检索结果-内蒙古大学图书馆

LEGF-DST:LLMs-Enhanced Graph-Fusion Dual-Stream Transformer for Fine-Grained Chinese Malicious SMS Detection

computers, Materials & Continua 2025年第2期82卷 1901-1924页

作者： Xin Tong Jingya Wang Ying Yang Tian Peng Hanming Zhai Guangming Ling School of Information and Cybersecurity People’s Public Security University of ChinaBeijing100038China Cyber Investigation Technology Research and Development Center The Third Research Institute of the Ministry of Public SecurityShanghai201204China Department of Cybersecurity Defense Beijing Police CollegeBeijing102202China School of Computer Science Henan Institute of EngineeringZhengzhou451191China

With the widespread use of SMS(Short Message Service),the proliferation of malicious SMS has emerged as a pressing societal *** deep learning-based text classifiers offer promise,they often exhibit suboptimal performance in fine-grained detection tasks,primarily due to imbalanced datasets and insufficient model representation *** address this challenge,this paper proposes an LLMs-enhanced graph fusion dual-stream Transformer model for fine-grained Chinese malicious SMS *** the data processing stage,Large Language Models(LLMs)are employed for data augmentation,mitigating dataset *** the data input stage,both word-level and character-level features are utilized as model inputs,enhancing the richness of features and preventing information loss.A dual-stream Transformer serves as the backbone network in the learning representation stage,complemented by a graph-based feature fusion *** the output stage,both supervised classification cross-entropy loss and supervised contrastive learning loss are used as multi-task optimization objectives,further enhancing the model’s feature *** results demonstrate that the proposed method significantly outperforms baselines on a publicly available Chinese malicious SMS dataset.

关键词： Transformers malicious SMS multi-task learning large language models

来源：评论

学校读者我要写书评

暂无评论

cyber Threat Intelligence (CTI): An Analysis on the Use of Artificial Intelligence and Machine Learning to Identify cyber Hazards

Cyber Threat Intelligence (CTI): An Analysis on the Use of A...

引用

International Conference on Recent Developments in cyber security, ReDCySec 2023

作者： Kant, Neelima Amrita Department of Computer Science and Engineering School of Engineering and Technology Sharda University Uttar Pradesh Greater Noida201306 India Department of Computer Science and Engineering School of Engineering and Technology Center for Cyber Security and Cryptology Sharda University Uttar Pradesh Greater Noida201306 India

ISBN: (纸本)9789819998104

Data transport volume and scope on networks are growing daily due to the quick advancements in network technology. It is challenging for cybersecurity specialists to keep track of every action taking place on the network because of the constantly growing density of networks. This circumstance has led to an increase in the complexity and intensity of threats and attacks. It is harder to detect and identify irregularities in network activities because of frequent and sophisticated cyberattacks. A well-crafted cybersecurity strategy now includes cyber threat intelligence (CTI), which is a crucial foundation. Automating the detection of cyberattacks as well as speedy attack type analysis and predication are all made possible by machine learning (ML), which offers a number of tools and techniques. The strategies for using machine learning (ML) to identify assaults are discussed in this article. Threat intelligence can help security teams defend against a constantly evolving threat environment before, during, and after an attack if used properly. By analyzing attackers and comprehending their tactics and goals, groups may create cyber defenses that are more effective, delicate, and resilient. However, due to two significant flaws, its usefulness is still in question. First, current methods are unable to detect unknown Indicator of Compromise (IoC), and second, they are unable to automatically produce categorized CTIs which renders CTI sharing. As a result, the objective of this paper is to present a complete analysis of cyber threat identification using intelligent techniques. Additionally, we covered the issues and solutions related to machine learning applications utilized in network assaults. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Approach to Spectrum Sharing Process for Coexistence Between Cellular and Wi-Fi Technologies 6

An Intelligent Approach to Spectrum Sharing Process for Coex...

引用

6th International Conference on Electrical engineering and Information and Communication technology, ICEEICT 2024

作者： Nasim, Gazi Md Hossen, Monir Hossain, Moinul Khulna University of Engineering & Technology Department of Electronics and Communication Engineering Khulna Bangladesh George Mason University Department of Cyber Security Engineering United States

