检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Yang, Kaylee Yingxi Wibisono, Andre Department of Statistics and Data Science Yale University United States Department of Computer Science Yale University United States

We study the Inexact Langevin Dynamics (ILD), Inexact Langevin Algorithm (ILA), and Score-based Generative Modeling (SGM) when utilizing estimated score functions for sampling. Our focus lies in establishing stable biased convergence guarantees in terms of the Kullback-Leibler (KL) divergence. To achieve these guarantees, we impose two key assumptions: 1) the target distribution satisfies the log-Sobolev inequality (LSI), and 2) the score estimator exhibits a bounded Moment Generating Function (MGF) error. Notably, the MGF error assumption we adopt is more lenient compared to the L∞ error assumption used in existing literature. However, it is stronger than the L2 error assumption utilized in recent works, which often leads to unstable bounds. We explore the question of how to obtain a provably accurate score estimator that satisfies the MGF error assumption. Specifically, we demonstrate that a simple estimator based on kernel density estimation fulfills the MGF error assumption for sub-Gaussian target distribution, at the population level. Copyright © 2022, The Authors. All rights reserved.

关键词： Errors

来源：评论

学校读者我要写书评

暂无评论

Generative Adversarial Networks for Semi-Supervised Learning in Clinical Decision Support Systems 3rd

Generative Adversarial Networks for Semi-Supervised Learning...

引用

3rd International Conference on Intelligent Systems and Sustainable Computing, ICISSC 2023

作者： Sengan, Sudhakar Katbar, Nek Muhammad Kothalanka, Amarendra Shelke, Mayura Chinnathambi, Kalimuthan Sharma, Dilip Kumar Ammasai, Viswanathan Department of Computer Science and Engineering PSN College of Engineering and Technology Tamil Nadu Tirunelveli627451 India School of Mathematics and Statistics Central South University Hunan Changsha410083 China Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra Pradesh Vaddeswaram522502 India Department of AI & DS AISSMS IOIT Pune411001 India Deprtment of Information Technology V.S.B. College of Engineering Technical Campus Tamil Nadu Coimbatore642109 India Department of Mathematics Jaypee University of Engineering and Technology Madhya Pradesh Guna473226 India School of Computer Science and Engineering Vellore Institute of Technology Tamil Nadu Vellore632014 India

ISBN: (纸本)9789819783540

Clinical Decision Support Systems (CDSS) have emerged as crucial tools in healthcare, leveraging vast amounts of data to aid clinicians in their decision-making processes. However, the challenge of efficiently utilizing limited labelled medical data remains. This paper introduces a novel approach using Semi-supervised DC-GAN to address this challenge. Building upon the foundational principles of Generative Adversarial Networks (GANs), this proposed model augments the capabilities of traditional semi-supervised methods, optimizing their performance in the CDSS landscape. Through extensive evaluations, they compared this approach with three widely accepted semi-supervised algorithms: MixMatch, Mean Teacher, and Π-Model. The results indicate that Semi-supervised DC-GAN enhances the data utilization process and significantly improves model training, especially when working with limited labelled datasets. This research underscores the potential of GANs in revolutionizing semi-supervised learning within CDSS, paving the way for more efficient and informed clinical decisions. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Multi-View Oriented GPLVM: Expressiveness and Efficiency

arXiv

引用

arXiv 2025年

作者： Yang, Zi Li, Ying Lin, Zhidi Zhang, Michael Minyi Olmos, Pablo M. School of Computing and Data Science City University of Hong Kong Hong Kong Department of Statistics and Data Science National University of Singapore Singapore Department of Signal Theory and Communications Universidad Carlos III de Madrid Madrid Spain

The multi-view Gaussian process latent variable model (MV-GPLVM) aims to learn a unified representation from multi-view data but is hindered by challenges such as limited kernel expressiveness and low computational efficiency. To overcome these issues, we first introduce a new duality between the spectral density and the kernel function. By modeling the spectral density with a bivariate Gaussian mixture, we then derive a generic and expressive kernel termed Next-Gen Spectral Mixture (NG-SM) for MV-GPLVMs. To address the inherent computational inefficiency of the NG-SM kernel, we propose a random Fourier feature approximation. Combined with a tailored reparameterization trick, this approximation enables scalable variational inference for both the model and the unified latent representations. Numerical evaluations across a diverse range of multi-view datasets demonstrate that our proposed method consistently outperforms state-of-the-art models in learning meaningful latent representations. Copyright © 2025, The Authors. All rights reserved.

