检索结果-内蒙古大学图书馆

2025 ieee/CVF Winter Conference on applications of Computer Vision, WACV 2025

作者： Shukla, Shubhi Dalui, Subhadeep Alam, Manaar Datta, Shubhajit Mondal, Arijit Mukhopadhyay, Debdeep Chakrabarti, Partha Pratim Centre for Computational and Data Sciences Iit Kharagpur India Iit Kharagpur Computer Science and Engineering Department India Center for Cyber Security New York University Abu Dhabi United Arab Emirates Iit Kharagpur Department of Artificial Intelligence India Iit Patna Computer Science and Engineering Department India

ISBN: (纸本)9798331510831

Adversarial attacks rely on transferability, where an adversarial example (AE) crafted on a surrogate classifier tends to mislead a target classifier. Recent ensemble methods demonstrate that AEs are less likely to mislead multiple classifiers in an ensemble. This paper proposes a new ensemble training using a Pairwise Adversarially Robust Loss (PARL) that by construction produces an ensemble of classifiers with diverse decision boundaries. PARL utilizes outputs and gradients of each layer with respect to network parameters in every classifier within the ensemble simultaneously. PARL is demonstrated to achieve higher robustness against black-box transfer attacks than previous ensemble methods as well as adversarial training without adversely affecting clean example accuracy. Extensive experiments using standard Resnet20, WideResnet28-10 classifiers demonstrate the robustness of PARL against state-of-the-art adversarial attacks. While maintaining similar clean accuracy and lesser training time, the proposed architecture has a 24.8% increase in robust accuracy (= 0.07) from the state-of-the art method. Code is available at: https://***/shubhishukla10/PARL © 2025 ieee.

关键词： adversarial attack decision-boundary diversity defense ensemble models

来源：评论

学校读者我要写书评

暂无评论

A Decentralized Greedy Assignment-Learning Spiking Neural Network-based Solution for A Perimeter defense Problem

A Decentralized Greedy Assignment-Learning Spiking Neural Ne...

引用

computational intelligence in security, Defence and Biometrics (CISDB), ieee symposium on

作者： P Mohammed Thousif Shridhar Velhal Suresh Sundaram Narasimhan Sundararajan Department of Aerospace Engineering Indian Institute of Science Bangalore India Department of Computer Electrical and Space Engineering Luleå University of Technology Luleå Sweden

ISBN: (数字)9798331508296

ISBN: (纸本)9798331508302

In this paper, a decentralized Greedy Assignment Learning solution framework using a Spiking neural network (de-GALS) is proposed for solving a Perimeter defense Problem (PDP). A typical PDP scenario considers defenders protecting the perimeter of a convex territory from intruders. The region between the perimeter and the sensing ranges of defenders is divided into two different layers namely a sensing layer and a capture layer respectively. These layers are further divided into multiple angular segments. The layer closest to the perimeter is termed the capture layer in which the defenders operate and capture the intruders. The next layer is the sensing layer in which the intruders’ arrivals are sensed based on the defender’s sensors. The spatiotemporal movements of the defenders and intruders are converted into spikes and given as input to a Spiking Neural Network (SNN). In the SNN, the segments in the capture layer are assigned to a defender in a decentralized fashion to capture the intruders before they enter the territory. The SNN is trained in a supervised manner where the expert assignments are generated greedily based on the location of a defender. Based on the performance studies the proposed de-GALS framework shows better performance than other existing state-of-the-art solutions for PDP with the added advantage of requiring a less computational greedy approach for generating the expert data.

