检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Shi, Liangliang Fan, Jack Yan, Junchi School of Artificial Intelligence Department of Computer Science and Engineering MoE Lab of AI Shanghai Jiao Tong University Shanghai China Department of Computer Science University of North Carolina Chapel Hill United States

We propose to understand Contrastive Language-Image Pretraining model (CLIP) from the Optimal Transport (OT) perspective. Specifically, we show that training of CLIP is an embodiment of inverse OT and the adopted two InfoNCE losses in CLIP correspond to a special case of bilevel optimization of a modified entropic OT. We then generalize the original CLIP loss to an OT-based loss family using variants of Regularized OT (e.g. Fused Gromov OT, unbalanced OT, etc.), and show their superior performance on public datasets for downstream tasks in both image and text domain. We also rethink the inference stage of CLIP by using the tool of OT, and propose to adopt the fused Gromov OT for (zero-shot) classification, in which the prediction is based on the graph representation whereby images and texts are nodes for graph matching. By our new technique, we show how to generalize zero-shot classification to other more flexible zero-shot tasks with competitive performance: long-tailed classification and selective classification. The former assumes the known prior distribution of labels, while in the latter case, only a subset of samples are asked to predict, yet with the need of high prediction confidence. The code is available at https://***/fan23j/ICML2024-OT-CLIP. Copyright 2024 by the author(s)

关键词： Inverse problems

来源：评论

学校读者我要写书评

暂无评论

Data-Driven Approach for Condition Monitoring and Improving Power Output of Photovoltaic Systems

引用

computers, Materials & Continua 2023年第3期74卷 5757-5776页

作者： Nebras M.Sobahi Ahteshamul Haque V S Bharath Kurukuru Md.Mottahir Alam Asif Irshad Khan Department of Electrical and Computer Engineering(ECE) King Abdulaziz UniversityJeddah 21589Saudi Arabia Advance Power Electronics Research Lab Department of Electrical EngineeringJamia Millia IslamiaNew DelhiIndia Computer Science Department Faculty of Computing and Information TechnologyKing Abdulaziz UniversityJeddah 21589Saudi Arabia

Increasing renewable energy targets globally has raised the requirement for the efficient and profitable operation of solar photovoltaic(PV)*** light of this requirement,this paper provides a path for evaluating the operating condition and improving the power output of the PV system in a grid integrated *** achieve this,different types of faults in grid-connected PV systems(GCPVs)and their impact on the energy loss associated with the electrical network are analyzed.A data-driven approach using neural networks(NNs)is proposed to achieve root cause analysis and localize the fault to the component level in the *** localized fault condition is combined with a parallel operation of adaptive neurofuzzy inference units(ANFIUs)to develop a power mismatch-based control unit(PMCU)for improving the power output of the *** develop the proposed framework,a 10-kW single-phase GCPV is simulated for training the NN-based anomaly detection approach with 14 deviation ***,the developed algorithm is combined with the PMCU implemented with the experimental setup of *** results identified 98.2%training accuracy and 43000 observations/sec prediction speed for the trained classifier,and improved power output with reduced voltage and current harmonics for the grid-connected PV operation.

关键词： Condition monitoring anomaly detection performance evaluation fault classification optimization

来源：评论

学校读者我要写书评

暂无评论

Predicting Pre-Owned Bike Values with Machine Learning Techniques in Bangladesh 2

Predicting Pre-Owned Bike Values with Machine Learning Techn...

引用

2nd IEEE International Conference on Data science and Network Security, ICDSNS 2024

作者： Ferdous, Kazi Hassanul Arman, Md.Shohel Ahmed, Kazi Rifat Tasnim, Nusrat Mahmud, Imran Revathi, P. Daffodil International University Diu Data Science Lab Department of Software Engineering Dhaka Bangladesh Holy Cross College Department of Computer Science Tamilnadu Tiruchirapalli India

