检索结果-内蒙古大学图书馆

Robust federated contrastive recommender system against targeted model poisoning attack

Science China(Information Sciences) 2025年第4期68卷 50-65页

作者： Wei YUAN Chaoqun YANG Liang QU Guanhua YE Quoc Viet Hung NGUYEN Hongzhi YIN School of Electrical Engineering and Computer Science The University of Queensland School of Information and Communication Technology Griffith University Deep Neural Computing Company Limited

Federated recommender systems(FedRecs) have garnered increasing attention recently, thanks to their privacypreserving benefits. However, the decentralized and open characteristics of current FedRecs present at least two ***, the performance of FedRecs is compromised due to highly sparse on-device data for each client. Second, the system's robustness is undermined by the vulnerability to model poisoning attacks launched by malicious users. In this paper, we introduce a novel contrastive learning framework designed to fully leverage the client's sparse data through embedding augmentation, referred to as CL4FedRec. Unlike previous contrastive learning approaches in FedRecs that necessitate clients to share their private parameters, our CL4FedRec aligns with the basic FedRec learning protocol, ensuring compatibility with most existing FedRec implementations. We then evaluate the robustness of FedRecs equipped with CL4FedRec by subjecting it to several state-of-the-art model poisoning attacks. Surprisingly, our observations reveal that contrastive learning tends to exacerbate the vulnerability of FedRecs to these attacks. This is attributed to the enhanced embedding uniformity, making the polluted target item embedding easily proximate to popular items. Based on this insight, we propose an enhanced and robust version of CL4FedRec(rCL4FedRec) by introducing a regularizer to maintain the distance among item embeddings with different popularity levels. Extensive experiments conducted on four commonly used recommendation datasets demonstrate that rCL4FedRec significantly enhances both the model's performance and the robustness of FedRecs.

关键词： federated recommender system contrastive learning model poisoning attack and defense

来源：评论

学校读者我要写书评

暂无评论

EPCNN: Efficient and Practical Privacy-Preserving Convolutional Neural Network Inference

引用

IEEE Transactions on Network Science and engineering 2025年第3期12卷 1567-1580页

作者： He, Gang Ren, Yanli Zhao, Jun Feng, Guorui Zhang, Xinpeng Shanghai University School of Communication and Information Engineering Shanghai200444 China Nanyang Technological University School of Computing and Data Science 639798 Singapore

Convolutional Neural Networks (CNNs), as a powerful tool for efficient inference, have rapidly developed into a Machine Learning as a Service paradigm facilitated by cloud computing. Nevertheless, this service model raises privacy concerns, particularly in scenarios where relying on two non-colluding servers is unfeasible. To address this issue, we present EPCNN, an Efficient and Practical CNN inference scheme, which concurrently ensures data privacy, model privacy, and inference results with only a single server. EPCNN leverages Paillier homomorphic encryption for secure convolution operations on encrypted data and involves minimal client-server interactions. The client only participates in evaluating non-linear activation functions by judging the signs of blinded convolution results to streamline the interactions and enhance the system's practicality. Our security analysis validates the reliability of the proposed scheme, while experimental results demonstrate high inference accuracy comparable to plaintext-based methods. Compared to the state-of-the-art work, EPCNN attains huge improvements in both runtime and communication overhead. © 2013 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

HEAT: Efficient Vision Transformer Accelerator With Hybrid-Precision Quantization

引用

IEEE Transactions on Circuits and Systems II: Express Briefs 2025年第5期72卷 758-762页

作者： Zhao, Pan Xue, Donghui Wu, Licheng Chang, Liang Tan, Haining Han, Yinhe Zhou, Jun University of Electronic Science and Technology of China School of Information and Communication Engineering Chengdu611731 China Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China

Quantization is an important technique for the acceleration of transformer-based neural networks. Prior related works mainly consider quantization from the algorithm level. Their hardware implementation is inefficient. In this brief, we propose an efficient vision transformer accelerator with retraining-free and finetuning-free hybrid-precision quantization. At the algorithm level, the features and weights are divided into two parts: normal values and outlier values. These two parts are quantized with different bit widths and scaling factors. We use matrix transformation and group-wise quantization policy to improve hardware utilization. At the hardware level, we propose a two-stage FIFO group structure and a hierarchical interleaving data flow to further improve the utilization of the PE array. As a result, the input and weight matrices are quantized to 5.71 bits on average with 0.526 % accuracy loss on Swin-T. The accelerator achieves a frame rate of 118.9 FPS and an energy efficiency of 43.58 GOPS/W on the ZCU102 FPGA board, better than state-of-the-art works. © 2004-2012 IEEE.

