检索结果-内蒙古大学图书馆

Conference on Computer Vision and Pattern recognition (CVPr)

作者： Huantong Li Xiangmiao Wu Fanbing Lv daihai Liao Thomas H. Li Yonggang Zhang Bo Han Mingkui Tan South China University of Technology Information Technology R&D Innovation Center of Peking University Changsha Hisense Intelligent System Research Institute Co. Ltd School of Electronic and Computer Engineering Peking University Shenzhen Graduate School Shenzhen China Hong Kong Baptist University Key Laboratory of Big Data and Intelligent Robot Ministry of Education PengCheng Laboratory

Zero-shot quantization (ZSQ) is promising for compressing and accelerating deep neural networks when the data for training full-precision models are inaccessible. In ZSQ, network quantization is performed using synthetic samples, thus, the performance of quantized models depends heavily on the quality of synthetic samples. Nonetheless, we find that the synthetic samples constructed in existing ZSQ methods can be easily fitted by models. Accordingly, quantized models obtained by these methods suffer from significant performance degradation on hard samples. To address this issue, we propose HArd sample Synthesizing and Training (HAST). Specifically, HAST pays more attention to hard samples when synthesizing samples and makes synthetic samples hard to fit when training quantized models. HAST aligns features extracted by full-precision and quantized models to ensure the similarity between features extracted by these two models. Extensive experiments show that HAST significantly outperforms existing ZSQ methods, achieving performance comparable to models that are quantized with real data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Security Challenges and Elucidations in Cloud Storage and File Systems: An Advanced Investigation review

Security Challenges and Elucidations in Cloud Storage and Fi...

引用

Ubiquitous Computing and Intelligent Information Systems (ICUIS), International Conference on

作者： S. P. Santhoshkumar S. Hariharasudhan S. Poorna Prakash Jim Mathew Philip S. Haritha T. Nalini Department of Computer Science and Engineering Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Avadi Chennai India Computer Science and Engineering Vel Tech Rangarajan Dr. Sagunthala R & D Institute of Science and Technology Avadi Chennai India Artificial Intelligence & Data Science Akshaya College of Engineering and Technology Coimbatore India Department of Computer Science and Engineering Sri Ramakrishna Institute of Technology Coimbatore Tamil Nadu India Master of Computer Application Rathinam Technical Campus. Coimbatore India Computer Science and Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Chennai

ISBN: (数字)9798331529635

ISBN: (纸本)9798331529642

With the widespread adoption of cloud storage and file systems, security concerns have become paramount. This paper presents an advanced investigation into the security challenges encountered in cloud storage and file systems. We explore the vulnerabilities and threats that these systems face, ranging from unauthorized access to data breaches and insider attacks. Additionally, we examine the evolving landscape of security solutions, including encryption techniques, access control mechanisms, and intrusion detection systems. Through a comprehensive review of existing literature and case studies, we identify emerging trends and best practices for mitigating security risks in cloud storage and file systems. Our analysis provides valuable insights for researchers, practitioners, and organizations seeking to enhance the security posture of their cloud-based infrastructure. Cloud storage and file systems have revolutionized data management, offering scalability, flexibility, and accessibility. However, alongside their benefits, they present formidable security challenges. This paper conducts an advanced investigation into the security landscape of cloud storage and file systems, aiming to interpret key challenges and propose effective solutions. Through a thorough examination of prevalent security threats, including unauthorized access, data breaches, and insider attacks, this study identifies vulnerabilities inherent in cloud environments. Moreover, it scrutinizes emerging security technologies such as encryption, access control mechanisms, and intrusion detection systems, assessing their efficacy in mitigating risks. By synthesizing insights from existing literature and real-world case studies, this investigation provides valuable guidance for enhancing the security posture of cloud-based infrastructure. Ultimately, this paper contributes to the ongoing discourse on cloud security by offering nuanced analyses and actionable recommendations to researchers, practitioners,