关键词： Gaussian distribution

来源：评论

学校读者我要写书评

暂无评论

Speed Bump Detection Through Inertial Sensors and Deep Learning in a Multi-contextual Analysis

引用

SN computer science 2023年第1期4卷 18页

作者： Menegazzo, Jeferson von Wangenheim, Aldo Graduate Program in Computer Science (PPGCC) Image Processing and Computer Graphics Lab (LAPIX) Brazilian National Institute for Digital Convergence (INCoD) Department of Informatics and Statistics (INE) Federal University of Santa Catarina (UFSC)-Florianópolis Santa Catarina Brazil

Speed bumps are vertical raisings of the road pavement used to force drivers to slow down to ensure greater safety in traffic. However, these obstacles have disadvantages in terms of efficiency and safety, where the presence of speed bumps can affect travel time and fuel consumption, cause traffic jams, delay emergency vehicles, and cause vehicle damage or accidents when not properly signaled. Due to these factors, the availability of geolocation information for these obstacles can benefit several applications in Intelligent Transportation System (ITS), such as Advanced Driver Assistance Systems (ADAS) and autonomous vehicles, allowing to trace more efficient routes or alert the driver of the presence of the obstacle ahead. Speed bump detection applications described in the literature employ cameras or inertial sensors, represented by accelerometers and gyroscopes. While camera-based solutions are mature with evaluation in different contextual conditions, those based on inertial sensors do not offer multi-contextual analyses, being mostly simple applications of proof of concept, not applicable in real-world scenarios. For this reason, in this work, we propose the development of a reliable speed bump detection model based on inertial sensors, capable of operating reliably in contextual variations: different vehicles, driving styles, and environments in which vehicles can travel to. For the model development and validation, we collect nine datasets with contextual variations, using three different vehicles, with three different drivers, in three different environments, in which there are three different surface types, in addition to variations in conservation state and the presence of obstacles and anomalies. The speed bumps are present in two different pavement types, asphalt and cobblestone. We use the collected data in experiments to evaluate aspects such as the influence of the placement of the sensors for vehicle data collection and the data window size. Afterwar

关键词： Autonomous vehicles Deep learning Inertial sensors Intelligent transport systems Road conditions Speed bump

来源：评论

学校读者我要写书评

暂无评论

Design Principle Transfer in Neural Architecture Search via Large Language Models

arXiv

引用

arXiv 2024年

作者： Zhou, Xun Wu, Xingyu Feng, Liang Lu, Zhichao Tan, Kay Chen Department of Data Science and Artificial Intelligence The Hong Kong Polytechnic University Hong Kong College of Computer Science Chongqing University China Department of Computer Science City University of Hong Kong Hong Kong

Transferable neural architecture search (TNAS) has been introduced to design efficient neural architectures for multiple tasks, to enhance the practical applicability of NAS in real-world scenarios. In TNAS, architectural knowledge accumulated in previous search processes is reused to warm up the architecture search for new tasks. However, existing TNAS methods still search in an extensive search space, necessitating the evaluation of numerous architectures. To overcome this challenge, this work proposes a novel transfer paradigm, i.e., design principle transfer. In this work, the linguistic description of various structural components’ effects on architectural performance is termed design principles. They are learned from established architectures and then can be reused to reduce the search space by discarding unpromising architectures. Searching in the refined search space can boost both the search performance and efficiency for new NAS tasks. To this end, a large language model (LLM)-assisted design principle transfer (LAPT) framework is devised. In LAPT, LLM is applied to automatically reason the design principles from a set of given architectures, and then a principle adaptation method is applied to refine these principles progressively based on the new search results. Experimental results show that LAPT can beat the state-of-the-art TNAS methods on most tasks and achieve comparable performance on others. Code is available at https://***/milkmilk511/LAPT. Copyright © 2024, The Authors. All rights reserved.

关键词： Architectural design

来源：评论

学校读者我要写书评

暂无评论

Enabling Protection for Critical Infrastructure through Security and Privacy in the Industrial Internet of Things 9

Enabling Protection for Critical Infrastructure through Secu...

