检索结果-内蒙古大学图书馆

A CMA-ES-Based Adversarial Attack Against Black-Box Object Detectors

Chinese Journal of Electronics 2021年第3期30卷 406-412页

作者： LYU Haoran TAN Yu'an XUE Yuan WANG Yajie XUE Jingfeng School of Computer Science and Technology Beijing Institute of Technology Academy of Military Science

Object detection is one of the essential tasks of computer vision. Object detectors based on the deep neural network have been used more and more widely in safe-sensitive applications, like face recognition, video surveillance, autonomous driving, and other tasks. It has been proved that object detectors are vulnerable to adversarial attacks. We propose a novel black-box attack method, which can successfully attack regression-based and region-based object detectors. We introduce methods to reduce search dimensions, reduce the dimension of optimization problems and reduce the number of queries by using the Covariance matrix adaptation Evolution strategy(CMA-ES) as the primary method to generate adversarial examples. Our method only adds adversarial perturbations in the object box to achieve a precise *** proposed attack can hide the specified object with an attack success rate of 86% and an average number of queries of 5, 124, and hide all objects with a success rate of74% and an average number of queries of 6, 154. Our work illustrates the effectiveness of the CMA-ES method to generate adversarial examples and proves the vulnerability of the object detectors against the adversarial attacks.

关键词： regression analysis video surveillance precise attack computer vision optimisation specified object essential tasks adversarial attacks region-based object detectors deep neural network object box adversarial perturbations primary method safe-sensitive applications attack regression-based black-box attack method CMA-ES-based adversarial attack evolutionary computation neural nets face recognition adversarial examples CMA-ES method attack success rate covariance matrices black-box object detectors object detection

来源：评论

学校读者我要写书评

暂无评论

Integrating Decision Trees with Cloud computing for Predicting Cognitive Decline in Alzheimer's Patients 2

Integrating Decision Trees with Cloud Computing for Predicti...

引用

2nd IEEE International Conference on Integrated Circuits and Communication Systems, ICICACS 2024

作者： Meenakshi, K. Ravinder, Bathini Pasha, Mohammed Rehaman Thinakaran, K. Jayanayudu, Dabbara Maniraj, S.P. SRM Institute of Science and Technology Faculty of Engineering and Technology Department of Computer Science and Engineering Tamil Nadu Vadapalani India Koneru Lakshmaiah Education Foundation Department of Computer Science and Engineering Andhra Pradesh Guntur India Malla Reddy Institute of Engineering and Technology Department of Computer Science and Engineering Telangana Hyderabad India Saveetha Institute of Medical and Technical Sciences Department of Computer Science and Engineering Saveetha School of Engineering Tamil Nadu Chennai India Siddartha Institute of Science and Technology Department of Electronics and Communication Engineering Andhra Pradesh Puttur India SRM Institute of Science and Technology Department of Computer Science and Engineering Tamil Nadu Chennai India

ISBN: (纸本)9798350317558

The proposed research work suggests a combined technique based on decision trees and cloud computing for predicting cognitive decline in Alzheimer's patients. Medical records and data from wearable devices are among the many sources that perform preprocessing and feature engineering. Decision tree models can make the most efficient use of computing resources. Decision tree topologies are tested for performance criteria, including accuracy and F1 score. The system's predictive powers are rather encouraging since it frequently achieves accuracy rates over 90%. However, limitations in data quality and the intricacy of decision tree interpretation provide challenges. The prediction of cognitive deterioration in Alzheimer's patients may be made more scalable and interpretable by integrating decision trees with cloud computing. This method has promise for early identification and intervention despite its limitations. To further improve prediction accuracy and clinical value, future research should increase data quality and make it easier to understand. © 2024 IEEE.

关键词： Alzheimer's disease Cloud computing Decision trees Early detection Healthcare

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Based Scientific Document Summarization Considering Citation

引用

SN computer science 2025年第4期6卷 1-13页

作者： Jyoti, Divya Mahato, Dharmendra Prasad Srivastava, Jyoti Department of Computer Science and Engineering National Institute of Technology Hamirpur Himachal Pradesh Hamirpur 177 005 India