关键词： Cloud computing File systems reviews Scalability Cloud computing security Intrusion detection data breach Market research Encryption Security

来源：评论

学校读者我要写书评

暂无评论

Analysis of cloud services using OWASP security design

Analysis of cloud services using OWASP security design

引用

research Methodologies in Knowledge Management, Artificial Intelligence and Telecommunication Engineering (rMKMATE), International Conference on

作者： Elumalaivasan Poongavanam Munirathinam. T Kayalvizhi. V Sankar Murugesan Department of Data Science and Business System SRM Institute of Science and Technology Chennai Tamil Nadu India Department of Information Technology Sona College of Technology Salem Tamil Nadu India School of Computer Science and Engineering and Information Science Presidency University Bangalore Karnataka India Department of Computer Science and Engineering Vel Tech Rangarajan Dr.Sagunthala R&D Institute of Science and Technology Chennai Tamil Nadu India

Cloud computing is extremely popular model among enterprises. It provides more availability of cloud resources when needed with high scalability and extremely cost-effective model. Security is one of the least understood concepts to provide and access for OWASP applications in cloud computing. Many enterprises especially banking clients are risk-averse and have less cloud adoption rate. In this study, security research in the area of cloud security is reviewed and AWS (Amazon Web Service) is the top provider of cloud computing. Since, we have also demonstrated how it works in the security exploration. This paper's major goal is to make cloud computing security is a key activity and prevent unauthorized access to the apps used in the real world.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Hybrid Machine Learning Approach of “KNN LSTM” for detection of Brain Tumour in Mr images

Hybrid Machine Learning Approach of “KNN LSTM” for detecti...

引用

International Conference on Computing and Networking technology (ICCNT)

作者： G. Uganya Shruthi V Kulkarni Vani Valsaraj Anuja Govindrao Gunale d. rajalakshmi B Srividhya Department of ECE Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Chennai India Department of AIML HKBK College of Engineering opposite manyata tech park Nagawara Bengaluru Karnataka Department of Computer Engineering UPL University of Sustainable Technology Block no 402 Ankleshwar-Valia Road Vataria Gujarat Department of Data Science and Business Systems SRM Institute of Science and Technology Kattankulathur Chennai Department of ComputerApplication Faculty of Humanities and Sciences SRM Institute Of Science and Technology Chengalpattu Tamil Nadu

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

Now a days, more existing techniques are available to found tumours on brain from Mr images. But still the researchers are concentrated on to detect the brain tumour accurately with less time. So, this work proposed the hybrid machine learning algorithm of “K-Nearest Neighbour - Long Short-Term Memory” (“KNN-LSTM”) with increased ‘accuracy’, ‘sensitivity’, ‘specificity’, with less ‘time consumption’. In this work, the brain tumour is identified by predicting the Mr images from huge dataset with usage of machine learning approach of “KNN-LSTM”. Brain tumour detection is difficult when to compare the abnormal and typical tissues of large capacity. In order to enable a speedy recovery following therapy, brain tumours should be identified as soon as feasible. If the brain tumour is discovered early on, patients can benefit from improved care. due to the irregular spreading of the abrasions, brain tumours are still difficult to identify even with intensive study. Sometimes the tiny tumours look like normal, but it is challenging to find it. So, the techniques are need to increase the classification accuracy, and selecting the useful features. This paper proposes a hybrid machine learning model “KNN-LSTM” for categorizing the tumours with predicting through Mr images. MrI brain imaging collection is used for our experimentation. The “KNN” algorithm usage is to extract the important characteristics from pictures once the data has been pre-processed effectively. The proposed work predicts the brain tumour with 99.15% of ‘accuracy’, 98% of ‘sensitivity’, 97.12% of ‘specificity’ and testing time consumption of 0.5 minutes.

关键词： Machine learning algorithms Accuracy Sensitivity Magnetic resonance imaging Machine learning Nearest neighbor methods Brain modeling Long short term memory Tumors Testing

来源：评论

学校读者我要写书评

暂无评论

Experimental Analysis of radioscopy Images based Pneumonia disease detection using Improved Neural Network Methodology

Experimental Analysis of Radioscopy Images based Pneumonia D...

