检索结果-内蒙古大学图书馆

Optimization of Binary Randomized Response Based on Lanke Privacy and Utility Analysis

Chinese Journal of Electronics 2025年第1期34卷 242-256页

作者： Yihui Zhou Wenli Wang Jun Yan Zhenqiang Wu Laifeng Lu School of Computer Science Shaanxi Normal University School of Mathematics and Statistics Shaanxi Normal University School of Mathematics and Computer Applications Shangluo College

Currently, it has become a consensus to enhance privacy protection. Randomized response(RR) technique, as the mainstream perturbation mechanism for local differential privacy, has been widely studied. However,most of the research in literature managed to modify existing RR schemes and propose new mechanisms with better privacy protection and utility, which are illustrated only by numerical experiments. We study the properties of generalized binary randomized response mechanisms from the perspectives of Lanke privacy and utility. The mathematical expressions of privacy and utility for the binary RR mechanism are given respectively. Moreover, the comparison principle for privacy and utility of any two mechanisms is proved. Finally, the optimization problem of the binary RR mechanism is discussed. Our work is based on a rigorous mathematical proof of privacy and utility for the general binary RR mechanism, and numerical verification illustrates the correctness of the conclusions. It can provide theoretical support for the design of binary RR mechanism and can be applied in data collection, analysis and publishing.

关键词： Privacy Differential privacy Publishing Perturbation methods Data collection Protection Optimization

来源：评论

学校读者我要写书评

暂无评论

Free-text keystroke authentication using transformers: a comparative study of architectures and loss functions

引用

Soft Computing 2025年第2期29卷 1259-1272页

作者： Momeni, Saleh BabaAli, Bagher School of Mathematics Statistics and Computer Science University of Tehran Tehran Iran

Keystroke biometrics is a promising approach for user identification and verification, leveraging the unique patterns in individuals’ typing behavior. In this paper, we propose a Transformer-based network that employs self-attention to extract informative features from keystroke sequences, surpassing the performance of traditional Recurrent Neural Networks. We explore two distinct architectures, namely bi-encoder and cross-encoder, and compare their effectiveness in keystroke authentication. Furthermore, we investigate different loss functions, including triplet, batch-all triplet, and WDCL loss, along with various distance metrics such as Euclidean, Manhattan, and cosine distances. These experiments allow us to optimize the training process and enhance the performance of our model. To evaluate our proposed model, we employ the Aalto desktop keystroke dataset. The results demonstrate that the bi-encoder architecture with batch-all triplet loss and cosine distance achieves the best performance, yielding an exceptional Equal Error Rate of 0.0186%. Furthermore, alternative algorithms for calculating similarity scores are explored to enhance accuracy. Notably, the utilization of a one-class Support Vector Machine reduces the Equal Error Rate to an impressive 0.0163%. The outcomes of this study indicate that our model surpasses the previous state-of-the-art in free-text keystroke authentication. These findings contribute to advancing the field of keystroke authentication and offer practical implications for secure user verification systems. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Attack Detection Using Item Vector Shift in Matrix Factorisation Recommenders

引用

ACM Transactions on Privacy and Security 2025年第2期28卷 1-22页

作者： Shams, Sulthana Leith, Douglas School of Computer Science and Statistics Trinity College Dublin Dublin Ireland

This article proposes a novel method for detecting shilling attacks in Matrix Factorization (MF)-based Recommender Systems (RSs), in which attackers use false user-item feedback to promote a specific item. Unlike existing methods that use either supervised learning to distinguish between attack and genuine profiles or analyze target item rating distributions to detect false ratings, our method uses an unsupervised technique to detect false ratings by examining shifts in item preference vectors that exploit rating deviations and user characteristics, making it a promising new direction. The experimental results demonstrate the effectiveness of our approach in various attack scenarios, including those involving obfuscation techniques. © 2025 Copyright held by the owner/author(s).

关键词： Matrix factorization

来源：评论

学校读者我要写书评

暂无评论

Identifying the Intents Behind Website Visits by Employing Unsupervised Machine Learning Models

引用

Annals of Data science 2025年第1期12卷 413-437页

作者： Soobramoney, Judah Chifurira, Retius Zewotir, Temesgen Chinhamu, Knowledge School of Mathematics Statistics and Computer Science University of KwaZulu-Natal Durban South Africa

With digitisation globally on the rise, corporates are compelled to better understand the usage of their websites. In doing so, corporates will be empowered to better understand consumers, and make necessary adjustments to ultimately improve the corporate’s stance in the competitive global landscape of this modern age. However, the online website visit data has proven to be highly complex, big in data volume, and highly transactional with users expressing unique behaviours. Thus, extracting insight can be a complex problem to solve. This study aimed to employ unsupervised machine learning models to identify the intentions behind the visits on the observed website. The data studied was sourced from the Google Analytics tracking tool that was deployed on a corporate informative website. The study employed a k-means, hierarchical and dbscan unsupervised machine learning models to understand the intents behind visitors on the studied website. All three models detected five major intents that were expressed within the observed data. The intents identified were labelled as "accidentals", "drop-offs", "engrossed", "get-in-touch" and "seekers". On the observed data, all three unsupervised machine learning methods have performed well. However, in the context of the study, which investigated the intents that drove online visits, the hierarchical clustering method yielded superior results by maintaining the best balance between cluster homogeneity (stronger silhouette coefficients) and cluster size. © The Author(s) 2025.

