检索结果-内蒙古大学图书馆

Feature Reconstruction Attacks and Countermeasures of DNN Training in Vertical Federated Learning

IEEE Transactions on Dependable and Secure Computing 2025年第3期22卷 2659-2669页

作者： Ye, Peng Jiang, Zhifeng Wang, Wei Li, Bo Li, Baochun Hong Kong University of Science and Technology Department of Computer Science and Engineering Hong Kong University of Toronto Department of Electrical and Computer Engineering TorontoM5S 1A1 Canada

Federated learning (FL) has increasingly been deployed, in its vertical form, among organizations to facilitate secure collaborative training. In vertical FL (VFL), participants hold disjoint features of the same set of sample instances. The one with labels - the active party, initiates training and interacts with other participants - the passive parties. It remains largely unknown whether and how an active party can extract private feature data owned by passive parties, especially when training deep neural network (DNN) models. This work examines the feature security problem of DNN training in VFL. We consider a DNN model partitioned between active and passive parties, where the passive party holds a subset of the input layer with some features of binary values. Though proved to be NP-hard. we demonstrate that, unless the feature dimension is exceedingly large, it remains feasible, both theoretically and practically, to launch a reconstruction attack with an efficient search-based algorithm that prevails over current feature protection. We propose a novel feature protection scheme by perturbing intermediate results and fabricated input features, which effectively misleads reconstruction attacks towards pre-specified random values. The evaluation shows it sustains feature reconstruction attack in various VFL applications with negligible impact on model performance. © 2004-2012 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Photovoltaic power one-day and multistep-hourly AI predictions using node-by-node evolved binomial tree structures to form L-transformed PDE modules

引用

Soft Computing 2025年第5期29卷 2483-2495页

作者： Zjavka, Ladislav Department of Computer Science Faculty of Electrical Engineering and Computer Science VŠB-Technical University of Ostrava 17. Listopadu 15/2172 Ostrava Czech Republic

Photovoltaic (PV) power is generated by two common types of solar components that are primarily affected by fluctuations and development in cloud structures as a result of uncertain and chaotic processes. Local PV forecasting is unavoidable in supply and load planning necessary in integration of smart systems into electrical grids. Intra- or day-ahead modelling of weather patterns based on Artificial Intelligence (AI) allows one to refine available 24 h. cloudiness forecast or predict PV production at a particular plant location during the day. AI usually gets an adequate prediction quality in shorter-level horizons, using the historical meteo- and PV record series as compared to Numerical Weather Prediction (NWP) systems. NWP models are produced every 6 h to simulate grid motion of local cloudiness, which is additionally delayed and usually scaled in a rough less operational applicability. Differential Neural Network (DNN) is based on a newly developed neurocomputing strategy that allows the representation of complex weather patterns analogous to NWP. DNN parses the n-variable linear Partial Differential Equation (PDE), which describes the ground-level patterns, into sub-PDE modules of a determined order at each node. Their derivatives are substituted by the Laplace transforms and solved using adapted inverse operations of Operation Calculus (OC). DNN fuses OC mathematics with neural computing in evolution 2-input node structures to form sum modules of selected PDEs added step-by-step to the expanded composite model. The AI multi- 1…9-h and one-stage 24-h models were evolved using spatio-temporal data in the preidentified daily learning sequences according to the applied input–output data delay to predict the Clear Sky Index (CSI). The prediction results of both statistical schemes were evaluated to assess the performance of the AI models. Intraday models obtain slightly better prediction accuracy in average errors compared to those applied in the second-day-ahead

关键词： Laplace transforms

来源：评论

学校读者我要写书评

暂无评论

Developing a Comprehensive NLP Framework for Indigenous Dialect Documentation and Revitalization

引用

International Journal of Advanced computer science and Applications 2025年第4期16卷 815-823页

作者： Fakhreldin, Mohammed Department of Computer Science-College of Engineering and Computer Science Jazan University Jazan45142 Saudi Arabia