引用

Image Processing and Capsule Networks (ICIPCN), International Conference on

作者： Sajja Suneel B.M.S. rani M. Amina Begum Jayendra Gopal Thatipudi Ajay Singh Yadav d. Muthukumaran Department of CSE(Data Science) Institute of Aeronautical Engineering Hyderabad Telangana India Department of ECE Vignan’s Nirula Institute of technology and science for women Department of ECE Oxford College of Engineering Venmani Tamilnadu India Dept. of Computer Science University of North Texas Department of Mathematics SRM Institute of Science and Technology Delhi-NCR Campus Ghaziabad U.P India Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology

ISBN: (数字)9798350367171

ISBN: (纸本)9798350367188

The bacterium Streptococcus pneumonia is the most prevalent causative agent of pneumonia, a severe and potentially fatal lung infection in humans. According to the World Health Organization, pneumonia accounts for one in three deaths in India. Assessing chest radiographs to detect pneumonia typically requires skilled radiologists, making an automatic detection method crucial for rapid treatment, especially in rural areas. This study introduces an Improved Neural Network (INN) learning method, cross-validated against existing Convolutional Neural Networks (CNNs), reflecting the success of deep-learning algorithms in medical image analysis. The research highlights the utility of these methods for disease classification. The features learned by INN and CNN models on large datasets are particularly valuable for image classification tasks. By using feature extractors programmed on trained INN and CNN frameworks, this study compares the performance of various classifiers in identifying normal and abnormal chest radiographs. Statistical results indicate that pre-trained INN models combined with supervised classifier approaches significantly enhance the evaluation of chest radiographs, particularly in detecting pneumonia. This advancement is pivotal for improving diagnosis and treatment outcomes in resource-limited environment.

关键词： Learning systems Pneumonia Neural networks Lung Organizations Feature extraction Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Prototype-Guided Continual Adaptation for Class-Incremental Unsupervised domain Adaptation

arXiv

引用

arXiv 2022年

作者： Lin, Hongbin Zhang, Yifan Qiu, Zhen Niu, Shuaicheng Gan, Chuang Liu, Yanxia Tan, Mingkui South China University of Technology China National University of Singapore Singapore Information Technology R&D Innovation Center Peking University China MIT-IBM Watson AI Lab United States Key Laboratory of Big Data and Intelligent Robot Ministry of Education China

This paper studies a new, practical but challenging problem, called Class-Incremental Unsupervised domain Adaptation (CI-UdA), where the labeled source domain contains all classes, but the classes in the unlabeled target domain increase sequentially. This problem is challenging due to two difficulties. First, source and target label sets are inconsistent at each time step, which makes it difficult to conduct accurate domain alignment. Second, previous target classes are unavailable in the current step, resulting in the forgetting of previous knowledge. To address this problem, we propose a novel Prototype-guided Continual Adaptation (ProCA) method, consisting of two solution strategies. 1) Label prototype identification: we identify target label prototypes by detecting shared classes with cumulative prediction probabilities of target samples. 2) Prototype-based alignment and replay: based on the identified label prototypes, we align both domains and enforce the model to retain previous knowledge. With these two strategies, ProCA is able to adapt the source model to a class-incremental unlabeled target domain effectively. Extensive experiments demonstrate the effectiveness and superiority of ProCA in resolving CI-UdA. The source code is available at https://***/Hongbin98/***. Copyright © 2022, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Automated Classification and diagnosis of Focal and Non-Focal EEG Signals using a Hybrid Classification Approach

Automated Classification and Diagnosis of Focal and Non-Foca...