The rapid expansion of scientific publications poses a significant challenge for researchers who need to stay updated on recent advancements and efficiently extract crucial information. Summarizing scientific documents can address this challenge by providing concise overviews that highlight the main contributions of the original texts. This study introduces a cutting-edge deep learning approach that employs the sBERT model to enhance the accuracy and relevance of document summaries. Our method treats extractive summarization as a classification task, implementing a dual BERT model configuration to effectively pinpoint and extract critical information. Tested on the CL-SciSumm dataset, our approach demonstrates marked improvements over existing techniques, achieving higher ROUGE scores. These results affirm the potential of our model to generate accurate and dependable summaries, assisting researchers in managing and interpreting the growing volume of scientific literature. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： BERT and sBERT model Citation based summarization Deep learning Natural language processing Scientific documentation Spacy

来源：评论

学校读者我要写书评

暂无评论

SocialChain: A Decentralized Social Media Platform on the Blockchain 1

引用

2nd EAI International Conference on Pervasive knowledge and collective intelligence on Web and Social Media, PerSOM 2023

作者： Prashanth, Mallellu Sai Maheswari, V. Uma Aluvalu, Rajinikanth Kantipudi, M. V. V. Prasad Departmentof Computer Science and Engineering Vardhaman College of Engineering Hyderabad India Departmentof Computer Science and Engineering Chaitanya Bharathi Institute of Technology Hyderabad India Departmentof Information Technology Chaitanya Bharathi Institute of Technology Hyderabad India Pune India

ISBN: (数字)9783031660443

ISBN: (纸本)9783031660436

The emergence of blockchain technology has paved the way for innovative solutions in various domains, including social media. This research paper presents the design and implementation of "SocialChain," a decentralized social media platform built on the Ethereum blockchain using the Solidity programming language. SocialChain aims to address issues related to data privacy, content ownership, and censorship prevalent in traditional social media platforms. The proposed system leverages blockchain’s inherent properties such as immutability, transparency, and decentralization to create a trustless and secure environment for social interactions. The platform allows users to create posts, share content, and engage in discussions while maintaining ownership and control over their data. The underlying blockchain ensures that the history of posts and interactions remains tamper-proof and verifiable. The core components of the SocialChain smart contract include data structures for posts and comments, functions for creating posts and adding comments, as well as mechanisms for retrieving post-related information. Each user’s identity is associated with their Ethereum address, eliminating the need for traditional authentication methods and enhancing user pseudonymity. To ensure efficiency, the contract optimizes data storage by using mappings and arrays, minimizing gas costs and improving scalability. The contract also incorporates modifiers to validate post and comment existence, enhancing security and preventing unauthorized access. The results of this research showcase the feasibility of integrating blockchain technology into social media platforms. However, it is important to note that while the implementation provides a foundational framework, practical deployment requires addressing challenges such as user experience, scalability, and incentivization mechanisms. In conclusion, SocialChain exemplifies the potential of blockchain technology to revolutionize the landscape of s

关键词： Decentralized finance

来源：评论

学校读者我要写书评

暂无评论

Leveraging contrastive learning for enhanced node representations in tokenized graph transformers 24

Leveraging contrastive learning for enhanced node representa...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Jinsong Chen Hanpeng Liu John E. Hopcroft Kun He School of Computer Science and Technology and Institute of Artificial Intelligence and Hopcroft Center on Computing Science Huazhong University of Science and Technology School of Computer Science and Technology and Hopcroft Center on Computing Science Huazhong University of Science and Technology Hopcroft Center on Computing Science Huazhong University of Science and Technology and Department of Computer Science Cornell University

ISBN: (纸本)9798331314385

While tokenized graph Transformers have demonstrated strong performance in node classification tasks, their reliance on a limited subset of nodes with high similarity scores for constructing token sequences overlooks valuable information from other nodes, hindering their ability to fully harness graph information for learning optimal node representations. To address this limitation, we propose a novel graph Transformer called GCFormer. Unlike previous approaches, GCFormer develops a hybrid token generator to create two types of token sequences, positive and negative, to capture diverse graph information. And a tailored Transformer-based backbone is adopted to learn meaningful node representations from these generated token sequences. Additionally, GCFormer introduces contrastive learning to extract valuable information from both positive and negative token sequences, enhancing the quality of learned node representations. Extensive experimental results across various datasets, including homophily and heterophily graphs, demonstrate the superiority of GCFormer in node classification, when compared to representative graph neural networks (GNNs) and graph Transformers.

关键词：

来源：评论

学校读者我要写书评

暂无评论

BookGenius: Elevating Reading Exploration through Full-Stack Mastery 7

BookGenius: Elevating Reading Exploration through Full-Stack...

