检索结果-内蒙古大学图书馆

IEEE Conference on Global Communications (GLOBECOM)

作者： Rui Zhang Hongwei Li Wenbo Jiang Ran Zhang Jiaming He School of Computer Science and Engineering University of Electronic Science and Technology of China College of Computer Science and Cyber Security (Oxford Brookes College) Chengdu University of Technology China

ISBN: (数字)9798350351255

ISBN: (纸本)9798350351262

With the widespread use of deep learning systems in many applications, adversaries have strong incentives to perform attacks against these systems for their adversarial purposes. Reports have indicated that backdoor attacks on deep neural networks represent a novel form of threat. In this attack, the adversary will inject backdoors into the benign model and then mislead the model to classify the input containing backdoor triggers as a target label specified by the adversary. Existing research mainly focuses on backdoor attacks in image and text models, little attention has been paid to the backdoor attacks on text-to-speech (TTS) models. We conduct a systematic investigation of backdoor attacks on text-to-speech models and propose BadTTS, the first backdoor attack against TTS models, which is a general backdoor attack framework that tampers with input texts in three semantic levels to generate malicious output speech, including Char-Backdoor, Word-Backdoor, and Sentence-Backdoor. Our method not only efficiently injects backdoors into a TTS model but is also stealthy and has little impact on the synthesized speech quality. We implement the backdoor attack in a black-box fine-tuning setting, where the adversary has no knowledge of model architectures except for a small amount of training data. We perform empirical experiments on three representative and widely studied TTS models, indicating that backdoors can be injected into TTS models within a few fine-tuning steps. Additionally, we conduct experiments to explore the impact of different types of triggers, as well as the intermediate outputs of models, which provide insights for potential defenses against backdoor attacks.

关键词： Training Deep learning Systematics Semantics Training data Closed box Artificial neural networks Data models Text to speech Global communication

来源：评论

学校读者我要写书评

暂无评论

Multi-objective optimization of task assignment in distributed mobile edge computing

引用

Journal of Reliable Intelligent Environments 2022年第1期8卷 21-33页

作者： Almasri, Sanaa Jarrah, Moath Al-Duwairi, Basheer Department of Computer Engineering Jordan University of Science and Technology P.O. Box 3030 Irbid22110 Jordan Department of Computer Networks and Security Jordan University of Science and Technology P.O. Box 3030 Irbid22110 Jordan

Traditional computing models and centralized cloud computing are not capable of meeting today’s application requirements, especially when deploying technologies, such as the Internet of things (IoT), 5G, and wearable devices, on a large scale. Mobile edge computing (MEC) introduces the feasibility of using edge and smart devices, such as gateways and smart phones, to perform task execution of different applications. Moreover, an efficient task scheduling approach should consider the deadlines requirements and the power consumption of the edge devices. This paper proposes a multi-objective optimization solution to assign different application tasks to different edge devices while minimizing the energy consumption of edge devices and the computation time of tasks. Task dependencies and data distribution are considered within a new and more general MEC model. Multi-objective evolutionary algorithm (MOEA) framework is used to solve the optimization problem subject to deadline and power consumption constraints. Results show that the proposed multi-objective approach achieves better performance in terms of energy and computation time when compared to a single objective approach. © 2021, The Author(s), under exclusive licence to Springer Nature Switzerland AG.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

WGAN-AFL: Seed Generation Augmented Fuzzer with Wasserstein-GAN

arXiv

引用

arXiv 2024年

作者： Yang, Liqun Li, Chunan Qiu, Yongxin Wei, Chaoren Yang, Jian Guo, Hongcheng Ma, Jinxin Li, Zhoujun School of Cyber Science and Technology Beihang University China School of Computer Science and Engineering China Information Technology Security Evaluation Center China

The importance of addressing security vulnerabilities is indisputable, with software becoming crucial in sectors such as national defense and finance. Consequently, The security issues caused by software vulnerabilities cannot be ignored. Fuzz testing is an automated software testing technology that can detect vulnerabilities in the software. However, most previous fuzzers encounter challenges that fuzzing performance is sensitive to initial input seeds. In the absence of high-quality initial input seeds, fuzzers may expend significant resources on program path exploration, leading to a substantial decrease in the efficiency of vulnerability detection. To address this issue, we propose WGAN-AFL. By collecting high-quality testcases, we train a generative adversarial network (GAN) to learn their features, thereby obtaining high-quality initial input seeds. To overcome drawbacks like mode collapse and training instability inherent in GANs, we utilize the Wasserstein GAN (WGAN) architecture for training, further enhancing the quality of the generated seeds. Experimental results demonstrate that WGAN-AFL significantly outperforms the original AFL in terms of code coverage, new paths, and vulnerability discovery, demonstrating the effective enhancement of seed quality by WGAN-AFL. Copyright © 2024, The Authors. All rights reserved.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Orbit Weighting Scheme in the Context of Vector Space Information Retrieval