关键词： Particle accelerators

来源：评论

学校读者我要写书评

暂无评论

DDoS Detection in SDN Environments: Leveraging Machine Learning Algorithms with Local Interpretable Model-Agnostic Explanations (LIME) for Enhanced Security

DDoS Detection in SDN Environments: Leveraging Machine Lear...

引用

International Conference on Business Intelligence and Data Analytics, BIDA 2024

作者： Karna, Siwani Jahanve, Poluru Reddy Yadukul, S. Bhaskaran, Sreebha Vekkot, Susmitha Rajagopal, Shinu M. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bangalore India Department of Electronics and Communication Engineering Amrita School of Engineering Amrita Vishwa Vidyapeetham Bangalore India

ISBN: (纸本)9789819777167

SDN, short for Software Defined Networking, represents a network architecture controlled by software applications. Distributed Denial-of-Service (DDoS) attacks create a challenge to these SDN environments and lead to many issues like loss of data and network instability, and takes a lot of time to address requests made by human users. It is very crucial to detect such attacks, for users to seamlessly use the network and take advantage of what is available online. This study demonstrates the significance of Machine Learning algorithms in identifying malicious traffic. Employing algorithms Gradient Boosting Classifier (GBC), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Logistic Regression (LR) proves effective in detecting complex and dynamic attacks. The study utilizes Local Interpretable Model-agnostic Explanation (LIME) to analyze the strength of each feature that contributes towards the prediction of classification. The results highlight the superior performance of the Gradient Boosting Classifier, achieving an accuracy of 99.75% and an F1-score of 99.68% in detecting network anomalies and threats. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Application programs

来源：评论

学校读者我要写书评

暂无评论

Integrated AI-Based Pipeline for Law Enforcement Application: FIR Assistance 5th

Integrated AI-Based Pipeline for Law Enforcement Application...

引用

5th International Conference on computing and Network communications, CoCoNet 2023

作者： Amruth, A. Ramanan, R. Paul, Rhea Pranav, R. Gupta, Deepa Vekkot, Susmitha Nair, Priyanka C. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India Department of Electronics and Communication Engineering Amrita School of Engineering Amrita Vishwa Vidyapeetham Bengaluru India

ISBN: (纸本)9789819747108

This paper presents an integrated AI-based pipeline for law enforcement applications, specifically targeting the creation of first information report (FIR) in the Kannada language. The pipeline comprises a speech processing block utilizing automatic speech recognition (ASR) models to convert spoken Kannada into text, and a natural language processing block employing advanced AI techniques for information extraction and generating concise FIR summaries. The research focuses on optimizing and streamlining the FIR creation process by leveraging AI technologies and overcoming data scarcity in the Kannada language through data augmentation techniques. OpenAI’s Whisper is used for ASR, resulting in two-word error rate in this use case. For information extraction, a Bi-LSTM neural network is employed for named entity recognition (NER) with a performance of 94%. The proposed NER model utilizes three tags, which can be expanded to accommodate additional tags as per FIR requirements. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： FIR filters

来源：评论

学校读者我要写书评

暂无评论

Multi-Scale Time Series Segmentation Network Based on Eddy Current Testing for Detecting Surface Metal Defects

引用

IEEE/CAA Journal of Automatica Sinica 2025年第3期12卷 528-538页

作者： Xiaorui Li Xiaojuan Ban Haoran Qiao Zhaolin Yuan Hong-Ning Dai Chao Yao Yu Guo Mohammad S.Obaidat George Q.Huang the School of Intelligence Science and Technology University of Science and Technology Beijing the Beijing Advanced Innovation Center for Materials Genome Engineering the Key Laboratory of Intelligent Bionic Unmanned Systems and the Institute of Materials Intelligent Technology Liaoning Academy of Materials IEEE the Department of Computer Science Hong Kong Baptist University the School of Computer and Communication Engineering Key Laboratory of Advanced Materials and Devices for Post-Moore Chips Ministry of Education University of Science and Technology Beijing the Beijing Advanced Innovation Center for Materials Genome Engineering University of Science and Technology Beijing the School of Computer and Communication Engineering University of Science and Technology Beijing the King Abdullah Ⅱ School of Information Technology The University of Jordan the Department of Computational Intelligence the School of Computing SRM University the School of Engineering The Amity University The Hong Kong Polytechnic University