关键词： Training Biometrics Uncertainty Shape Spiking neural networks Sensors Spatiotemporal phenomena security computational intelligence

来源：评论

学校读者我要写书评

暂无评论

2012 ieee symposium on computational intelligence for security and Defence applications (ieee CISDA 2012)

引用

ieee computational intelligence MAGAZINE 2013年第1期8卷 12-13页

作者： Abielmona, Rami Larus Technologies CANADA

The 2012 ieee symposium on computational intelligence for security and Defence applications (ieee CISDA 2012) was held from 11 to 13 July 2012 in the Delta Ottawa City Centre Hotel in Ottawa, Canada. This article overviews the happenings at the conference. ieee CISDA 2012 was the fifth forum (the first was held in 2007) on CI for security and defence applications. Since that inaugural year, new problems have emerged within the broad areas of security and defence that are increasingly more difficult to tackle with conventional methods, thus requiring new techniques for detecting and adapting to emerging threats. The purpose of the symposium is to present current efforts in CI (e.g., neural networks, fuzzy systems, evolutionary computation, swarm intelligence, and other emerging learning or optimization techniques) as applied to security, defence and military problems. The following topics are typically discussed at the symposium: complex adaptive systems, intelligence gathering and exploitation, modeling and simulation of military operations, network security, automatic target recognition and tracking, biometrics, as well as employment and control of autonomous vehicles.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploring Open Dimensions in Artificial Immune Systems for High-Dimensional Tasks

Exploring Open Dimensions in Artificial Immune Systems for H...

引用

computational intelligence in Artificial Life and Cooperative Intelligent Systems (ALIFE-CIS), ieee symposium on

作者： Axel Luiggi-Gørrissen Trygve Johansen Woldseth Pauline Catriona Haddow Depart. of Computer Science CRAB Lab NTNU Trondheim Norway

ISBN: (数字)9798331508371

ISBN: (纸本)9798331508388

High-dimensional (HD) data is relevant to a wide range of fields e.g. remote sensing, network security, medicine and e-commerce. Improved methods are needed to handle such HD data. Artificial Immune Systems (AIS) have shown promise across various classification tasks. However, they are challenged for applications with HD spaces that thus limit their application to such fields. The work herein proposes a novel AIS-based model, the Flexible-Dimension-AIS (FD-AIS), incorporating open dimensions into antibody recognition regions (where antigens can be discovered) so as to address this challenge. Open dimensions enable ABs to extend coverage without being constrained by fixed boundaries, thereby allowing for improved coverage with reduced computational overhead. Key contributions of the model include the introduction of flexible, adaptable recognition regions. Experimental results highlighted the positive effect of unbounded recognition regions in FD-AIS, particularly with respect to efficiency - important when working with high-dimensional data.

关键词： Antigens High dimensional data computational modeling Artificial immune systems Network security Antibodies Electronic commerce Intelligent systems Remote sensing computational intelligence

来源：评论

学校读者我要写书评

暂无评论

RSA Cryptanalysis: An Elitist Genetic Algorithm for Integer Factorization

RSA Cryptanalysis: An Elitist Genetic Algorithm for Integer ...

引用

computational intelligence in security, Defence and Biometrics Companion (CISDB Companion), ieee symposium on

作者： Aarti Kumari Utpal Garain Sanghamitra Bandyopadhyay Girish Mishra Saibal K. Pal Malay Bhattacharyya CAIML Indian Statistical Institute CVPRU Indian Statistical Institute MIU Indian Statistical Institute Scientific Analysis Group DRDO

ISBN: (数字)9798331508470

ISBN: (纸本)9798331508487

The problem of Integer Factorization deals with the disintegration of an integer into prime factors. This has direct applications in cryptanalysis, particularly in breaking the RSA. Lately, scientists have investigated the scope of employing metaheuristic approaches to perform integer factorization in bounded time. This paper improves the state-of-the-art by incorporating the concept of elitism, removal of redundancy, as well as diminishing the effect of exploitation in a genetic algorithm (GA) model. This enabled us to factor integers up to 22 decimal digits (≈ 70 bits) with a higher rate of success and quicker convergence on a standard desktop, thereby breaking the bottleneck of existing metaheuristic approaches. By design, the proposed metaheuristic approach is population based and can process the intermediate steps by means of independent calculations. Hence, it can be easily parallelized and scaled up for factorizing larger integers.

关键词： Biometrics computational modeling Metaheuristics Redundancy Wheels Minimization security Standards Genetic algorithms Convergence

来源：评论

学校读者我要写书评

暂无评论

Imbalanced malware classification: an approach based on dynamic classifier selection

Imbalanced malware classification: an approach based on dyna...