引用

computers, Materials & Continua 2024年第7期80卷 1347-1379页

作者： Ahmad Ababneh Yousef Sanjalawe Salam Fraihat Salam Al-E’mari Hamzah Alqudah Department of Computer Science Faculty of Information TechnologyAmerican University of MadabaAmman11821Jordan Department of Cybersecurity Faculty of Information TechnologyAmerican University of MadabaAmman11821Jordan Artificial Intelligence Research Center(AIRC) College of Engineering and Information TechnologyAjman UniversityP.O.Box 346Ajman13306United Arab Emirates Information Security Department Faculty of Information TechnologyUniversity of PetraAmman11196Jordan Department of Data Science and Artificial Intelligence Faculty of Information TechnologyAmerican University of MadabaAmman11821Jordan

This study introduces the Orbit Weighting Scheme(OWS),a novel approach aimed at enhancing the precision and efficiency of Vector Space information retrieval(IR)models,which have traditionally relied on weighting schemes like tf-idf and *** conventional methods often struggle with accurately capturing document relevance,leading to inefficiencies in both retrieval performance and index size *** proposes a dynamic weighting mechanism that evaluates the significance of terms based on their orbital position within the vector space,emphasizing term relationships and distribution patterns overlooked by existing *** research focuses on evaluating OWS’s impact on model accuracy using Information Retrieval metrics like Recall,Precision,InterpolatedAverage Precision(IAP),andMeanAverage Precision(MAP).Additionally,we assessOWS’s effectiveness in reducing the inverted index size,crucial for model *** compare OWS-based retrieval models against others using different schemes,including tf-idf variations and *** reveal OWS’s superiority,achieving a 54%Recall and 81%MAP,and a notable 38%reduction in the inverted index *** highlights OWS’s potential in optimizing retrieval processes and underscores the need for further research in this underrepresented area to fully leverage OWS’s capabilities in information retrieval methodologies.

关键词： Information retrieval orbit weighting scheme semantic text analysis Tf-Idf weighting scheme vector space model

来源：评论

学校读者我要写书评

暂无评论

Deployment of Scalable Cloud Based Image Processing Application Using Fuzzy Coupled Learning Technique 1

Deployment of Scalable Cloud Based Image Processing Applicat...

引用

1st International Conference on Computational Science and technology, ICCST 2022

作者： Mathivanan, B. Nandhashree, K.R. Radhakrishnan, C. Ananthi, S. Suganthi, P. Asokan, R. Sri Ramakrishna Engineering College Department of Cse Coimbatore India Srm Valliammai Engineering College Department of Cyber Security India Kongunadu College of Engineering and Technology Department of Cse Trichy India

ISBN: (纸本)9781665476553

Because image processing typically uses a lot of memory and computational time, a sophisticated platform is needed. As a result, it's critical to adopt cost-effective solutions to replace outdated ones. We chose to use cloud computing to fulfill our needs for large-scale data processing as a result of these factors. In the meantime, this strategy offers quick access to on-demand services that are highly available and scalable. By outsourcing computations to a third party and utilizing cloud services rather than internal applications, healthcare organizations can reduce operating costs. However, robust data protection against both unauthorized users and unreliable clouds is required to avoid the disclosure of malicious data. Users can now assemble and execute their data using the cloud thanks to the development of numerous frameworks. Distributed systems, cryptosystems, or a combination of the two are frequently used in their construction. Almost all current implementations use homomorphism cryptosystems, Service Oriented Architecture, secret share schemes, and secure multi-party computation as their primary security mechanisms. The main disadvantage of utilizing these techniques for large-scale data analysis on the cloud is the cost of computing image processing tasks. Keeping medical records & sensitive health-related information secure is a problem. We suggest a fresh method for protecting data processing in cloud environments that is based on machine learning methods. We typically employ Fuzzy C-means Clustering and Multi-Dense Convolution Neural Network to more accurately classify image pixels. To further reduce the danger of a possible leak of medical data, we add a third level, the CloudSaS-TPM module, to the conventional two-layered design. The results show some encouraging discoveries that provide new angles for creating cloud storage for the medical industry. © 2022 IEEE.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

IoT-Enhanced Gesture Recognition System for Healthcare Applications

IoT-Enhanced Gesture Recognition System for Healthcare Appli...