ISBN: (纸本)9798350385779

It is anticipated that New-Radio Unlicensed (NR-U) technology will offload its traffic to unlicensed spectrum, such as the 5 GHz band. Nevertheless, NR-U transmissions must coexist with the current Wi-Fi networks in the 5 GHz band. The majority of current coexistence plans take into account coordinated Wi-Fi and NR-U networks, where a centralized coordinator communicates the traffic demands of the co-located networks. However, using this technique to estimate Wi-Fi traffic increases the complexity, traffic overhead, and response times. Furthermore, the conventional Listen-Before- Talk (LBT) process of NR-U is unable to ensure enough fairness of the system. Therefore, an efficient coexistence mechanism must be required to improve the fairness. For this purpose, this paper presents an intelligent LBT based coexistence mechanism to maintain the fairness using Q-learning. In this proposed scheme, the NR-U base station optimizes its maximum channel occupancy time with respect to Wi-Fi access points, based on Q-learning algorithm to improve the system performance. We also make a performance comparison between the proposed and traditional LBT mechanisms. The simulation results show that our proposed mechanism provides better fairness compared to the conventional LBT mechanisms. © 2024 IEEE.

关键词： Wi-Fi

来源：评论

学校读者我要写书评

暂无评论

Multi-Layer Perceptrons and Convolutional Neural Networks Based Side-Channel Attacks on AES Encryption

Multi-Layer Perceptrons and Convolutional Neural Networks Ba...

引用

2023 International Conference on engineering technology and Technopreneurship, ICE2T 2023

作者： Ahmed, Amjed Abbas Hasan, Mohammad Kamrul Center for Cyber Security Faculty of Information Science and Technology Bangi 43600 Malaysia Department of Computer Techniques Engineering Baghdad Iraq

ISBN: (纸本)9798350304107

Side - channel attacks (SCA) exploit vulnerabilities introduced by insecure implementations in order to get information about the data that is being processed or the data itself via device leakages. The majority of modern attack techniques use AES systems, and several studies have exploited side-channel leakages. Exploiting leakages for the first and final rounds, often known as outer rounds, is frequently the foundation of the strategy. There is a possibility that outer round leakages cannot be attacked due to inadequate countermeasures or device design. The attacker has to focus their attention on the inner rounds. Deep learning side-channel attacks, whether they include profiling or not, have the potential to recover secret keys stored in cryptographic devices that use masked AES. Mask value profiling is being used in more recent techniques for the extraction of secret keys. The objective of our research is to defeat masked AES by deep learning methods like as multi-layer perceptrons and convolutional neural networks are used. In an environment when profiling is not taking place, we extract sensitive information using DDLA. The HW (Hamming Weight) model is used by our approach in order to carry out DDLA by means of an innovative binary labeling strategy. Using the ChipWhisperer power traces and the ASCAD dataset, we show profiling and non-profiling attacks. The accuracy of key recovery is at its maximum when the AES implementation is either masked or unmasked. © 2023 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A Literature Review on Predictive Classification of Major Depressive Disorder Using Advanced Data Analytics Techniques 5th

A Literature Review on Predictive Classification of Major De...

引用

5th International Conference on Data Science, Machine Learning and Applications, ICDSMLA 2023

作者： Mahender, Udutala Arivalagan, S. Sathiyasuntharam, V. Sudhakar, P. Annamalai University Tamilnadu Chidambaram India Department of Computer Science and Technology Annamalai University Tamilnadu Chidambaram India Department of Cyber Security CMR Engineering College Hyderabad India Department of Computer Science and Engineering Annamalai University Tamilnadu Chidambaram India

ISBN: (纸本)9789819780426

Major Depressive Disorder (MDD) is a significant global mental health concern predicted to become the leading mental health condition by 2030. It is characterised by persistent sadness, hopelessness, and a lack of interest in daily activities. The current diagnostic procedures for MDD are often unreliable, subjective, and prone to biases, emphasising the urgent need for accurate prediction methods. Predicting MDD is crucial for various reasons: it enables early detection, including identifying milder forms of depression, facilitating timely interventions and preventing the condition from worsening. Additionally, accurate prediction contributes to a deeper understanding of the disorder's neurobiological and psychological underpinnings, potentially leading to more effective treatments. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

Detecting Deepfake Videos using Face Recognition and Neural Networks 1

Detecting Deepfake Videos using Face Recognition and Neural ...