引用

distributed Computing and Electrical Circuits and Electronics (ICdCECE), IEEE International Conference on

作者： V. Kishore Kumar Anubhav Sharma Y. Prasanthi Priyadarshini M. C Karre ramesh Homera durani Department of Electronics and Communication Engineering Chennai Institute of Technology Chennai India Department of Computer Science and Engineering IES Insititute of Technology and Management IES University Bhopal India Department of Computer Science and Engineering-Data Science KKR & KSR Institute of Technology And Sciences Guntur India Department of English Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Chennai India Department of Computer Science and Engineering(DS) CMR Engineering College Hyderabad India Department of MCA School of Engineering RK University Rajkot India

ISBN: (数字)9798331533663

ISBN: (纸本)9798331533670

As a result of the non-linear in the categorizing technique, the initial section of the study, which sorts EEG data into focal and non-focal, is not suitable for severe evaluation. A combined classification system is therefore suggested for the purpose of automated epilepsy illness diagnosis and level of severity evaluation. Stages for breakdown, extraction of features, and categorization make up the suggested approach. After applying dTWCT to the EEG signals, we scale them utilizing a signal's scaling coefficient and use the wavelet transform to break them down into their component bands. Then, dTCWT takes these deconstructed sub band correlations and uses them to extract parameters like average, variance, skewing, average deviation, kurtosis, covariance, or period, and coefficients of correlation. The next step is to use a Neural Network (NN) categorization strategy to sort the EEG data into two categories: focal and non-focal. "Mild" and "Severe" are the additional diagnostic categories used for the categorized focused EEG data. A larger quantity of learned signals is needed by NN in order to perform the diagnostic procedure. If fuzzy limitations might improve the NN classifier's diagnostic accuracy, they aren't there. In comparison to Shearlet Transform, discrete Wavelet Transform (dWT), and Contourlet Transform, the proposed method that makes use of dTCWT maximizes accuracy to 100%. This establishes the hybrid categorization approach as the gold standard for automated seizure detection.

关键词： Correlation Accuracy Neural activity Epilepsy Transforms Artificial neural networks Feature extraction Electroencephalography Kurtosis discrete wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Synergizing Machine Learning Algorithms with Quantum Computing: A Path to Unprecedented Capabilities

Synergizing Machine Learning Algorithms with Quantum Computi...

引用

Advances in Computing, Communication & Materials (ICACCM), 2020 International Conference on

作者： dileep Pulugu K. Subathra Srichandana Abbineni S. Zulaikha Beevi B. Aishwarya Sananda Kumar Department of Computer Science and Engineering Malla Reddy College of Engineering and Technology Hyderabad Telangana Department of Mathematics Sri Sairam Engineering College Chennai Tamilnadu Department of ET-CSE (Data Science) CVR College of Engineering Hyderabad Telangana Department of CSE JCT College of Engineering and Technology Coimbatore Tamilnadu Department of Computer Science and Engineering (Cyber Security) Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology School of Electronics KIIT Deemed to be University Bhubaneswar Odisha

ISBN: (数字)9798350367973

ISBN: (纸本)9798350367980

Quantum machine learning has become a promising field due to the capability to solve the computational issues of classical machine learning models. This research investigates the integration of machine learning and quantum computing with the aim of establishing and utilizing the QML models including QSVM and QNN. Given the large number of experiments on both conventional and quantum enriched datasets, the authors show that quantum machine learning models can be faster and more accurate when compared to Classical models. Furthermore, they provide a practical way for extending current quantum models’ capabilities to fight against available quantum device limitations, achieve higher computation rates and better data handling. Savings in training times and better performance are well-demonstrated in the outcome – especially with large data sets like MNIST. As much as such issues as quantum hardware constraints persist, marrying quantum computing with machine learning expands the ability to solve problems that are big and complicated across several fields including finance, healthcare and AI. This paper therefore encourages more investment on the you need more research and development towards the realization of quantum computing in the realm of machine learning.

关键词： Training Support vector machines Quantum computing Machine learning algorithms Computational modeling Neural networks Machine learning Hardware data models Optimization

来源：评论

学校读者我要写书评

暂无评论

Creating a Secure data Sharing Network for disease Identification Using CT Images and rHC-Based Encryption Scheme

Creating a Secure Data Sharing Network for Disease Identific...