关键词： Cubic clustering criteria Dbscan Dendrogram Google analytics tracking Hierarchical clustering K-means Online website visits Silhouette’s coefficient Unsupervised machine learning

来源：评论

学校读者我要写书评

暂无评论

AN INERTIAL SCHEME FOR APPROXIMATING THE SOLUTIONS OF PSEUDO-MONOTONE EQUILIBRIUM AND COMMON FIXED POINT PROBLEMS IN BANACH SPACES

Applied Set-Valued Analysis and Optimization

引用

Applied Set-Valued Analysis and Optimization 2025年第2期7卷 223-249页

作者： Mewomo, Oluwatosin Temitope Godwin, Emeka Chigamezu Alakoya, Timilehin Opeyemi School of Mathematics Statistics and Computer Science University of KwaZulu-Natal Durban South Africa

In this paper, we propose a new iterative method for approximating a common solution of pseudo-monotone equilibrium problems and common fixed point problems of Bregman quasi-nonexpansive mappings in a p-uniformly convex and uniformly smooth Banach space. Under some weaker conditions, we obtain a strong convergence result. A variational inequality problem is also considered. We present numerical examples to illustrate and support our main convergence theorem. © 2025 Applied Set-Valued Analysis and Optimization.

关键词： Banach spaces

来源：评论

学校读者我要写书评

暂无评论

Finite Non-abelian p-groups All of Whose Non-abelian Proper Subgroups have Centers of the Same Order

引用

Acta Mathematica Sinica,English Series 2025年第4期41卷 1247-1268页

作者： Dandan Zhang Haipeng Qu Yanfeng Luo School of Data Sciences Zhejiang University of Finance & Economics School of Mathematics and Statistics Lanzhou University School of Mathematics and Computer Science Shanxi Normal University

In this paper,we classify the finite non-abelian p-groups all of whose non-abelian proper subgroups have centers of the same order.

关键词： 20D15 At-groups Finite p-groups the center of groups the order of centers

来源：评论

学校读者我要写书评

暂无评论

Unsupervised multiplex graph diffusion networks with multi-level canonical correlation analysis for multiplex graph representation learning

引用

science China(Information sciences) 2025年第3期68卷 66-79页

作者： Sichao FU Qinmu PENG Yange HE Baokun DU Bin ZOU Xiao-Yuan JING Xinge YOU School of Electronic Information and Communications Huazhong University of Science and Technology Platform Operation and Marketing Center JD Retail Faculty of Mathematics and Statistics Hubei Key Laboratory of Applied Mathematics Hubei University School of Computer Science Wuhan University

In recent years, unsupervised multiplex graph representation learning(UMGRL) has received increasing research interest, which aims to learn discriminative node features from the multiplex graphs supervised by data without the guidance of labels. Although these designed UMGRL methods have obtained great success in various graph-related tasks, most existing UMGRL models still have the following issues: highly depending on complex self-supervised strategies(i.e., data augmentation,pretext tasks, and negative pairs sampling), restricted receptive fields, and only aggregating low-frequency information between nodes. In this paper, we propose a simple unsupervised multiplex graph diffusion network(UMGDN) with the aid of multi-level canonical correlation analysis to solve the above issues. Specifically, we first decouple the feature transform and propagation processes of the graph convolution layer to further improve the generalization of the learnable parameters. And then, we propose adaptive diffusion propagation to capture long-range dependency relationships between nodes, not the local neighborhood interactions. Finally, a multi-level canonical correlation analysis loss on both the feature transform and propagation processes is proposed to maximize the correlation of the same node features from multiple graphs for guiding model optimization. Compared to the existing UMGRL models, our proposed UMGDN does not need to introduce any data augmentation, negative pairs sampling techniques, complex pretext tasks, and also adaptively aggregates the optimal frequency information between nodes to generate more robust node embeddings. Extensive experiments on four popular datasets and two graph-related tasks demonstrate the effectiveness of the proposed method.

关键词： unsupervised multiplex graph representation learning graph neural networks node classification node clustering

来源：评论

学校读者我要写书评

暂无评论

Do Not Skip over the Offline: Verifiable Silent Preprocessing from Small Security Hardware

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 4860-4873页

作者： Dong, Wentao Xu, Lei Zheng, Leqian Duan, Huayi Wang, Cong Wang, Qian Nanjing University of Science and Technology School of Mathematics and Statistics China ETH Zurich Department of Computer Science Switzerland City University of Hong Kong Department of Computer Science Hong Kong SAR Hong Kong Wuhan University School of Cyber Science and Engineering China