引用

1st IEEE International Conference on Cognitive Robotics and Intelligent Systems, ICC - ROBINS 2024

作者： Murugan, Muthu Aravind Mathu, T. Priya, S Jeba Karunya Institute of Technology and Sciences Computer Science and Engineering Coimbatore India Karunya Institute of Technology and Sciences Data Science and Cyber Security Coimbatore India

ISBN: (纸本)9798350372748

Deepfake videos created using advanced artificial intelligence techniques, pose a significant threat to digital media credibility. This project introduces a holistic strategy for identifying these videos, incorporating face recognition, feature extraction, and an innovative deep-learning model. The methodology involves pre-processing video data, extracting facial features using the face recognition library, and training a neural network model on processed face-only videos. The project filters videos based on frame count, extracts face, and creates a curated dataset for the detection model. Face-only videos are loaded and pre-processed to train a custom neural network model, which combines a pre-trained ResNext CNN with an LSTM layer for temporal feature extraction. The model is trained using Adam optimization with a cross-entropy loss function, and after completion, it has an accuracy of 95% and is capable of differentiate between fake and real videos using a confidence score. © 2024 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Advancing Multiclass Emotion Recognition with CNN-RNN Architecture and Illuminating Module for Real-time Precision using Facial Expressions

Advancing Multiclass Emotion Recognition with CNN-RNN Archit...

引用

2024 International Conference on Advances in Modern Age Technologies for Health and engineering Science, AMATHE 2024

作者： Prince, Simrith C. Babu, Nirmal Varghese Computer Science And Engineering Karunya Institute Of Technology And Sciences Coimbatore India Data Science And Cyber Security Karunya Institute Of Technology And Sciences Coimbatore India

ISBN: (纸本)9798350371567

Emotions, intrinsic to the human experience, serve as the catalyst for a groundbreaking technological innovation in multiclass emotion recognition through facial expressions. This study introduces a revolutionary system that harnesses advanced deep learning techniques, specifically Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN), to unravel the intricate interplay between emotions and cutting-edge technology. Commencing with a nuanced exploration of the evolutionary origins of emotions, the research lays a robust foundation for understanding the neural mechanisms behind emotional responses. The proposed algorithm unfolds in two pivotal stages: a meticulously designed CNN extracts intricate features from raw facial images, capturing nuanced expressions, while an adaptive RNN establishes temporal dependencies between sequential frames, allowing for a comprehensive understanding of the dynamic nature of emotional states. To address real world challenges, an innovative illuminating module is seamlessly integrated, accounting for contextual factors such as lighting conditions. Empirical evaluations on benchmark datasets underscore the system's exceptional accuracy, achieving an impressive 87.2% accuracy post hyperparameter tuning, surpassing state-of-the-art methods. This transformative method not only precisely captures static facial features but dynamically interprets evolving emotional expressions in real-time, marking a paradigm shift in the field of emotion recognition. Beyond its technical achievements, the proposed system holds profound implications across psychology, human-computer interaction, and market research, positioning itself as a pioneering force in the evolving landscape of emotion-centric technologies. © 2024 IEEE.

关键词： Emotion Recognition

来源：评论

学校读者我要写书评

暂无评论

Identifying The Coconut Leaf Disease Using Deep Transfer Learning Approach 4

Identifying The Coconut Leaf Disease Using Deep Transfer Lea...

引用

4th IEEE International Conference on Smart Information Systems and Technologies, SIST 2024

作者： Tuhin, Hemayet Hossain Kayum, Md Abdul Rahmatul Kabir Rasel Sarker, Md. Tanvir Siddiquee, Shah Md Pandey, Bishwajeet Department of Computer Science and Engineering Daffodil International University Dhaka Bangladesh Department of Intelligent System and Cyber Security Astana IT University Kazakhstan

ISBN: (纸本)9798350374865

Coconut is one of the main economic crops in Bangladesh. It is a tree whose every part is useful in one way or another in public life. The leaves, flowers, fruits, stems, and roots of this tree are used as raw materials for various small and large industries, materials for making various delicious foods, delicious drinks, and food for patients. This is the world’s most beautiful tree and is well known and appreciated by all as the’heavenly tree’. However, it has recently become known that most coconut trees suffer from illnesses that gradually weaken the trees’ health and coconut yield. Pest illnesses and nutrient deficits have an impact on the majority of the tree’s leaves. The main reason is that most of the farmers in our country are unaware. They do not know what kind of measures to take in the case of any disease. They are using pesticides, assuming ancient principles, sometimes benefiting and sometimes counterproductively, and making the farmers poor. As a result, farmers are showing disinterest in coconut cultivation. Our main objective is to increase the viability of coconut leaves and detect problems early on so that farmers can benefit more from the cultivation of coconuts. The study suggests examining diseases and detecting insect attacks and nutritional deficiencies in coconut leaves. This will help grow more coconuts. It is expected that this model will help the agricultural field on both an economic and ecological level. This study focuses on the three coconut diseases: WCLWD Yellowing, WCLWD Flaccidity, and CCI Caterpillars. Three neural networks have_ been chosen in this study to identify the best model. After evaluating the VGG19, MobileNet v2, Inception v3, and ResNet50 models, the accuracy for a collection of 3166 images was determined to be 96%, 96.02%, 97.77%, and 99.36%, respectively. As a result of this technology, farmers will be able to produce more coconuts, surely bringing about a revolution in the agricultural industry. © 2024 IEEE.

