检索结果-内蒙古大学图书馆

Make Shuffling Great Again: A Side-Channel-Resistant Fisher-Yates Algorithm for Protecting Neural Networks

IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS 2025年

作者： Puskac, Leonard Benovic, Marek Breier, Jakub Hou, Xiaolu Slovak Univ Technol Bratislava Fac Informat & Informat Technol Bratislava 84216 Slovakia TTControl GmbH A-1040 Vienna Austria

Neural network (NN) models implemented in embedded devices have been shown to be susceptible to side-channel attacks (SCAs), allowing recovery of proprietary model parameters, such as weights and biases. There are already available countermeasure methods currently used for protecting cryptographic implementations that can be tailored to protect embedded NN models. Shuffling, a hiding-based countermeasure that randomly shuffles the order of computations, was shown to be vulnerable to SCA when the Fisher-Yates algorithm is used. In this article, we propose a design of an SCA-secure version of the Fisher-Yates algorithm. By integrating the masking technique for modular reduction and Blakely's method for modular multiplication, we effectively remove the vulnerability in the division operation that led to side-channel leakage in the original version of the algorithm. We experimentally evaluate that the countermeasure is effective against SCA by implementing a correlation power analysis (CPA) attack on an embedded NN model implemented on ARM Cortex-M4. Compared to the original proposal, the memory overhead is 2x the biggest layer of the network, while the time overhead varies from 4% to 0.49% for a layer with 100 and 1000 neurons, respectively.

关键词： Artificial neural networks Neurons Hardware Proposals Computational modeling software Timing software algorithms Protection Power demand Countermeasures Fisher-Yates hiding neural networks (NNs) shuffling side-channel attacks (SCAs)

来源：评论

学校读者我要写书评

暂无评论

A software Implementation and Research of Ensemble algorithms for Solving the Different Types of Data Classification Problem

A Software Implementation and Research of Ensemble Algorithm...

引用

IEEE Region 8 International Conference on Computational Technologies in Electrical and Electronics Engineering, SIBIRCON

作者： Nikita A. Zelenchuk Olga K. Alsova Dept. of Computer Science Novosibirsk State Technical University Novosibirsk Russia

ISBN: (纸本)9781665464819

Many algorithms for solving classification problems have been studied and developed at the moment, but the problem of improving the quality (accuracy, stability) of classification solutions remains relevant. This problem can be solved by developing ensemble classification algorithms, as well as by performing a qualitative primary analysis and processing of the initial data. This research reports the results of research and implementation of ensemble algorithms, which are being developed in software to solve the problem of classification of different types of data. Homogeneous and heterogeneous ensemble classification models were implemented by using R language. Several classifiers were built on the basis of real agricultural data using the developed models. Comparative testing showed the advantages of using ensemble algorithms compared to single classifiers (increased classification accuracy, reduced logarithmic losses). According to the results of the study, the conclusion about the prospects of development and application of ensemble models to solve the problem of classification and software to solve the problems of analysis and data processing in various fields using modern technologies and software tools.

关键词： R language software algorithms Prediction algorithms Data processing Data models Stability analysis Classification algorithms

来源：评论

学校读者我要写书评

暂无评论

An Empirical Study on software Aging of Long-Running Object Detection algorithms

An Empirical Study on Software Aging of Long-Running Object ...

引用

IEEE International Conference on software Quality, Reliability and Security (QRS)

作者： Roberto Pietrantuono Domenico Cotroneo Ermeson Andrade Fumio Machida University of Naples Federico II Naples Italy Federal Rural University of Pernambuco Recife Brazil University of Tsukuba Tsukuba Japan

Efficient and effective object detection is a key problem in Computer Vision. Numerous object detection algorithms have been developed, whose aim is to achieve two conflicting goals, namely accuracy and efficiency, while being executed in real-time with high robustness. Many of these algorithms must run for an extended period of time, i.e., in video surveillance or in self-driving cars – a working condition that make them subject to the risk of software *** this work, we focus on evaluating several object detection algorithms to understand if and to what extent they are affected by software aging. A measurement-based aging approach was adopted, with a series of long-running tests and subsequent data analysis. The results report significant trends of performance degradation, sometimes leading to aging-related failures, as well as memory consumption trends, which turned out to be the main issue across all the experiments.

关键词： Statistical analysis software algorithms Memory management Computer bugs Object detection Aging Market research

来源：评论

学校读者我要写书评

暂无评论

Applications of the Partial Fraction Decomposition to Pole Variation Rate for Root Locus Construction

引用

IEEE CONTROL SYSTEMS LETTERS 2024年 8卷 2925-2930页

作者： Tebaldi, Davide Zanasi, Roberto Univ Modena & Reggio Emilia Dept Engn Enzo Ferrari I-41125 Modena Italy

The applications of the partial fraction decomposition in control and systems engineering are several. In this letter, we use the partial fraction decomposition to address the pole variation rate problem, namely to study the rate of variation of the system poles when the control parameter changes and when the system is subject to variations of its own parameters, which has led to the proposal of a new algorithm for the construction of the root locus. The new algorithm is proven to be much more efficient in terms of execution time than the dedicated MATLAB function, while providing the same output results.