ISBN: (纸本)9798350373110

As the importance of sustainable practices in the automobile sector grows, it's critical to anticipate motorcycle prices and offerings. With so many variables to consider when buying a secondhand motorcycle-condition, mileage, brand reputation, model characteristics, etc.-accurate pricing prediction is essential for both customers looking for good values and sellers hoping to maximize profits. This study investigates the effectiveness of four machine learning techniques Decision Tree models, Support Vector Regression (SVR), K-Nearest Neighbors (KNN), and Stochastic Gradient Descent (SGD) Regressions-in predicting motorcycle prices using a dataset obtained from motorcycle merchants. It's noteworthy that these methods are used with a hybrid architecture to improve prediction accuracy. Different approaches are compared and contrasted to find the best fit for the given dataset, and the difficulties and roadblocks that come with them are also discussed. After extensive testing, the decision tree model's effectiveness is shown with an amazing 97% accuracy rate, providing accurate price estimates that are essential for encouraging sustainable habits in the ever-changing Internet of Things environment. With an impressive accuracy rate, this methodology provides precise pricing estimations, which are critical for supporting sustainable practices in the automotive industry. © 2024 IEEE.

关键词： Motorcycles

来源：评论

学校读者我要写书评

暂无评论

Privacy-Aware and AI Techniques for Healthcare Based on K-Anonymity Model in Internet of Things

引用

IEEE Transactions on engineering Management 2024年 71卷 12448-12462页

作者： Sangaiah, Arun Kumar Javadpour, Amir Ja'fari, Forough Pinto, Pedro Chuang, Huan-Ming National Yunlin University of Science and Technology International Graduate Institute of Artificial Intelligence Douliou64002 Taiwan Lebanese American University Department of Electrical and Computer Engineering Byblos1102 2801 Lebanon Harbin Institute of Technology Department of Computer Science and Technology Shenzhen150001 China Instituto Politécnico de Viana Do Castelo ADiT-Lab Electrotechnics and Telecommunications Department Porto4900-347 Portugal Sharif University of Technology Department of Computer Engineering Tehran11155-9161 Iran National Yunlin University of Science and Technology Department of Information Management Yunlin640 Taiwan

The government and industry have given the recent development of the Internet of Things in the healthcare sector significant respect. Health service providers retain data gathered from many sources and are useful for patient diagnostics and research for pivotal analysis. However, sensitive personal information about a person is contained in healthcare data, which must be protected. Individual privacy protection is a crucial concern for both people and organizations, particularly when those firms must send user data to data centers due to data mining. This article investigated two general states of increasing entropy by changing the entropy of the class set of characteristics based on artificial intelligence and the k-anonymity model in privacy in context, and also three different strategies have been investigated, i.e., the strategy of 'selecting the feature with the lowest number of distinct values,' 'selecting the feature with the lowest entropy,' and 'selecting the feature with the highest entropy.' For future tasks, we can find an optimal strategy that can help us to achieve optimal entropy in the least possible repetition. The results of our work have been compared by lightweight and MH-Internet of Things, FRUIT methods and shown that the proposed method has high efficiency in entropy criteria. © 1988-2012 IEEE.

关键词： Data mining

来源：评论

学校读者我要写书评

暂无评论

Multimodal Emotion Recognition through Deep Fusion of Audio-Visual Data 26

Multimodal Emotion Recognition through Deep Fusion of Audio-...

引用

26th International Conference on computer and Information Technology, ICCIT 2023

作者： Sultana, Tamanna Jahan, Meskat Uddin, Md. Kamal Kobayashi, Yoshinori Hasan, Mahmudul Comilla University Department of Computer Science and Engineering Cumilla Bangladesh NSTU Department of Computer Science & Telecommunication Engineering Noakhali Bangladesh Saitama University Interactive Systems Lab. Saitama Japan

ISBN: (纸本)9798350359015

The field of emotion recognition in artificial intelligence focuses on enabling machines to comprehend and react to the range of emotions experienced by humans. This paper presents a novel approach that integrates the Convolution Neural Network (CNN) with audio and visual modalities. The study employs the RAVDESS database as a resource to train two distinct models for the analysis of both video and audio data. When it comes to audio pre-processing, advanced signal-processing techniques are applied to extract relevant elements and capture basic acoustic characteristics. A one-dimensional Convolutional Neural Network (CNN) architecture receives the audio data as input, enabling the model to learn complicated patterns and representations from the audio domain. In the context of video pre-processing, sophisticated algorithms are employed to extract essential facial characteristics. In order to capture the changing periods of facial expressions, the video frames are analyzed using a three-dimensional CNN framework following that they have been compressed and converted to grayscale. The fusion technique involves concatenating and extending the outputs of the audio and visual models. The fused features are subsequently sent into a softmax layer, which facilitates the development of a resilient emotion identification system. © 2023 IEEE.