引用

9th International Conference on Communication and Electronics Systems, ICCES 2024

作者： Yasotha, B. Vijayanandh, T. Thirumoorthy, P. Priya, M.Sangeetha Sasidaran, S. Prakalya, S.B. Srmist Department of Data Science and Business Systems Kattankulathur India School of Computing Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Department of Computer Science & Engineering Avadi Tamil Nadu Chennai India Dhanalakshmi College of Engineering Department of Electrical and Electronics Engineering Chennai India Chennai India Panimalar Engineering College Department of Information Technology Chennai600123 India Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Department of Electronics and Communication Engineering Chennai India

ISBN: (纸本)9798350377972

Industrial Internet of Things can improve critical infrastructure in energy, transportation, and manufacturing. However, IIoT device and system integration opened security weaknesses that bad actors may exploit, causing huge disruptions and financial losses. IoT network security and privacy must be prioritized to protect key infrastructure. IIoT security issues and solutions are examined in this research. First, we take a look at the state of security and the unique risks that industrial IoT devices are up against, including hacking, illegal access, and cyber-physical assaults. Then, to lessen the impact of these dangers, we investigate cutting-edge security solutions, such as authentication, encryption, and access control systems. Building a robust security architecture specifically for IIoT applications is one of the main goals of the project. To improve safety and anonymity, this system uses cutting-edge innovation like blockchain, AI, and edge computing. We also talk about how privacy-preserving methods like differential privacy and anonymization are important for keeping sensitive information safe while allowing data sharing for analysis and operations. We assess the efficacy of the suggested security protocols and procedures using a battery of case studies and simulations. Reducing the attack surface and improving the resilience of IIoT systems can be achieved by a multi-layered security approach that includes effective privacy protections. We conclude by outlining our research's larger implications, which include the importance of standardization, regulatory compliance, and stakeholder collaboration in ensuring the safe and privacy-protecting deployment of IIoT technologies in vital infrastructure. Protecting vital infrastructure from cyber threats is our ultimate goal, and our work helps build IIoT systems that are more secure and dependable. © 2024 IEEE.

关键词： Industrial internet of things (IIoT)

来源：评论

学校读者我要写书评

暂无评论

An Improved Method for Fault Detection Electric Power System Protection using SVMCNN Model 1

An Improved Method for Fault Detection Electric Power System...

引用

1st IEEE International Conference on Electronics, Computing, Communication and Control Technology, ICECCC 2024

作者： Kumar, N.M.G. Shanmuga Priya, S. Varghese, Vineetha Renuga, K. Patra, Jyoti Prasad Nishant, Neerav Mohan Babu University Sree Vidyanikethan Engineering College Rangam Pet Dept. of EEE Tirupathi India Saveetha Engineering College Department of Artificial Intelligence and Data Science Tamilnadu Chennai India Rathinam Technical campus Department of Computer Science and Engineering Tamilnadu Coimbatore India St. Joseph's Institute of Technology OMR Department of Mathematics Tamilnadu Chennai India College-KEC EE And EEE Krupajal Engineering Odisha Puri India School of Engineering Babu Banarasi Das University Department of Computer Science and Engineering Lucknow India

ISBN: (纸本)9798350371802

Multiple factors outside our control can contribute to the unanticipated breakdown of electrical power networks. It is of the utmost importance to prevent unforeseen power system issues from affecting other components of the protective system. It is the job of the protective systems to identify and categorize potentially dangerous voltage and/or current lines. To isolate the malfunctioning line, utilize the protection relay to trigger the circuit breaker. An SVM-CNN architecture is presented in this study for the purpose of detecting power outages in transmission lines. When it comes to preprocessing, feature selection, and training models, the correct sequence of activities is crucial. For anomaly-based fault detection algorithms to accurately represent and understand the data, data preprocessing is essential. Feature selection makes the process more feasible by reducing the entire data volume and helping HIF detection home in on the most relevant features. The results were evident with an accuracy improvement reaching 95.70%. © 2024 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Machine Learning Models for Customer Churn Prediction in the U.S. Banking and Financial Services: Economic Impact and Industry-Specific Insights