关键词： Mathematical models MATLAB Vectors Approximation algorithms Proposals Polynomials Poles and zeros Electric motors Control theory software algorithms Partial fraction decomposition pole variation rate root locus root locus construction

来源：评论

学校读者我要写书评

暂无评论

A Novel 3D-to-3D Diffeomorphic Registration Algorithm With Applications to Left Ventricle Segmentation in MR and Ultrasound Sequences

引用

IEEE ACCESS 2023年 11卷 3144-3159页

作者： Krishnaswamy, Deepa Noga, Michelle Becher, Harald Boulanger, Pierre Punithakumar, Kumaradevan Univ Alberta Dept Radiol & Diagnost Imaging Edmonton AB T6G 2B7 Canada Mazankowski Alberta Heart Inst Servier Virtual Cardiac Ctr Edmonton AB T6G 2B7 Canada Mazankowski Alberta Heart Inst Edmonton AB T6G 2B7 Canada Univ Alberta Dept Comp Sci Edmonton AB T6G 2E8 Canada

Left ventricular segmentation is a difficult and time-consuming task performed by clinicians, requiring the use of manual contours. We propose a novel three-dimensional diffeomorphic registration algorithm for endocardial segmentation of the left ventricle in magnetic resonance images and ultrasound temporal sequences. The proposed diffeomorphic registration method computes a voxel-to-voxel correspondence in three-dimensional space and is parameterized by one radial and three curl components to emulate the cardiac deformations. In addition, the proposed method allows for enforcing constraints to control the amount of deformation. The method was evaluated on 521 temporal frames from 20 patients from the Automated Cardiac Diagnosis Challenge magnetic resonance imaging dataset and 213 frames from 10 patients undergoing ultrasound scans from the Mazankowski Alberta Heart Institute. The algorithm was compared against six other registration methods, the Symmetric Normalization algorithm from the Dipy package, two variants of the Demons algorithm from the Insight Toolkit software package, two versions of RealTiTracker, and Elastix. The proposed method yielded overall Dice scores of 98.10 (0.90)% for the MRI dataset and 92.90 (2.42)% for the ultrasound dataset. The robustness of the algorithm is demonstrated by the high performance on multiple imaging modalities and various patient abnormalities.

关键词： Cardiology Three-dimensional displays Magnetic resonance imaging Heart Ultrasonic imaging Image segmentation software algorithms Image registration image segmentation left ventricle MRI ultrasound

来源：评论

学校读者我要写书评

暂无评论

Graphical and Numerical Study of a Newly Developed Root-Finding Algorithm and Its Engineering Applications

引用

IEEE ACCESS 2023年 11卷 2375-2383页

作者： Naseem, Amir Rehman, M. A. Qureshi, Sania Ide, Nasr Al Din Univ Management & Technol Dept Math Lahore 54770 Pakistan Mehran Univ Engn & Technol Dept Basic Sci & Related Studies Jamshoro 76062 Pakistan Near East Univ Dept Math TR-99138 Mersin Turkiye Aleppo Univ Fac Sci Dept Math Aleppo Syria

The primary objective of this paper is to develop a new method for root-finding by combining forward and finite-difference techniques in order to provide an efficient, derivative-free algorithm with a lower processing cost per iteration. This will be accomplished by combining forward and finite-difference techniques. We also detail the convergence criterion that was devised for the root-finding approach, and we show that the method that was recommended is quintic-order convergent. We addressed a few engineering issues in order to illustrate the validity and application of the developed root-finding algorithm. The quantitative results justified the constructed root-finding algorithm's robust performance in comparison to other quintic-order methods that can be found in the literature. For the graphical analysis, we make use of the newly discovered method to plot some novel polynomiographs that are attractive to the eye, and then we evaluate these new plots in relation to previously established quintic-order root-finding strategies. The graphic analysis demonstrates that the newly created method for root-finding has better convergence with the larger area than the other comparable methods do.

关键词： Mathematical models Iterative methods Convergence Approximation algorithms Costs Taylor series software algorithms Computational algorithms convergence-order non-linearity Halley's scheme dynamical aspects

来源：评论

学校读者我要写书评

暂无评论

Hyperparameter Optimization for AST Differencing

引用

IEEE TRANSACTIONS ON software ENGINEERING 2023年第10期49卷 4814-4828页

作者： Martinez, Matias Falleri, Jean-Remy Monperrus, Martin Univ Politecn Cataluna Barcelona 08034 Spain Univ Bor deaux CNRS Bordeaux INP LaBRIUMR 5800 F-33400 Deaux France Inst Univ France Paris France KTH Royal Inst Technol S-11428 Stockholm Sweden

Computing the differences between two versions of the same program is an essential task for software development and software evolution research. AST differencing is the most advanced way of doing so, and an active research area. Yet, AST differencing algorithms rely on configuration parameters that may have a strong impact on their effectiveness. In this paper, we present a novel approach named DAT (Diff Auto Tuning) for hyperparameter optimization of AST differencing. We thoroughly state the problem of hyper-configuration for AST differencing. We evaluate our data-driven approach DAT to optimize the edit-scripts generated by the state-of-the-art AST differencing algorithm named GumTree in different scenarios. DAT is able to find a new configuration for GumTree that improves the edit-scripts in 21.8% of the evaluated cases.