引用

7th International Conference on Circuit Power and computing Technologies, ICCPCT 2024

作者： Ruchitha, Gangishetty Jyoshna, Dasari Galla, Gnana Sri Koushik Varma, Indukuri Sai Manikanta Babu, Ch. Madhu Pallavi, L. B V Raju Institute of Technology Narsapur Department of Computer Science and Engineering Telangana Narsapur India B V Raju Institute of Technology Department of Computer Science and Engineering Telangana Narsapur India

ISBN: (纸本)9798350372816

BookGenius introduces a groundbreaking platform designed to revolutionize book discovery. Leveraging full-stack development principles, BookGenius offers personalized book recommendations, detailed overview of books, audio book functionality, interactive quizzes, social recommendation functionalities and streak maintenance capabilities to enhance user engagement. BookGenius employs a full-stack development approach, integrating frontend technologies for the user interface and back-end technologies for server-side operations. Through data-driven algorithms and user interaction strategies, the platform delivers tailored book suggestions, summaries of books, audio book functionality, social recommendation functionalities and quizzes while incentivizing regular usage through streak maintenance features. By embracing full-stack development principles, BookGenius creates a dynamic and user-centric platform. The integration of personalized content, interactive quizzes, and streak maintenance functionalities enhances user satisfaction and engagement. BookGenius showcases the transformative potential of full-stack development in redefining the book browsing experience. By providing a seamless and immersive platform for discovering and engaging with literature, BookGenius aims to revolutionize the way users interact with books in the digital age. © 2024 IEEE.

关键词： Maintenance

来源：评论

学校读者我要写书评

暂无评论

Customer Phycological Behaviours Analysis Using Ml Algorithm to Customise the Service 2

Customer Phycological Behaviours Analysis Using Ml Algorithm...

引用

2nd IEEE International Conference on IoT, Communication and Automation technology, ICICAT 2024

作者： Balaji Srikaanth, P. Sudhakaran, Pradeep Shirisha, K. SRM Institute of Science and Technology Department of Networking and Communications Tamil Nadu Kattankulathur603203 India SRM Institute of Science and Technology Department of Computing Technologies Tamil Nadu Kattankulathur603203 India

ISBN: (纸本)9798350368109

Pursuing profitability and a loyal customer base in the current conflict via opposition banking structure requires understanding and predicting customer churn. This research focuses on the usage of Machine Learning (ML) techniques with a view of developing a predictive model regarding customer churn in the context of banking industries. The use of transactional patterns, previous customer data, and other behavioral traits enables the training of ML algorithms apropos to and deployed for usage. We are planning to discover early signs and trends that could help foresee that a customer is about to churn by applying the above algorithms. To this end, the following are some of the benefits financial institutions can harness from this kind of predictive capacity: The proposed model can help to enable such proactive approaches to retention and marketing because it envisages which of its consumers are likely to go where, and when their attrition is most probable. The ability of the project to, enhance overall organizational effectiveness, enhance the levels of customer satisfaction, and fashion better customer loyalty are the aspects that put the project in the right place within the banking industry. Considering the life cycle of market change and complex customer's behaviour, the use of machine learning to predict the customer churn is one of the strategic tools which allows financial institutions to build long-term customer relationship and make a proper decision based on the customer data gained. © 2024 IEEE.

关键词： Sales

来源：评论

学校读者我要写书评

暂无评论

Underwater Image Enhancement using CLAHE Fusion in YIQ and HSI Color Model 4

Underwater Image Enhancement using CLAHE Fusion in YIQ and H...

引用

4th IEEE International Conference on Emerging Frontiers in Electrical and Electronic Technologies, ICEFEET 2024

作者： Kumar, Ravi Kumar, Ashish Kumar, Manikant Rai, Mithilesh Kumar Kumari, Shweta Mourya, Vishal Electronics and Communication Engineering Muzaffarpur Institute of Technology Muzaffarpur India Computer Science Engineering Muzaffarpur Institute of Technology Muzaffarpur India Leather Technology Muzaffarpur Institute of Technology Muzaffarpur India Information Technology Muzaffarpur Institute of Technology Muzaffarpur India National Institute of Electronics and Information Technology Patna India