引用

Journal of data Analysis and Information Processing 2024年第3期12卷 388-418页

作者： Omoshola S. Owolabi Prince C. Uche Nathaniel T. Adeniken Oghenekome Efijemue Samuel Attakorah Oluwabukola G. Emi-Johnson Emmanuel Hinneh Department of Data Science Carolina University Winston-Salem NC USA Department of Computer Science Austin Peay State University Clarkesville TN USA Department of Statistical Sciences Wake Forest University Winston-Salem NC USA

Customer churn poses a significant challenge for the banking and finance industry in the United States, directly affecting profitability and market share. This study conducts a comprehensive comparative analysis of machine learning models for customer churn prediction, focusing on the U.S. context. The research evaluates the performance of logistic regression, random forest, and neural networks using industry-specific datasets, considering the economic impact and practical implications of the findings. The exploratory data analysis reveals unique patterns and trends in the U.S. banking and finance industry, such as the age distribution of customers and the prevalence of dormant accounts. The study incorporates macroeconomic factors to capture the potential influence of external conditions on customer churn behavior. The findings highlight the importance of leveraging advanced machine learning techniques and comprehensive customer data to develop effective churn prevention strategies in the U.S. context. By accurately predicting customer churn, financial institutions can proactively identify at-risk customers, implement targeted retention strategies, and optimize resource allocation. The study discusses the limitations and potential future improvements, serving as a roadmap for researchers and practitioners to further advance the field of customer churn prediction in the evolving landscape of the U.S. banking and finance industry.

关键词： Churn Prediction Machine Learning Economic Impact Industry-Specific Insights Logistic Regression Random Forest Neural Networks

来源：评论

学校读者我要写书评

暂无评论

Sustainable allocation of greenhouse gas emission permits for …rms with Leontief technologies

arXiv

引用

arXiv 2024年

作者： Gutiérrez, Elisabeth Llorca, Natividad Sánchez-Soriano, Joaquin Mosquera, Manuel A. CIO and Department of Statistics Mathematics and Computer Science University Miguel Hernández of Elche Spain Department of Statistics and Operations Research University of Vigo Spain

In this paper we deal with production situations where a cap or limit to the amount of greenhouse gas emissions permitted is imposed. Fixing a tax for each ton of pollutant emitted is also considered. We use bankruptcy rules to de…ne cooperative games with externalities associated with these situations and analyze the existence of coalitionally stable allocations of the emission permits. We prove that the constrained equal awards (CEA) rule provides stable allocations and as a direct mechanism, it is incentive compatible. These two facts have interesting managerial implications to control pollution *** Codes 91A12 (Primary) 90B30, 91A40, 91A80, 91B32 (Secondary) © 2024, CC BY-NC-ND.

关键词： Game theory

来源：评论

学校读者我要写书评

暂无评论

Thai Conversational Chatbot Classification Using BiLSTM and data Augmentation 1st

Thai Conversational Chatbot Classification Using BiLSTM and ...

引用

1st International Conference on data science and Artificial Intelligence, DSAI 2023

作者： Lhasiw, Nunthawat Tanantong, Tanatorn Sanglerdsinlapachai, Nuttapong Thammasat Research Unit in Data Innovation and Artificial Intelligence Department of Computer Science Faculty of Science and Technology Thammasat University Pathum Thani Thailand Strategic Analytics Networks with Machine Learning and AI Research Team National Electronics and Computer Technology Center Pathum Thani Thailand

ISBN: (纸本)9789819979684

Chatbot platforms, e.g., Facebook and Line, have revolutionized human interaction in the digital age. In order to develop an automatic chatbot classification, there are several challenges especially for Thai chat messages. Conversational messages are usually short and ambiguous. Therefore, it is difficult to find a dataset for constructing an effective classification model. To address the limited size of the dataset, data augmentation techniques can be possibly applied. data augmentation involves generating synthetic messages by applying various transformations to existing data samples while preserving their original meaning. In this study, the size and diversity of the dataset is increased by two methods, i.e., text augmentation using word2vec from Thai2Fit and English-Thai machine translation models proposed by VISTEC. Based on the augmented messages, a Deep Learning technique, BiLSTM, is used to construct a chatbot classification model. The experimental obtained results demonstrate that data augmentation can help to increase the classification performance. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2023.

关键词： Classification (of information)

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：