引用

Intelligent and Cloud Computing (ICoICC), International Conference on

作者： Vinston raja r Jose P rajakumar M P r. Balamurugan M. S. Malchijah raj M. robinson Joel Department of Computational Intelligence Faculty of Engineering and Technology School of Computing SRM Institute of Science and Technology Chennai India Department of CSE Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Chennai Tamil Nadu India Department of Artificial Intelligence and Data Science OMR St. Joseph’s College of Engineering Chennai Tamil Nadu India Dept. of Artificial Intelligence and Data Science New Prince Shri Bhavani College of Engineering & Technology Chennai Tamil Nadu India Department of Mathematics Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Chennai Tamil Nadu India Department of Computer Science and Engineering KCG College of Technology Chennai India

ISBN: (数字)9798331513023

ISBN: (纸本)9798331513030

This study integrates a dilated multiscale convolutional network with an encryption strategy based on reversible Hidden Communication (rHC) to present a novel framework for safe data sharing and illness diagnosis using CT images. The system meets the vital requirement for safe data exchange in healthcare by guaranteeing the security, integrity, and privacy of medical data throughout transmission. Sensitive CT scans may be encrypted and decrypted reversibly and losslessly using rHC-based encryption, which makes it appropriate for sending massive amounts of medical data over unsecure networks. We create a dilated multiscale network for illness identification, using dilated convolutions to extract fine-grained information from the CT images at several scales. This network maintains spatial resolution while broadening the receptive field, improving the detection accuracy of conditions like tumours and lesions. The framework is designed to offer end-to-end security, enabling the safe exchange of encrypted medical pictures, their decryption, and subsequent analysis for the purpose of detecting diseases. This method guarantees high detection accuracy while striking a balance between security and computational economy. The suggested concept seeks to improve data privacy, expedite medical diagnoses, and advance the development of safer and more efficient remote healthcare and telemedicine systems. Future research will concentrate on enhancing the detection model's generalisation over a variety of datasets and speeding up encryption. Proposed method ShuffleNetV3 gave good result like 95% against all the other methods.

关键词： data privacy Accuracy Computed tomography Telemedicine Medical services Encryption Convolutional neural networks Spatial resolution Medical diagnostic imaging diseases

来源：评论

学校读者我要写书评

暂无评论

Scaling Capability in Token Space: An Analysis of Large Vision Language Model

arXiv

引用

arXiv 2024年

作者： Li, Tenghui Zhou, Guoxu Zhao, Xuyang Zhao, Qibin School of Automation Guangdong University of Technology Guangzhou510006 China Tensor Learning Team RIKEN Center for Advanced Intelligence Project Tokyo103-0027 Japan Advanced Data Science Project RIKEN Information R&D and Strategy Headquarters Yokohama230-0045 Japan Department of Artificial Intelligence Medicine Graduate School of Medicine Chiba University Chiba260-0856 Japan

The scaling capability has been widely validated in neural language models with respect to the number of parameters and the size of training data. One important question is that does the scaling capability also exists similarly with respect to the number of vision tokens in large vision language Model? This study fills the gap by investigating the relationship between the number of vision tokens and the performance on vision-language models. Our theoretical analysis and empirical evaluations demonstrate that the model exhibits scalable performance S(Nl) with respect to the number of vision tokens Nl, characterized by the relationship S(Nl) ≈ (c/Nl)α. Furthermore, we also investigate the impact of a fusion mechanism that integrates the user’s question with vision tokens. The results reveal two key findings. First, the scaling capability remains intact with the incorporation of the fusion mechanism. Second, the fusion mechanism enhances model performance, particularly when the user’s question is task-specific and relevant. The analysis, conducted on fifteen diverse benchmarks spanning a broad range of tasks and domains, validates the effectiveness of the proposed approach. © 2024, CC BY.

关键词： Visual languages

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：