The disappearance of Indigenous languages results in a decrease in cultural diversity, hence making the preservation of these languages extremely important. Conventional methods of documentation are lengthy, and the present AI solutions somehow do not deliver due to data scarcity, dialectal variation, and poor adaptability to low-resource languages. A novel NLP framework is being proposed to solve the existing problems. This framework intermixes Meta-Learning and Contrastive Learning to counter these problems. Thus, adaptation to low-resourced languages becomes rapid via meta-learning (MAML), while dialect differentiation is enhanced through contrastive learning. The model training is carried out on Tatoeba (text) and Mozilla Common Voice (speech) datasets to ensure robust performance in both text and phonetic tasks. The results indicate that there is a reduction of 15% in Word Error Rate (WER), an 18% improvement in BLEU score corresponding to translation, and a 12% improvement in F1-score related to dialect classification. The testing was also done with native speakers to assess its practical viability. It is a real-time translation, transcription, and language documentation system deployed via a cloud-based platform, thereby reaching out to Indigenous communities globally. This dual-learning framework represents a scalable, adaptive, and cost-efficient solution for the revitalization of languages. The models proposed have been a game changer for language preservation, have set new standards for low-resource NLP, and have made some tangible contributions towards the digital sustainability of endangered dialects. © (2025), (science and Information Organization). All Rights Reserved.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

A Magnetic Signal Based Device Fingerprinting Scheme in Wireless Charging

引用

IEEE Transactions on Dependable and Secure Computing 2025年第3期22卷 2943-2958页

作者： Li, Jiachun Meng, Yan Zhang, Le Chen, Guoxing Tian, Yuan Zhu, Haojin Shanghai Jiao Tong University Department of Computer Science and Engineering Shanghai200240 China Electrical and Computer Engineering Department Institute for Technology Los AngelesCA90095 United States

Wireless charging is widely used to charge smart devices with limited battery capacity. However, it is susceptible to the identity spoofing attack, where adversaries can impersonate malicious devices as legitimate ones to gain unauthorized access and potentially disrupt the wireless charging system (e.g., resulting in incorrect billing, overheating, or even explosions). Device fingerprinting is a classical method for defending against identity spoofing attacks. However, applying existing schemes in wireless charging scenarios has drawbacks such as inconvenience (e.g., requiring specialized devices or user participation) and ineffectiveness (e.g., vulnerability to spoofing). Thus, we design a novel passive, effective, and robust device fingerprinting scheme called MagID for wireless charging systems. The insight of MagID lies in the fact that during wireless charging, the magnetic signal around a device can reflect inherent hardware differences. These differences can be extracted as unique fingerprints for authentication purposes. MagID leverages a novel scheme, SUPER-ARRAY, to precisely measure magnetic data and generate effective fingerprints for authenticating a device's identity before starting charging progress. Experimental results demonstrate that MagID achieves an accuracy rate of 98.14% across various charging devices. We have also tested its performance under different impact factors and verified its compatibility with various wireless charging pads. © 2004-2012 IEEE.

关键词： Charging stations

来源：评论

学校读者我要写书评

暂无评论

Drug–target interactions prediction based on similarity graph features extraction and deep learning

引用

Neural Computing and Applications 2025年第6期37卷 4303-4322页

作者： Torkey, Hanaa El-Behery, Heba Attia, Abdel-Fattah El-Fishawy, Nawal Computer Science and Engineering Department Faculty of Electronic Engineering Menoufia University Menouf Egypt Department of Computer Science and Engineering Faculty of Engineering Kafrelsheikh University Kafrelsheikh Egypt Department of Computer Science College of Computer Engineering and Sciences Prince Sattam Bin Abdulaziz University Al-Kharj Saudi Arabia

Identifying drug–target interactions (DTIs) is a critical step in both drug repositioning. The labor-intensive, time-consuming, and costly nature of classic DTI laboratory studies makes it imperative to create efficient computer algorithms to forecast possible DTIs. However, current computational approaches that predict potential drug–target interactions (DTIs) suffer from some limitations, like finding the best similarity measures or negative samples, and thus require substantial performance improvement. This study proposes an integrated approach based on feature representation and deep learning to predict DTIs. We extract the relevant features of drugs and proteins from heterogeneous networks using graph mining techniques. The proposed approach constructs a heterogeneous graph from the known drug–protein interactions, protein–protein, and drug–drug similarities. Then applying two feature extraction techniques to extract the features, then utilizing these features in training a deep learning model to predict the potential DTIs. Also, a novel algorithm is proposed to find the negative samples based on the drug and protein similarity matrices. Four Benchmark datasets are used to evaluate the proposed approach. Our approach achieves the highest AUC (area under the ROC curve) across all datasets (0.98) with around 2% increases over the existing methods. Experimental results demonstrate that our proposed approach outperforms the baseline methods in predicting DTI, and our negative sample-identifying algorithm could be established as a competitive solution. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