引用

computational intelligence in security, Defence and Biometrics Companion (CISDB Companion), ieee symposium on

作者： José Vinicius S. Souza Camila Barbosa Vieira George D. C. Cavalcanti Rafael M. O. Cruz Centro de Informática Universidade Federal de Pernambuco Recife PE Brazil École de Technologie Supérieure Université du Québec Montreal Canada

ISBN: (数字)9798331508470

ISBN: (纸本)9798331508487

In recent years, the rise of cyber threats has emphasized the need for robust malware detection systems, especially on mobile devices. Malware, which targets vulnerabilities in devices and user data, represents a substantial security risk. A significant challenge in malware detection is the imbalance in datasets, where most applications are benign, with only a small fraction posing a threat. This study addresses the often-overlooked issue of class imbalance in malware detection by evaluating various machine learning strategies for detecting malware in Android applications. We assess monolithic classifiers and ensemble methods, focusing on dynamic selection algorithms, which have shown superior performance compared to traditional approaches. In contrast to balancing strategies performed on the whole dataset, we propose a balancing procedure that works individually for each classifier in the pool. Our empirical analysis demonstrates that the KNOP algorithm obtained the best results using a pool of Random Forest. Additionally, an instance hardness assessment revealed that balancing reduces the difficulty of the minority class and enhances the detection of the minority class (malware). The code used for the experiments is available at https://***/jvss2/Machine-Learning-Empirical-Evaluation.

关键词： Machine learning algorithms Heuristic algorithms Focusing Malware Mobile handsets Classification algorithms security Ensemble learning Random forests computational intelligence

来源：评论

学校读者我要写书评

暂无评论

Feature Selection Methods for Intrusion Detection Systems in IoT

Feature Selection Methods for Intrusion Detection Systems in...

引用

International symposium on Applied Machine intelligence and Informatics (SAMI)

作者： Ons Saadallah Ikrame Nouar Department of Data Science and Engineering ELTE - Eotvös Loránd University Budapest Hungary

ISBN: (数字)9798350379365

ISBN: (纸本)9798350379372

IoT systems enhance daily life through smart applications but are vulnerable due to the increasing number of con-nected devices. To address security concerns, Intrusion Detection Systems (IDS) are crucial for detecting and preventing attacks. While various IDS techniques use computational datasets, many features are irrelevant, prompting the use of feature selection (FS) methods, such as wrapper and embedded techniques, to improve performance. This paper evaluates the impact of FS methods on machine learning models like Naive Bayes, Logistic Regression, Random Forest, Decision Tree, K-Nearest Neighbors, and SVM for binary and multi-class classification of network attacks. It also explores FS effects on a CNN-LSTM model using the BoTNeTIoT dataset.

关键词： Training Support vector machines Performance evaluation Logistic regression Intrusion detection Nearest neighbor methods Feature extraction security Random forests Machine intelligence

来源：评论

学校读者我要写书评

暂无评论

2012 ieee symposium on computational intelligence for security and defence applications (ieee CISDA 2012)

2012 IEEE symposium on computational intelligence for securi...

引用

作者： Abielmona, Rami Larus Technologies Canada

The 2012 ieee symposium on computational intelligence (CI) for security and defense applications (ieee CISDA 2012) was held in the Delta Ottawa City Center Hotel in Ottawa, Canada, from July 11-13, 2012. The 2012 event's motto was Creating Intelligent Surveillance where one selectively applied CI techniques to create intelligent surveillance solutions. This was demonstrated by papers on anomaly detection for explosive hazards, along with radioactive material concealment that utilized support vector machines (SVMs) and autoassociators. Other papers discussed topics such as the creation and simulation of courses of action using cloud-based planning, the detection of buried objects using features such as histogram of gradients (HOG) and an SVM classifier and the analysis of simulation and real-world complex systems data with heterogeneous structures to improve the surveillance of communication networks. ieee CISDA 2012 also offered four special sessions, including one that provided an interdisciplinary forum.