关键词： Audio-visual recognition Convolution Neural Networks Emotion recognition Multi-modal fusion

来源：评论

学校读者我要写书评

暂无评论

VQR: vulnerability analysis in quadratic residues-based authentication protocols

引用

Journal of Ambient Intelligence and Humanized Computing 2023年第6期14卷 7559-7574页

作者： Ghahramani, Meysam HaddadPajouh, Hamed Javidan, Reza Kumari, Saru Computer Engineering and IT Department Shiraz University of Technology Shiraz Iran Cyber Science Lab School of Computer Science University of Guelph 50 Stone Rd E GuelphON Canada Department of Mathematics Ch. Charan Singh University Uttar Pradesh Meerut India

Ensuring security and respect for users’ privacy, especially in electronic health-care systems, is an important task that is achievable by authentication protocols. The security of many protocols is based on public-key cryptography, the breaking of which depends on solving a hard problem, e.g. Quadratic Residue (QR) which is compatible with constrained devices and is based on integer factoring with the Mean time complexity of exp((c+ o(1)) log 1 / 3(n) log 2 / 3(log (n))) for some c2(n)) to calculate user’s session keys and finding confidential user’s data by passive monitoring of the transmitted data over public networks for 2048-bit public keys in 6.9 μs. To prove the efficiency of the proposed attacks, we address 12 vulnerable QR-based authentication protocols and show that the structure of all vulnerable transmitted messages is limited to 6 cases while messages in (n,n-n) are safe from the proposed method. Also, two recent authentication protocols have been analyzed in detail to show the consequences of the proposed vulnerability. © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.

关键词： Public key cryptography

来源：评论

学校读者我要写书评

暂无评论

LEO Satellite Oriented Wildfire Detection Model Using Deep Neural Networks: A Transfer Learning Based Approach 6

LEO Satellite Oriented Wildfire Detection Model Using Deep N...

引用

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

作者： Gain, Mrityunjoy Raha, Avi Deb Biswas, Bristy Bairagi, Anupam Kumar Adhikary, Apurba Debnath, Rameswar Khulna University Computer Science and Engineering Discipline Khulna Bangladesh Ai Innovation Lab Khulna Bangladesh Noakhali Science and Technology University Department of Information and Communication Engineering Bangladesh

ISBN: (纸本)9798350385779

Wildfires are a recurring concern in several regions around the globe, often leading to catastrophic outcomes. The increasing accessibility of small satellites operating in Low Earth Orbit (LEO) has led to the potential for enhancing the speed of wildfire detection b y utilizing constellations of m any miniature satellites in the areas of interest. This research presents a system model for detecting and reporting wildfires to fire stations using LEO satellites. In this paper, we propose a transfer learning-based lightweight deep learning wildfire detection model at the base station to analyze pictures captured by LEO satellites and make predictions on the occurrence of wildfires. We use MobileNetV2 as our backbone model. MobileNetV2 is a lightweight model with low parameters. Lightweight models for image categorization offer faster inference, lower power consumption, efficient memory u sage, a nd ease of deployment, making them well-suited for real-time applications. However, a highly parameterized model can perform better than a poorly parameterized model. We apply transfer learning to ensure high performance in a low-parameterized model. The server must continuously check the images of all forest regions within a limited interval period to detect wildfires quickly. The lightweight model is very suitable for continuous image checking. Our experimental setup uses four baselines: VGG16, ResNet152, DenseNet161, and MobileNetv2. This study shows the efficacy of models in acquiring information from many sources within a given topic, as opposed to starting with no prior knowledge. We can also find from t he experimental results that our proposed low-parameterized transfer learning-based mobileNetV2 model achieves near 100 % accuracy, outperforming the other high-parameterized baselines. © 2024 IEEE.