MLRT-UNet:An Efficient Multi-Level Relation Transformer Based U-Net for Thyroid Nodule Segmentation

引用

computer Modeling in engineering & sciences 2025年第4期143卷 413-448页

作者： Kaku Haribabu Prasath R Praveen Joe IR Department of Computer Science and Engineering RMK College of Engineering and TechnologyTiruvallur601206India School of Computer Science and Engineering Vellore Institute of TechnologyChennai600127India

Thyroid nodules,a common disorder in the endocrine system,require accurate segmentation in ultrasound images for effective diagnosis and ***,achieving precise segmentation remains a challenge due to various factors,including scattering noise,low contrast,and limited resolution in ultrasound *** existing segmentation models have made progress,they still suffer from several limitations,such as high error rates,low generalizability,overfitting,limited feature learning capability,*** address these challenges,this paper proposes a Multi-level Relation Transformer-based U-Net(MLRT-UNet)to improve thyroid nodule *** MLRTUNet leverages a novel Relation Transformer,which processes images at multiple scales,overcoming the limitations of traditional encoding *** transformer integrates both local and global features effectively through selfattention and cross-attention units,capturing intricate relationships within the *** approach also introduces a Co-operative Transformer Fusion(CTF)module to combine multi-scale features from different encoding layers,enhancing the model’s ability to capture complex patterns in the ***,the Relation Transformer block enhances long-distance dependencies during the decoding process,improving segmentation *** results showthat the MLRT-UNet achieves high segmentation accuracy,reaching 98.2% on the Digital Database Thyroid Image(DDT)dataset,97.8% on the Thyroid Nodule 3493(TG3K)dataset,and 98.2% on the Thyroid Nodule3K(TN3K)*** findings demonstrate that the proposed method significantly enhances the accuracy of thyroid nodule segmentation,addressing the limitations of existing models.

关键词： Thyroid nodules endocrine system multi-level relation transformer U-Net self-attention external attention co-operative transformer fusion thyroid nodules segmentation

来源：评论

学校读者我要写书评

暂无评论

Monitoring COVID-19 environment: a real-time facial mask detection using YOLO models

引用

Research on Biomedical engineering 2025年第1期41卷 1-18页

作者： Pandiyan, P. Thangaraj, Rajasekaran V, Balasubramanian V. K, Manavalasundaram S, Balakrishnan Department of Electrical and Electronics Engineering KPR Institute of Engineering and Technology Coimbatore India Department of Computer Science and Engineering Nandha Engineering College Erode India Department of Information Technology Velalar College of Engineering and Technology Erode India Department of Computer Science and Engineering Aarupadai Veedu Institute of Technology Chennai India Department of Computer Science and Engineering Sri Ramakrishna Institute of Technology Coimbatore India

Purpose: The rapid spread of COVID-19 has resulted in significant harm and impacted tens of millions of people globally. In order to prevent the transmission of the virus, individuals often wear masks as a protective measure for themselves and others. Coronavirus protection guidelines have been published by the World Health Organization (WHO). According to WHO standards, COVID-19 can be prevented by wearing a mask in public places and congested regions. In these places, it is very difficult to personally check to see if people are wearing face masks or not. Methods: The objective of this research work is to build a powerful, efficient, and real-time approach for detecting people not wearing masks. Three cutting-edge object identification models, namely YOLOv4, Tiny-YOLOv4, and YOLOv5, are employed in this study for the identification of masked faces. Result: The proposed YOLOv5 model is evaluated using real-time images captured using a smartphone or tablet. The test images include both single and multiple people with and without masks. The YOLOv5 model achieved recognition accuracy of 88.90% with an average detection speed of 0.0316 s per image, whereas the YOLOv4 and Tiny-YOLOv4 produced recognition accuracy of 82.24% and 74.80% with an average detection speed of 0.0530 s and 0.0541 s per image, respectively. Conclusion: The comparative performance suggests that the YOLOv5 model has a maximum recognition accuracy of 88.90% in face mask identification tasks compared to other models such as the YOLOv4 and Tiny-YOLOv4. © The Author(s), under exclusive licence to The Brazilian Society of Biomedical engineering 2025.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