In high-risk industrial environments like nuclear power plants, precise defect identification and localization are essential for maintaining production stability and safety. However, the complexity of such a harsh environment leads to significant variations in the shape and size of the defects. To address this challenge, we propose the multivariate time series segmentation network(MSSN), which adopts a multiscale convolutional network with multi-stage and depth-separable convolutions for efficient feature extraction through variable-length templates. To tackle the classification difficulty caused by structural signal variance, MSSN employs logarithmic normalization to adjust instance distributions. Furthermore, it integrates classification with smoothing loss functions to accurately identify defect segments amid similar structural and defect signal subsequences. Our algorithm evaluated on both the Mackey-Glass dataset and industrial dataset achieves over 95% localization and demonstrates the capture capability on the synthetic dataset. In a nuclear plant's heat transfer tube dataset, it captures 90% of defect instances with75% middle localization F1 score.

关键词： Eddy current testing nondestructive testing semantic segmentation time series analysis

来源：评论

学校读者我要写书评

暂无评论

Signal Accumulation and Parameter Estimation for Target Detection on Dual-Function Radar and communication System

引用

IEEE Internet of Things Journal 2025年第11期12卷 15957-15969页

作者： Ji, Wenshuai Wang, Yu Tang, Yanqun Liu, Fan Tian, Biao Gu, Tao Sun Yat-sen University School of Electronics and Communication Engineering Shenzhen Campus Shenzhen518107 China Southern University of Science and Technology School of System Design and Intelligent Manufacturing Shenzhen518055 China Macquarie University School of Computing SydneyNSW2109 Australia

Spectrum competition and hardware complexity inherent in communication and radar systems can be alleviated by a dual-function radar communication (DFRC) systems. However, enhancing detection capabilities for maneuvering or weak targets remains a significant challenge, as traditional radar signal accumulation algorithms are not directly applicable to DFRC systems. This paper proposes a novel method that integrates signal accumulation and parameter estimation for target detection in DFRC systems. The approach employs an orthogonal frequency division multiplexing (OFDM) waveform within a multiple-input multiple-output (MIMO) framework, enabling the detection of high-speed or weak targets through a computationally efficient signal accumulation process. The proposed method comprises three key steps: first, the designed signals combined with the multiple signal classification (MUSIC) algorithm for target angle estimation in the spatial domain. Second, a two-step signal accumulation process is introduced, which separately extracts range and velocity information for weak or high-speed targets. Third, acceleration is derived based on velocity information and accumulation time. We present explicit expressions and detailed analyses of various performance metrics, including accumulation output response for high-speed targets, multiple target scenarios, and cases with low signal-to-noise ratio (SNR). Additionally, Cram4er-Rao bounds (CRBs) for azimuth, range cell, and velocity cell estimation in DFRC MIMO-OFDM systems are derived. Simulation results validate the proposed method, demonstrating its superior target detection performance compared to existing techniques. © 2014 IEEE.

关键词： Orthogonal frequency division multiplexing

来源：评论

学校读者我要写书评

暂无评论

Distributed Distortion-Aware Beamforming Designs for Cell-Free mMIMO Systems

引用

IEEE Journal on Selected Topics in Signal Processing 2025年第2期19卷 381-397页

作者： Liu, Mengzhen Li, Ming Liu, Rang Liu, Qian Dalian University of Technology School of Information and Communication Engineering Dalian116024 China University of California Center for Pervasive Communications and Computing IrvineCA92697 United States Dalian University of Technology School of Computer Science and Technology Dalian116024 China