关键词： Orbits

来源：评论

学校读者我要写书评

暂无评论

Natural Image Matting with Attended Global Context

引用

Journal of computer science & Technology 2023年第3期38卷 659-673页

作者：张億一牛力 Yasushi Makihara 张健夫赵维杰 Yasushi Yagi 张丽清 MoE Key Lab of Artificial Intelligence Department of Computer Science and EngineeringShanghai Jiao Tong University Shanghai 200240China Department of Intelligent Media Institute of Scientific and Industrial ResearchOsaka UniversityOsaka 565-0871Japan

Image matting is to estimate the opacity of foreground objects from an image. A few deep learning based methods have been proposed for image matting and perform well in capturing spatially close information. However, these methods fail to capture global contextual information, which has been proved essential in improving matting performance. This is because a matting image may be up to several megapixels, which is too big for a learning-based network to capture global contextual information due to the limit size of a receptive field. Although uniformly downsampling the matting image can alleviate this problem, it may result in the degradation of matting performance. To solve this problem, we introduce a natural image matting with the attended global context method to extract global contextual information from the whole image, and to condense them into a suitable size for learning-based network. Specifically, we first leverage a deformable sampling layer to obtain condensed foreground and background attended images respectively. Then, we utilize a contextual attention layer to extract information related to unknown regions from condensed foreground and background images generated by a deformable sampling layer. Besides, our network predicts a background as well as the alpha matte to obtain more purified foreground, which contributes to better qualitative performance in composition. Comprehensive experiments show that our method achieves competitive performance on both Composition-1k and the *** benchmark quantitatively and qualitatively.

关键词： image matting global context deformable sampling

来源：评论

学校读者我要写书评

暂无评论

Intrusion Detection in the Internet of Things Using Fusion of GRU-LSTM Deep Learning Model

引用

Intelligent Automation & Soft Computing 2023年第8期37卷 2279-2290页

作者： Mohammad S.Al-kahtani Zahid Mehmood Tariq Sadad Islam Zada Gauhar Ali Mohammed ElAffendi Department of Computer Engineering Prince Sattam Bin Abdulaziz UniversityAl-Kharj16273Saudi Arabia Department of Computer Engineering University of Engineering and TechnologyTaxila47050Pakistan The FAMLIR Group The University of LahoreLahore54000Pakistan Department of Computer Science University of Engineering&TechnologyMardan23200Pakistan Faculty of Computing International Islamic UniversityIslamabad44000Pakistan EIAS Data Science and Blockchain Lab College of Computer and Information SciencesPrince Sultan UniversityRiyadh11586Saudi Arabia

Cybersecurity threats are increasing rapidly as hackers use advanced *** a result,cybersecurity has now a significant factor in protecting organizational *** detection systems(IDSs)are used in networks to flag serious issues during network management,including identifying malicious traffic,which is a *** remains an open contest over how to learn features in IDS since current approaches use deep learning *** learning,which combines swarm intelligence and evolution,is gaining attention for further improvement against cyber *** this study,we employed a PSO-GA(fusion of particle swarm optimization(PSO)and genetic algorithm(GA))for feature selection on the CICIDS-2017 *** achieve better accuracy,we proposed a hybrid model called LSTM-GRU of deep learning that fused the GRU(gated recurrent unit)and LSTM(long short-term memory).The results show considerable improvement,detecting several network attacks with 98.86%accuracy.A comparative study with other current methods confirms the efficacy of our proposed IDS scheme.

关键词： Cyber security deep learning intrusion detection PSO-GA CICIDS-2017 intelligent system security and privacy IoT

来源：评论

学校读者我要写书评

暂无评论

Feature Selection for Peer-to-Peer Lending Default Risk Using Boruta and mRMR Approach

Feature Selection for Peer-to-Peer Lending Default Risk Usin...

引用

IEEE India Conference (INDICON)

作者： H Anusha Hegde Biswajit Bhowmik Department of Computer Science and Engineering Ishwarchandra Vidyasagar AIT Lab BRICS Lab National Institute of Technology Karnataka Mangalore Bharat

Peer-to-peer (P2P) lending in the Financial Technology (FinTech) sector is increasingly gaining attention from people where the online platform enables lenders to offer loans to borrowers. The platform as a much needed mechanism targets to reduce the risk of default and increase profitability for lenders and the platform. Each loan record maintains a variety of attributes, including details about the loan, the borrower, their credit history, their finances, and public data. If all the features are considered, the performance of the lending platform may decline. Finding the necessary characteristics more helpful in forecasting loan default is a concern. This paper investigates essential features of the P2P lending mechanism with adequate performance in lending money to individuals or businesses. We employ two algorithms to find the pertinent features: Boruta and Max-Relevance and Min-Redundancy (mRMR). Further, we use two classifiers-decision tree and XGBoost that exercise the selected elements to predict the loan defaults.

关键词：

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：