ISBN: (纸本)9798331528614

- Underwater image enhancement has been employed in a variety of areas, such as terrain scanning, underwater microscopic detection, telecommunication cables, mine detection, and autonomous underwater vehicles. However, underwater imagery suffers from strong scattering, noise, color distortion, and absorption from artificial light sources, causing haziness, image blur, and a greenish or bluish texture. To address these issues, the proposed work introduces a new enhancement approach based on Contrast Limited Adaptive Histogram Equalization (CLAHE) used for images taken underwater. The proposed method employs CLAHE to adaptively enhance local contrast while constraining the amplification of noise and artifacts. The fusion method is proposed to enhance the complementary advantages of different color spaces in YIQ and HSI color models. The YIQ color space separates intensity from chrominance, allowing for independent processing of luminance information, which is crucial for preserving image details and contrast. The proposed algorithm improves low-contrast underwater images with intact color and a natural appearance with reliability according to the results of the qualitative analysis. The proposed method outperforms the state-of-the-art methods in terms of image performance measures like GMSD, VSI, MCSD, and PSNR. ©2024 IEEE.

关键词： Underwater imaging

来源：评论

学校读者我要写书评

暂无评论

A Novel Split Learning-Based Consumer Electronics Network Traffic Anomaly Detection Framework for Smart City Environment

引用

IEEE Transactions on Consumer Electronics 2024年第1期70卷 4197-4204页

作者： Kim, Dong-Jo Amma, N. G. Bhuvaneswari Sarveshwaran, Velliangiri Sunchon National University Department of Image Design Sunchon57922 Korea Republic of Vellore Institute of Technology School of Computer Science and Engineering Chennai600 127 India Srm Institute of Science and Technology Department of Computational Intelligence Chennai603 203 India National Chung Cheng University Department of Computer Science and Information Engineering Chiayi621301 Taiwan

Due to technology advancement, connectivity and intelligence have revolutionized the consumer electronics (CE) world and IoT systems pose a new cybersecurity threat because smart devices are constantly connected to sensors, which are connected directly to enormous cloud servers. To address the cybersecurity vulnerabilities associated with IoT in smart cities, we provide a Software Defined Network (SDN) orchestrated anomaly detection system based on Split Learning (SL). The traffic generated by the CE is pre-processed and the relevant traffic features are selected. Then the model is split into at least two sub-models using SL to detach model training from the necessity for direct access to raw data. Client-side sub-models are trained on local training data on client devices, whereas server-side sub-models are trained on servers. Experiments show that the proposed SL-based CE network anomaly detection framework decreases computation needs on the client side by just computing a sub-model rather than the complete computationally costly model. The proposed solution has the potential to detect compromised CE network traffic in smart city environments. As evidenced by our results, the proposed framework performed significantly better than the existing state-of-the-art anomaly detection system in terms of accuracy. © 1975-2011 IEEE.

关键词： Smart city

来源：评论

学校读者我要写书评

暂无评论

Advancing Multilingual Communication NLP based Translational Speech to Speech Dialogue System for Indian Languages 1

Advancing Multilingual Communication NLP based Translational...

引用

1st IEEE International Conference on Electronics, computing, Communication and Control technology, ICECCC 2024

作者： Agarwal, Umansh Gupta, Rishika Saranya, G. School of Computing College of Engineering and Technology SRM Institute of Science and Technology Department of Networking and Communications Tamil Nadu Kattankulathur Chennai603203 India

ISBN: (纸本)9798350371802

Today it is more crucial than ever to remove language barriers in the linked world of today. This study compares and contrasts on-demand live translation services with an emphasis on effectiveness and accuracy. It presents a workable approach to smooth multilingual communication, with a focus on Indian languages. A state-of-the-art on-call speech translation system built on a sentence-level Neural Machine Translation (NMT) model is revealed in this work. The NMT approach translates individual sentences, as opposed to previous algorithms that work with full documents, guaranteeing accuracy and context awareness. NMT, BERT, and XL Net are also compared, emphasizing their unique qualities and areas of use. To facilitate easy communication in chosen languages, this model supports Hindi, Tamil, Telugu, Malayalam, Urdu, Bengali, and English. This helps address the linguistic variety of India. Its versatility and efficacy are a result of intensive training on a wide range of language pairings, highlighting the contribution of technology to intercultural communication. The study emphasizes how the NMT-based on-call voice translation system improves commu-nication accessibility in our multicultural, globalized society by bridging linguistic gaps, particularly in Indian languages. In addition, a unique application of multithreading has been employed to lower latency by concurrently executing the transcription and translation operations. Research shows that when multithreading is used, the time required for a translation decreases to 93.82 seconds from an initial 137 seconds. This helps to some extent with the latency issue © 2024 IEEE.

关键词： Neural machine translation

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：