-dimensional nature of PDE systems.The contributions of this paper consist of the following aspects:(1)The differential Riccati equations and the solvability condition of the LQ optimal control problems are obtained via the discretization-then-continuousization method.(2)A numerical calculation way of the differential Riccati equations and a practical design way of the optimal controller are proposed.Meanwhile,the relationship between the optimal costate and the optimal state is established by solving a set of forward and backward partial difference equations(FBPDEs).(3)The correctness of the method used in this paper is verified by a complementary continuous method and the comparative analysis with the existing operator results is presented.It is shown that the proposed results not only contain the classic results of the standard LQ control problem of systems governed by ordinary differential equations as a special case,but also support the existing operator results and give a more convenient form of computation.

关键词： Discretization-then-continuousization method first-order hyperbolic partial differential equations forward and backward partial difference equations linear quadratic optimal control.

Two-Way Neural Network Performance PredictionModel Based onKnowledge Evolution and Individual Similarity

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

computer Modeling in engineering & Sciences 2024年第2期138卷 1183-1206页

作者： Xinzheng Wang Bing Guo Yan Shen College of computer Science Sichuan UniversityChengdu610065China school of Information Science and engineering Guilin University of TechnologyGuilin541004China school of control engineering Chengdu University of Information TechnologyChengdu610225China

Predicting students’academic achievements is an essential issue in education,which can benefit many stakeholders,for instance,students,teachers,managers,etc.Compared with online courses such asMOOCs,students’academicrelateddata in the face-to-face physical teaching environment is usually sparsity,and the sample size is relativelysmall.It makes building models to predict students’performance accurately in such an environment even morechallenging.This paper proposes a Two-WayNeuralNetwork(TWNN)model based on the bidirectional recurrentneural network and graph neural network to predict students’next semester’s course performance using only theirprevious course achievements.Extensive experiments on a real dataset show that our model performs better thanthe baselines in many indicators.

关键词： computer education performance prediction deep learning

Cross-Dimension Attentive Feature Fusion Network for Unsupervised Time-Series Anomaly Detection

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

computer Modeling in engineering & Sciences 2024年第6期139卷 3011-3027页

作者： Rui Wang Yao Zhou Guangchun Luo Peng Chen Dezhong Peng school of computer Science and engineering University of Electronic Science and Technology of ChinaChengdu611731China school of computer and Software engineering Xihua UniversityChengdu610039China school of computer Science Sichuan UniversityChengdu610065China National Innovation Center for UHD Video Technology Chengdu610095China

Time series anomaly detection is crucial in various industrial applications to identify unusual behaviors within the time series data.Due to the challenges associated with annotating anomaly events,time series reconstruction has become a prevalent approach for unsupervised anomaly detection.However,effectively learning representations and achieving accurate detection results remain challenging due to the intricate temporal patterns and dependencies in real-world time series.In this paper,we propose a cross-dimension attentive feature fusion network for time series anomaly detection,referred to as CAFFN.Specifically,a series and feature mixing block is introduced to learn representations in 1D space.Additionally,a fast Fourier transform is employed to convert the time series into 2D space,providing the capability for 2D feature extraction.Finally,a cross-dimension attentive feature fusion mechanism is designed that adaptively integrates features across different dimensions for anomaly detection.Experimental results on real-world time series datasets demonstrate that CAFFN performs better than other competing methods in time series anomaly detection.

关键词： Time series anomaly detection unsupervised feature learning feature fusion