The Duality Upper Bound for Finite-State Channels With Feedback

引用

IEEE Transactions on Information Theory 2025年第5期71卷 3255-3270页

作者： Huleihel, Bashar Sabag, Oron Aharoni, Ziv Permuter, Haim H. Ben-Gurion University of the Negev Department of Electrical and Computer Engineering Be’er Sheva84105 Israel The Hebrew University of Jerusalem Rachel and Selim Benin School of Computer Science and Engineering Jerusalem91904 Israel Duke University Department of Electrical and Computer Engineering DurhamNC27708 United States

This paper investigates the capacity of finite-state channels (FSCs) with feedback. We derive an upper bound on the feedback capacity of FSCs by extending the duality upper bound method from mutual information to the case of directed information. The upper bound is expressed as a multi-letter expression that depends on a test distribution on the sequence of channel outputs. For any FSC, we show that if the test distribution is structured on a Q-graph, the upper bound can be formulated as a Markov decision process (MDP) whose state being a belief on the channel state. In the case of FSCs and states that are either unifilar or have a finite memory, the MDP state simplifies to take values in a finite set. Consequently, the MDP consists of a finite number of states, actions, and disturbances. This finite nature of the MDP is of significant importance, as it ensures that dynamic programming algorithms can solve the associated Bellman equation to establish analytical upper bounds, even for channels with large alphabets. We demonstrate the simplicity of computing bounds by establishing the capacity of a broad family of Noisy Output is the State (NOST) channels as a simple closed-form analytical expression. Furthermore, we introduce novel, nearly optimal analytical upper bounds on the capacity of the Noisy Ising channel. © 1963-2012 IEEE.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

CPT:A Configurable Predictability Testbed for DNN Inference in AVs

引用

Tsinghua science and Technology 2025年第1期30卷 87-99页

作者： Liangkai Liu Yanzhi Wang Weisong Shi Department of Computer and Information Science University of DelawareNewarkDE 19713USA Department of Electrical&Computer Engineering Northeastern UniversityBostonMA 02115USA

Predictability is an essential challenge for autonomous vehicles(AVs)’*** neural networks have been widely deployed in the AV’s perception ***,it is still an open question on how to guarantee the perception predictability for AV because there are millions of deep neural networks(DNNs)model combinations and system configurations when deploying DNNs in *** paper proposes configurable predictability testbed(CPT),a configurable testbed for quantifying the predictability in AV’s perception *** provides flexible configurations of the perception pipeline on data,DNN models,fusion policy,scheduling policies,and predictability *** top of CPT,the researchers can profile and optimize the predictability issue caused by different application and system *** has been open-sourced at:https://***/Torreskai0722/CPT.

关键词： predictability deep neural network autonomous driving

来源：评论

学校读者我要写书评

暂无评论

Visually enhanced multi-secret sharing scheme with arithmetic modulo, DWT, and particle swarm optimization for Meaningful Shares

引用

Cluster Computing 2025年第5期28卷 1-28页

作者： Rawat, Arjun Singh Deshmukh, Maroti Singh, Maheep Department of Computer Science and Engineering National Institute of Technology Uttarakhand Srinagar Garhwal India Department of Computer Science Doon University Dehradun India Department of Computer Science and Engineering National Institute of Technology Delhi India

Traditional multi-secret sharing (MSS) schemes generate random shares to secure secrets, but their noisy appearance can raise suspicion. To address this, we present an advanced (n+1,n+1) MSS scheme that generates meaningful, high-quality visual shares, reducing the risk of detection by attackers. Our scheme combines arithmetic Modulo, discrete wavelet transform (DWT), and particle swarm optimization (PSO), balancing security and visual appeal. The scheme operates through two main processes: the first involves the generation of randomized shares and the embedding of watermarks in the meaningful share generation process, while the second focuses on the extraction of randomized shares or watermark images from meaningful shares and the reconstruction of the secret during the secret reconstruction process. PSO optimizes embedding factors, achieving an ideal balance between imperceptibility and visual quality, often exceeding that of the meaningful cover images. Quantitative assessments with metrics like correlation, mean squared error (MSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM) show marked improvements in both robustness and visual quality of the shared images. This solution addresses the limitations of traditional schemes, providing a secure and visually appealing approach to multi-secret image sharing. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Discrete wavelet transforms

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：