Multi-party computation (MPC) has garnered growing research and industry attention, with many protocols adhering to the preprocessing model to prioritize online performance via offline-generated, data-independent correlated randomness (or correlation for short). However, existing studies have predominantly focused on the online phase, and the equally critical offline phase remains overlooked. This oversight can lead to significant yet unaddressed challenges in both security and efficiency. To address these challenges, we start from the pseudorandom correlation generator (PCG) notion by Boyle et al. (CRYPTO'19, FOCS'20) and propose a programmable, verifiable, and efficient PCG construction using dedicated, small security hardware. Our technique, termed HPCG, bridges cryptography and hardware worlds for both strong security and high performance. It enables parties to generate virtually unbounded target correlations and verify their well-formedness at a low cost. Compared to existing cryptographic PCGs, HPCG restrainedly relies on minimal hardware trust, while advancing: 1) support for versatile correlations;2) well-formedness enforcement for any correlation regardless of its form;and 3) reduced offline costs further. In contrast to prior hardware-based MPC solutions, HPCG 1) minimizes the trusted computing base;2) eliminates single-vendor trust assumptions;and 3) avoids hosting sensitive user inputs directly on security chips. Furthermore, although HPCG operates solely in the offline phase and requires no direct access to user inputs, it enables performant online MPC protocols by efficiently generating some new, highly customized correlations. Our results demonstrate that HPCG outperforms prior preprocessing techniques in performance, security, simplicity, or a combination thereof. © 2025 IEEE.

关键词： Digital storage

来源：评论

学校读者我要写书评

暂无评论

SEIR Mathematical Model for Influenza-Corona Co-Infection with Treatment and Hospitalization Compartments and Optimal Control Strategies

引用

computer Modeling in Engineering & sciences 2025年第2期142卷 1899-1931页

作者： Muhammad Imran Brett McKinney Azhar Iqbal Kashif Butt Tandy School of Computer Science University of TulsaTulsaOK 74104USA Department of Mathematics and Statistics College of ScienceKing Faisal UniversityAl-Ahsa31982Saudi Arabia

The co-infection of corona and influenza viruses has emerged as a significant threat to global public health due to their shared modes of transmission and overlapping clinical *** article presents a novel mathematical model that addresses the dynamics of this co-infection by extending the SEIR(Susceptible-Exposed-Infectious-Recovered)framework to incorporate treatment and hospitalization *** population is divided into eight compartments,with infectious individuals further categorized into influenza infectious,corona infectious,and co-infection *** proposed mathematical model is constrained to adhere to fundamental epidemiological properties,such as non-negativity and boundedness within a feasible ***,the model is demonstrated to be well-posed with a unique *** points,including the disease-free and endemic equilibria,are identified,and various properties related to these equilibrium points,such as the basic reproduction number,are *** and global sensitivity analyses are performed to identify the parameters that highly influence disease dynamics and the reproduction *** the most influential parameters is crucial for understanding their impact on the co-infection’s spread and ***,an optimal control problem is defined to minimize disease transmission and to control strategy *** purpose of our study is to identify the most effective(optimal)control strategies for mitigating the spread of the co-infection with minimum cost of the *** results illustrate the effectiveness of the implemented control strategies in managing the co-infection’s impact on the population’s *** mathematical modeling and control strategy framework provides valuable tools for understanding and combating the dual threat of corona and influenza co-infection,helping public health authorities and policymakers make informed decisions in the face of these intertwined epidemics.

关键词： Influenza-corona co-infection stability analysis sensitivity analysis treatment self-precaution optimal control

来源：评论

学校读者我要写书评

暂无评论

Stacking Factorizing Partitioned Expressions in Hybrid Bayesian Network Models

引用

ACM Transactions on Knowledge Discovery from Data 2025年第3期19卷 1-28页

作者： Lin, Peng Neil, Martin Fenton, Norman School of Statistics Capital University of Economics and Business Beijing China School of Electronic Engineering and Computer Science Queen Mary University of London London United Kingdom School of Electronic Engineering and Computer Science Queen Mary University of London London United Kingdom

Hybrid Bayesian networks (HBN) contain complex conditional probability distributions (CPD) specified as partitioned expressions over discrete and continuous variables. The size of these CPDs grows exponentially with the number of parent nodes, and when using discrete inference methods, it results in significant execution time and space inefficiency. To reduce the CPD size, a binary factorization (BF) algorithm can be used to decompose the statistical or arithmetic functions in the CPD by factorizing the number of connected parent nodes into sets of size two. However, the BF algorithm was not designed to handle partitioned expressions. Therefore, we propose a new stacking factorization (SF) algorithm to decompose partitioned expressions. The SF algorithm creates intermediate nodes to incrementally reconstruct the conditional densities in the original partitioned expression, ensuring that no more than two continuous parent nodes are connected to each child node in the resulting HBN. It generally applies to both discrete and continuous child nodes with complex partitioned expressions. When we combine SF with a dynamic discretization (DD) inference algorithm, we achieve a significant improvement in inference efficiency. Experimental results demonstrate that the combination of SF and DD can effectively manage HBNs with complex CPDs that may challenge other algorithms, which also outperform competing inference algorithms in accuracy. © 2025 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Bayesian networks

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：