Cell-free massive multi-input multi-output (CF-mMIMO) systems have emerged as a promising paradigm for next-generation wireless communications, offering enhanced spectral efficiency and coverage through distributed antenna arrays. However, the non-linearity of power amplifiers (PAs) in these arrays introduce spatial distortion, which may significantly degrade system performance. This paper presents the first investigation of distortion-aware beamforming in a distributed framework tailored for CF-mMIMO systems, enabling pre-compensation for beam dispersion caused by nonlinear PA distortion. Using a third-order memoryless polynomial distortion model, the impact of the nonlinear PA on the performance of CF-mMIMO systems is firstly analyzed by evaluating the signal-to-interference-noise-and-distortion ratio (SINDR) at user equipment (UE). Then, we develop two distributed distortion-aware beamforming designs based on ring topology and star topology, respectively. In particular, the ring-topology-based fully-distributed approach reduces interconnection costs and computational complexity, while the star-topology-based partially-distributed scheme leverages the superior computation capability of the central processor to achieve improved sum-rate performance. Extensive simulations demonstrate the effectiveness of the proposed distortion-aware beamforming designs in mitigating the effect of nonlinear PA distortion, while also reducing computational complexity and backhaul information exchange in CF-mMIMO systems. © 2025 IEEE.

关键词： Beamforming

来源：评论

学校读者我要写书评

暂无评论

A Feature Transformation Technique for Improving Ensemble Learning Systems 17th

A Feature Transformation Technique for Improving Ensemble L...

引用

17th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2025

作者： Nguyen, Truong Thanh Vu, Hieu Dang, Truong Elyan, Eyad Vu, Thanh Son Nguyen, Tien Thanh School of Computing Engineering and Technology RGU Aberdeen United Kingdom School of Information and Communication Technology Thuyloi University Hanoi Viet Nam

ISBN: (纸本)9789819660049

In this study, we propose a feature transformation approach to improve the performance of Ensemble Learning Systems. Our method operates on the predictions of base classifiers within an ensemble system, known as meta-data, which are produced by applying a Cross-Validation procedure on the training data using various learning algorithms. The goal is to transform the meta-data to obtain a better combined output. Specifically, we compute the center vectors associated with each class label on the meta-data. We then increase the dimensionality of the meta-data by concatenating it with residuals, which are obtained by subtracting each vector in the meta-data from the associated center vector. A classifier is then trained on this transformed meta-data to create the combiner which will be used to combine outputs of base classifiers for ensemble prediction. Our method is compared with three well-known ensemble methods: deep forest (gcForest), Decision Template (DT), Sum rule, and the same ensemble without transformation. We evaluated the experimental methods on 22 UCI datasets and used the Friedman and Nemenyi tests to statistically compare their performances. The results show that our method outperforms all base classifiers and the benchmark algorithms on experimental datasets. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Enhance Statistical Features with Changepoint Detection for Driver Behaviour Analysis 21st

Enhance Statistical Features with Changepoint Detection for ...

引用

21st Pacific Rim International Conference on Artificial Intelligence, PRICAI 2024

作者： Maktoubian, Jamal Tran, Son N. Shillabeer, Anna Amin, Muhammad Bilal Sambrooks, Lawrence School of Information and Communication Technology University of Tasmania Hobart Australia School of Information Technology Deakin University VIC Australia School of Computing Mathematics and Engineering Charles Sturt University NSW Australia

ISBN: (纸本)9789819601240

Driver behaviour modelling is a critical field that addresses complex and dynamic driving behaviours on roads with the goal of enhancing road safety, reducing air pollution, and improving vehicle performance. Recent advancements in sensor technology and machine learning (ML) techniques have facilitated the capture and analysis of driver behaviour patterns. Nonetheless, the efficacy of ML models heavily relies on the quality of the data used. Therefore, developing feature extraction techniques that provide high-quality inputs is crucial. In this paper, we conceptualised, implemented, and evaluated a novel feature model called Changepoint-based Statistical Feature (C-bSF). Initially, we extracted various statistical functions from raw sensor data, which were then aggregated using lagging windows. Following this, a changepoint detection method was used to derive the C-bSF feature. We compared the performance metrics of this new approach with other feature extraction methods, demonstrating the superiority of C-bSF in driver behaviour classification tasks across three datasets. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Adversarial machine learning

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：