关键词： coconut leaf disease Inception v3 MobileNet v2 nutrient deficiency pest attacks ResNet50 VGG19

来源：评论

学校读者我要写书评

暂无评论

Multiclass Brain Tumor Detection using Convolutional Neural Network and Conventional Machine Learning 2

Multiclass Brain Tumor Detection using Convolutional Neural ...

引用

2nd IEEE International Conference on Emerging Research in Computational Science, ICERCS 2024

作者： Vidhya, K. Vamsi Krishna, B. Haritha, M. Rajasekaran, Arun Sekar Roopa Chandrika, R. Velusamy, Jayaraj Karunya Institute of Technology and Sciences Division of Data Science and Cyber Security School of Computer Science and Technology Coimbatore India QIS College of Engineering and Technology Department of AIMLDS Andhra Pradesh India QIS College of Engineering and Technology Department of Mechanical Engineering Andhra Pradesh India SR University Department of Electronics and Communication Engineering Telangana India Karpagam Academy of Higher Education Department of Cyber Security Coimbatore India Nehru Institute of Engineering and Technology Department of Electronics and Communication Engineering Coimbatore India

ISBN: (纸本)9798331534950

Brain tumor (BT) affect the brain and reduce the lifespan of humans. Therefore, diagnosis and treatment of illnesses in its early stage is required. The objective of this paper is to evaluate the effectiveness of the Convolutional Neural Network (CNN) and Conventional Machine Learning (CML) strategies for identifying multiple classes of brain tumor. In the research, a comparison is made between these two systems for the picture classification of brain tumor. The first step is the collection and preprocessing of images of several classes of brain tumor. The CNN and CML models are trained using the preprocessed dataset. CML makes use of traditional machine learning (ML) algorithms for the purpose of feature extraction and classification, whereas CNN relies on convolutional layers to automatically extract significant features from the images it is fed as input. Precision, accuracy, F1 score, and recall are the criteria that are utilized to evaluate methods. In terms of accuracy and overall performance, CNN is superior to CML. CNN provides a more precise classification of a variety of tumor classifications. Based on these findings, convolutional neural networks (CNN) are likely superior than machine learning for the identification of multiclass brain tumor. CNNs are superior because of their ability to directly understand intricate patterns and characteristics from photographs. This work demonstrates the usefulness of CNNs for multiclass classification in brain tumor detection systems because of their high accuracy and efficiency. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Circumventing Google Play vetting policies: a stealthy cyberattack that uses incremental updates to breach privacy

引用

Journal of Ambient Intelligence and Humanized Computing 2023年第5期14卷 4785-4794页

作者： Muhammad, Zia Amjad, Faisal Iqbal, Zafar Javed, Abdul Rehman Gadekallu, Thippa Reddy Department of Computer Science North Dakota State University Fargo United States Department of Cyber Security PAF Complex E-9 Air University Islamabad Pakistan Islamabad Pakistan School of Information Technology and Engineering Vellore Institute of Technology Vellore India Department of Electrical and Computer Engineering Lebanese American University Byblos Lebanon

Today digital technologies are evolving to accommodate small businesses and young entrepreneurs by reducing their time-to-market while encouraging rapid innovation in mobile, Extended Reality (XR), Internet of Things (IoT), cloud, and edge devices. The leading operating system Android typically takes one to a few days to perform application vetting and go to production by leveraging code analysis technologies in their Play Protect anti-malware program. However, developers with malicious intent are looking to circumvent this detection mechanism by exploiting Google’s relatively lenient trust policies that allow for package distribution and feature updates. This paper develops a proof-of-concept malware that exploits customers’ trust and Google’s policies to circumvent popular voice search applications. Our results show that attackers can initially circumvent Play Protect by uploading benign applications to build trust and then add malicious feature updates incrementally to distribute highly intrusive malware into user systems. This malware can scan and collect private user data from the device and exfiltrate it to the command-and-control server. The contributions are three-fold. (1) A proof-of-concept stealthy malware and publishing mechanism has developed that highlights the relative ease with which Google Play Protect policies may be subverted. (2) a comprehensive evaluation has been performed using major publicly available anti-malware solutions. (3) Recommendations and policies have been suggested to prevent this attack and ensure users’ privacy concerns (IMUTA is a novel attack in which malicious functionality is slowly added to a benign application through updates. This attack evades malware detection tools and exploits user trust. The attack can be launched against any application distribution platform like the Play Store). © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.

关键词： Mobile security

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：