引用

2024 International Telecommunications Conference, ITC-Egypt 2024

作者： Krithika, L.B. Nallakaruppan, M.K. Baza, Mohamed Badr, Mahmoud Deora, Yuvraj Singh School of Computer Science Engineering and Information Systems Vellore Institute of Technology Vellore India College of Charleston Deparment of Computer Science CharlestonSC United States Polytechnic Institute Department of Networks and Computer Security UticaNY12201 United States

ISBN: (纸本)9798350351408

In response to the prevalent language barriers in medical settings, this paper proposes an innovative solution leveraging gesture recognition technology. Through the integration of deep learning models and Internet of Things (IoT) tools, the system aims to facilitate seamless communication between patients and healthcare providers. Initially, a deep learning model is developed using a personalized dataset containing various medical gestures, allowing for the recognition of nuanced patient expressions such as 'I have Fever' or 'Shoulder pain.' Additionally, the system incorporates IoT sensors to monitor vital signs like body temperature and pulse in real-time, ensuring comprehensive patient assessment. With the capability to continuously evolve and adapt to new gestures, this solution not only enhances accessibility but also improves the quality of healthcare delivery, particularly in underserved communities. By bridging the gap between language barriers and medical communication, this innovative approach holds promise for fostering inclusive and effective healthcare practices worldwide. In addressing the challenges posed by language barriers in medical interactions, this paper presents a comprehensive solution that combines gesture recognition technology with deep learning models and IoT devices. By capturing and interpreting patient gestures, such as indicating specific symptoms or discomfort, the system facilitates clear communication between patients and doctors. Furthermore, the integration of IoT sensors enables continuous monitoring of vital signs, allowing for timely interventions and personalized care. With its ability to adapt and expand to accommodate new gestures and medical scenarios, this solution represents a significant advancement in healthcare accessibility and quality. By leveraging cutting-edge technology, this approach has the potential to revolutionize healthcare delivery, particularly in areas with limited access to linguistic resources or medi

关键词： Gesture recognition

来源：评论

学校读者我要写书评

暂无评论

Efficient Wildcard Searchable Symmetric Encryption with Forward and Backward security 1st

Efficient Wildcard Searchable Symmetric Encryption with For...

引用

1st International Conference on Artificial Intelligence security and Privacy, AIS and P 2023

作者： Zhang, Xi Su, Ye Wei, Zhongkai Shen, Wenting Qin, Jing College of Computer and Cyber Security Hebei Normal University Shijiazhuang050024 China Hebei Provincial Key Laboratory of Network and Information Security Hebei Normal University Shijiazhuang050024 China School of Information Science and Engineering Shandong Normal University Jinan250358 China School of Mathematics Shandong University Jinan250100 China College of Computer Science and Technology Qingdao University Qingdao266071 China

ISBN: (纸本)9789819997848

Wildcard Searchable Symmetric Encryption can achieve flexibility and pattern matching while protecting data privacy. There is a promising future for combining secure wildcard search with emerging technologies such as AI. However, there are challenges in reducing the communication cost and improving security. In this paper, we propose an efficient wildcard searchable symmetric encryption with forward and backward security. The complexity of communication costs in the search protocol is O(1), independent of the number of characters of wildcard keywords, or files in the search result. It is achieved by a double-compressed index in which the character set and file identifiers are encoded simultaneously. Then, the double-compressed index provides a possibility to achieve oblivious keyword query and update, so that the proposed scheme only reveals the query type (search or update) without anything else. It achieves forward and backward security using the distributed multi-point function and an additively homomorphic symmetric encryption scheme. Detailed security proof and the theoretical comparison show the improvement in security and efficiency. The acceptable overheads of the proposed scheme are presented by the extensive performance evaluation. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Pattern matching

来源：评论

学校读者我要写书评

暂无评论

Convolutional Neural Network Architecture Training Parameters Impact on Multi-Frequency Propagation Channel Model in the VHF and UHF Bands 16

Convolutional Neural Network Architecture Training Parameter...

引用

16th IEEE AFRICON, AFRICON 2023

作者： Adebowale, Quadri Ramon Udensi, Ferguson O. Tunde, Adubi Faruk, Nasir Yusuf Olayinka, Imam-Fulani Sowande, Olugbenga A. Onidare, Samuel O. Oloyede, Abdulkarim A. Olawoyin, L.A. Garba, Salisu University of Ilorin Department of Telecommunication Science Ilorin Nigeria Ohio University Department of Electrical Engineering and Computer Science Athens United States Colorado State University Department of Electrical and Computer Engineering Fort Collins United States Sule Lamido University Department of Information Technology Kafin Hausa Nigeria Sule Lamido University Department of Software Engineering Kafin Hausa Nigeria Sule Lamido University Department of Cyber Security Kafin Hausa Nigeria