关键词： Training software algorithms Computer bugs Syntactics Maintenance engineering Hyperparameter optimization software software evolution Tree differencing Abstract Syntax Trees (AST) hyperparameter optimization, edit-script

来源：评论

学校读者我要写书评

暂无评论

Don't Lie to Me: Avoiding Malicious Explanations With STEALTH

引用

IEEE software 2023年第3期40卷 43-53页

作者： Alvarez, Lauren Menzies, Tim North Carolina State Univ Raleigh NC 27606 USA North Carolina State Univ Comp Sci Raleigh NC 27606 USA

STEALTH is a method for using some artificial intelligence-generated models without suffering from malicious attacks or associated unfairness issues. STEALTH asks so few queries (one per data cluster) that malicious a... 详细信息

关键词： software algorithms Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Air-Ground Collaborative Edge Intelligence for Future Generation Networks

引用

IEEE NETWORK 2023年第2期37卷 118-125页

作者： Tang, Jianhang Nie, Jiangtian Zhang, Yang Duan, Yiqun Xiong, Zehui Niyato, Dusit Guizhou Univ Guiyang Peoples R China Nanyang Technol Univ Singapore Singapore Nanjing Univ Aeronaut & Astronaut Nanjing Peoples R China Univ Technol Sydney Sydney NSW Australia Singapore Univ Technol & Design Singapore Singapore

The air-ground integrated mobile edge computing (MEC) is expected to fulfill the ever-growing resource demands of artificial intelligence (AI)-enabled applications in sixth-generation (6G) wireless networks, ranging from computer vision to natural language processing. Nevertheless, it is still challenging to offer high-quality AI services by fully exploring the advantages of terrestrial MEC networks and unmanned aerial vehicles (UAVs), especially as they have to share resources collaboratively. To meet this challenge, we propose a novel framework termed air-ground collaborative edge intelligence (EI), featuring the collaboration of terrestrial and aerial resources as a potential solution to enable persistent and ubiquitous Al services. By installing various modules on UAVs, three distinct air-ground collaboration schemes are considered and discussed, where these UAVs can provide communication, computation, and energy resources in different use cases. Next, we elaborate on two potential applications and some open research issues for the proposed airground collaborative EI framework. Specifically, we develop a novel machine learning model caching approach, where a popular deep neural network (DNN) model is cached on proper terrestrial edge devices and UAVs to relieve network congestion. Finally, we provide extensive simulation results to demonstrate that the proposed air-ground collaborative caching algorithm can improve inference efficiency dramatically.

关键词： Machine learning algorithms Atmospheric modeling Computational modeling Wireless networks software algorithms Collaboration Machine learning

来源：评论

学校读者我要写书评

暂无评论

Analyzing the Behavioral Artifacts for Malicious software Detection Using Machine Learning algorithms

Analyzing the Behavioral Artifacts for Malicious Software De...

引用

Advances in Computing, Communication Control and Networking (ICAC3N), International Conference on

作者： Nikkam Chandini C. Kiran Mai Department of Computer Science and Engineering VNRVJIET Hyderabad India

Malicious software intentionally affects your computer system. Malware is analyzed using static or dynamic analysis techniques. These techniques are used to classify and predict unique patterns for correctly identifying malware. Over the last decade, many malware detection techniques have been proposed using a variety of techniques. In most cases, network-based attacks such as DDoS attacks are known network security risks. These causes usually cause network devices to overflow with more requests than they can handle, preventing the server from responding to legitimate requests. The biggest security risk is related to software. software attacks can exploit the entire system, steal information, modify data, deny services, and endanger or damage devices. In particular, I am interested in dynamic analysis for developing malware detection systems that utilize machine learning techniques. This document proposes behavior-based malware detection techniques. To develop this model, we arrange a dynamic analysis approach and run malware samples using classifiers. To extracted for the malware detection system. To develop this technique, we create a dynamic analysis approach and run malware using classification algorithms. Computer programs connected to Internet are increasingly needed to analyze and prevent malware. The software exploits system vulnerabilities to unknowingly steal sensitive information and secretly send it to an attacker-controlled remote server. This approach uses the ML classifier to detect malware and identify the type of malware attacked in the dataset.

关键词： Machine learning algorithms Computer hacking Heuristic algorithms software algorithms Prediction algorithms Malware Classification algorithms

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：