关键词： Radioactive materials

来源：评论

学校读者我要写书评

暂无评论

Benchmarking Cryptographic Protocols for (IoT) Malware Defence

Benchmarking Cryptographic Protocols for (IoT) Malware Defen...

引用

International symposium on Applied Machine intelligence and Informatics (SAMI)

作者： Khalil Ben Khaled Dönìz Borsos Jozsef Katona Kandó Kalmán Faculty of Electrical Engineering Institute of Microelectronics and Technology Obuda University Budapest Hungary Kándó Kálmán Faculty of Electrical Engineering Institute of Electronics and Communication Systems Obuda University Budapest Hungary Department of Software Development and Application Institute of Computer Engineering University of Dunaujvaros Dunaujvaros Hungary GAMF Faculty of Engineering and Computer Science John Von Neumann University Kecskemet Hungary Department of Applied Quantitative Methods Faculty of Finance and Accountancy Budapest Business University Budapest Hungary

ISBN: (数字)9798350379365

ISBN: (纸本)9798350379372

This study presents a comprehensive benchmarking analysis of cryptographic protocols for Internet of Things (IoT) malware defense. The framework was specifically tailored to evaluate cryptographic protocols such as AES-128, AES-256, ChaCha20, RSA (1024-bit, 2048-bit, and 4096-bit), SHA256, SHA512, and HMAC-SHA256 were tested in both standalone and emulated environments. The primary objective was to evaluate the performance and resource consumption of these protocols, focusing on their encryption and hashing efficiency. Key innovations include the comparative analysis of resource consumption and performance efficiency across diverse cryptographic operations, under both real-world and emulated conditions. By identifying protocols like ChaCha20 for high efficiency and minimal resource usage, and RSA 4096-bit for enhanced security at higher computational costs, this study provides actionable insights into the trade-offs between security and performance. These findings offer a foundational reference for selecting optimized cryptographic protocols, advancing IoT malware defense strategies through informed decision-making.

关键词： Performance evaluation Technological innovation Benchmark testing Malware Encryption Internet of Things Reliability Informatics Machine intelligence Cryptographic protocols

来源：评论

学校读者我要写书评

暂无评论

Enhancing Intrusion Detection Systems with representation methods: A comparative study

Enhancing Intrusion Detection Systems with representation me...

引用

computational intelligence in security, Defence and Biometrics (CISDB), ieee symposium on

作者： Bruno H. Meyer Aurora T. R. Pozo Michele Nogueira Wagner M. Nunan Zola Department of Informatics Federal University of Paraná Brazil Department of Computer Science Federal University of Minas Gerais Brazil

ISBN: (数字)9798331508296

ISBN: (纸本)9798331508302

This paper presents a comparative analysis of three data representation methods for improving Intrusion Detection Systems (IDS). The methods compared are autoencoders, Generative Adversarial Networks (GANs), and contrastive learning. Additionally, a baseline approach using raw input data is evaluated. The study is conducted on three well-known IDS datasets: NSL-KDD, Ton-IoT, and Bot-IoT, each with distinct characteristics. Our results demonstrate that representational methods significantly enhance classification performance, particularly when ample unlabeled data is available. Among the methods, GANs achieved the highest f1-score improvements in the Ton-IoT dataset, while contrastive learning excelled in the Bot-IoT dataset. The experiments also reveal that the choice of classifier impacts performance, with Random Forest performing best on raw data and Multi-Layer Perceptrons (MLP) excelling with transformed data. The study highlights the importance of selecting appropriate representation learning techniques and classifiers based on dataset characteristics. It emphasizes the potential of unsupervised learning methods to utilize large volumes of unlabeled data, a common scenario in real-world cybersecurity applications. The findings provide a foundation for future research in leveraging unsupervised learning for IDS and other cybersecurity challenges.

关键词： Representation learning Autoencoders Intrusion detection Data visualization Contrastive learning NSL-KDD Generative adversarial networks Computer security Unsupervised learning Random forests

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：