ISBN: (纸本)9798350336214

A path loss estimation model that is both computationally efficient, and precise is required for link budgeting, system performance optimization, base station selection, and coverage analysis. The limitations of empirical and deterministic models, on the other hand, necessitate the incorporation of computational intelligence into the path loss channel models development for multi-frequency band propagation channels. The principle and technique of Deep Neural Network (DNN) were applied in this paper for the modelling and development of a multi-frequency Convolutional Neural Network (CNN)-based path loss estimation model. The CNN architecture employed is a one-dimensional convolution - consisting of a convolution layer, a flattened layer, a dense layer and two fully connected layers. Filter sizes and epochs were varied to examine the effect on accuracy with respect to RMSE, and MSE during training. Results from the CNN model show that there is a significant impact of the training parameters in the development of the CNN model. © 2023 IEEE.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Fuzzy with Metaheuristics Based Routing for Clustered Wireless Sensor Networks

引用

Intelligent Automation & Soft Computing 2023年第1期35卷 367-380页

作者： Ashit Kumar Dutta Yasser Albagory Majed Alsanea Abdul Rahaman Wahab Sait Hazim Saleh AlRawashdeh Department of Computer Science and Information Systems College of Applied SciencesAlMaarefa UniversityAd DiriyahRiyadh13713Kingdom of Saudi Arabia Department of Computer Engineering College of Computers and Information TechnologyTaif UniversityTaif21944Kingdom of Saudi Arabia Department of Computing University Name:Arabeast CollegesRiyadh11583Kingdom of Saudi Arabia Department of Archives and Communication King Faisal UniversityAl AhsaHofuf31982Kingdom of Saudi Arabia Department of Cyber Security Onaizah Private CollegesUnaizah51911Kingdom of Saudi Arabia

Wireless sensor network(WSN)plays a vital part in real time tracking and data collection *** incorporates a set of numerous sensor nodes(SNs)commonly utilized to observe the target *** SNs operate using an inbuilt battery and it is not easier to replace or charge ***,proper utilization of available energy in the SNs is essential to prolong the lifetime of the *** this study,an effective Type-II Fuzzy Logic with Butterfly Optimization Based Route Selection(TFL-BOARS)has been developed for clustered *** TFL-BOARS technique intends to optimally select the cluster heads(CHs)and routes in the clustered ***,the TFL-BOARS technique incorporates Type-II Fuzzy Logic(T2FL)technique with distinct input parameters namely residual energy(RE),link quality(LKQ),trust level(TRL),inter-cluster distance(ICD)and node degree(NDE)to select CHs and construct ***,the butterfly optimization algorithm based route selection(BOARS)technique is derived to select optimal set of routes in the *** addition,the BOARS technique has computed afitness function using three parameters such as communication cost,distance and *** order to demonstrate the improved energy effectiveness and prolonged lifetime of the WSN,a wide-ranging simulation analysis was implemented and the experimental results reported the supremacy of the TFL-BOARS technique.

关键词： Type II-fuzzy logic uncertainty WSN energy lifetime route selection

来源：评论

学校读者我要写书评

暂无评论

Joint optimization for secure ambient backscatter communication in NOMA-enabled IoT networks

引用

Digital Communications and Networks 2023年第1期9卷 264-269页

作者： Wali Ullah Khan Furqan Jameel Asim Ihsan Omer Waqar Manzoor Ahmed Interdisciplinary Centre for Security Reliability and Trust(SnT)/SigComUniversity of LuxembourgLuxembourg Department of Communications and Networking Aalto University02150EspooFinland Department of Communication and Information Engineering Shanghai Jiao Tong UniversityShanghai200240China Department of Engineering Thompson Rivers University(TRU)British Columbia(BC)Canada College of Computer Science and Technology Qingdao UniversityQingdao266071China

Non-Orthogonal Multiple Access(NOMA)has emerged as a novel air interface technology for massive connectivity in Sixth-Generation(6G)*** recent integration of NOMA in Backscatter Communication(BC)has triggered significant research interest due to its applications in low-powered Internet of Things(IoT)***,the link security aspect of these networks has not been well *** article provides a new optimization framework for improving the physical layer security of the NOMA ambient BC *** system model takes into account the simultaneous operation of NOMA IoT users and the Backscatter Node(BN)in the presence of multiple EavesDroppers(EDs).The EDs in the surrounding area can overhear the communication of Base Station(BS)and BN due to the wireless broadcast ***,the chief aim is to enhance link security by optimizing the BN reflection coefficient and BS transmit *** gauge the performance of the proposed scheme,we also present the suboptimal NOMA and conventional orthogonal multiple access as benchmark *** Carlo simulation results demonstrate the superiority of the NOMA BC scheme over the pure NOMA scheme without the BC and conventional orthogonal multiple access schemes in terms of system secrecy rate.

关键词： 6G Non-orthogonal multiple access Ambient backscatter communication Internet-of-things Joint optimization Physical layer 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：