检索结果-内蒙古大学图书馆

Proceedings of the 41st International Conference on Machine Learning

作者： Nico Alvarado Hans Lobel Department of Computer Science Pontificia Universidad Católica de Chile and National Center of Artificial Intelligence Chile and Millenium Institute Foundational Research on Data Chile Department of Computer Science Pontificia Universidad Católica de Chile and National Center of Artificial Intelligence Chile and Millenium Institute Foundational Research on Data Chile and Department of Transport and Logistics Engineering Pontificia Universidad Católica de Chile

During the last few years, the field of dynamical systems has been developing innovative tools to study the asymptotic behavior of different optimizers in the context of neural networks. In this work, we redefine an extensively studied optimizer, employing classical techniques from hyperbolic geometry. This new definition is linked to a non-linear differential equation as a continuous limit. Additionally, by utilizing Lyapunov stability concepts, we analyze the asymptotic behavior of its critical points.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Unbiased Parameter Estimation for Bayesian Inverse Problems

arXiv

引用

arXiv 2025年

作者： Chada, Neil K. Jasra, Ajay Maama, Mohamed Tempone, Raul Department of Mathematics City University of Hong Kong China School of Data Science The Chinese University of Hong Kong Shenzhen CN Shenzhen China Applied Mathematics and Computational Science Program Computer Electrical and Mathematical Sciences and Engineering Division King Abdullah University of Science and Technology Thuwal23955-6900 Saudi Arabia

In this paper we consider the estimation of unknown parameters in Bayesian inverse problems. In most cases of practical interest, there are several barriers to performing such estimation, This includes a numerical approximation of a solution of a differential equation and, even if exact solutions are available, an analytical intractability of the marginal likelihood and its associated gradient, which is used for parameter estimation. The focus of this article is to deliver unbiased estimates of the unknown parameters, that is, stochastic estimators that, in expectation, are equal to the maximizer of the marginal likelihood, and possess no numerical approximation error. Based upon the ideas of [4] we develop a new approach for unbiased parameter estimation for Bayesian inverse problems. We prove unbiasedness and establish numerically that the associated estimation procedure is faster than the current state-of-the-art methodology for this problem. We demonstrate the performance of our methodology on a range of problems which include a PDE and ODE. Copyright © 2025, The Authors. All rights reserved.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Automated Plant Leaf Classification using Ensemble Transfer Learning in CNN model

Automated Plant Leaf Classification using Ensemble Transfer ...

引用

North Karnataka Subsection Flagship International Conference (NKCon), IEEE

作者： S. Yuvalatha J. Keerthika S. Prabhavathy M. Banupriya R. Priyadharshini Department of Computer Science and Business Systems Bannari Amman Institute of Technology Sathyamangalam Department of Computer Science and Engineering Sri Eshwar College of Engineering Coimbatore Department of Electronics and Communication Engineering Christ the King Engineering College Coimbatore Department of Artificial Intelligence and Data Science Hindustan Institute of Technology Coimbatore Department of Computer Science and Engineeeirng Sri Ranganathar Institute of Engineering Coimbatore

In the agriculture production, one of the biggest challenges is detecting the plant leaf disease in their early stages which significantly affects the farmer's earnings. In this paper various transfer learning CNN model of mobile NetV2, ResNet, InceptionV3, VGG16, Densenet are used to categories the potato leaf disease into healthy, Early blight, Late blight. data fragmentation strategies are applied to expand the number of images for the collected dataset to train the model. The mobileNetV2 achieve the accuracy of 86.8%. In this work the ensemble transfer learning for CNN model is proposed to identify the leaf disease. For the classification, majority voting ensemble techniques are used and obtained the accuracy of 94.8%.

关键词： Plants (biology) Transfer learning Production data models Agriculture Diseases

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

Machine Learning for Network Traffic in Home Automation IoT Systems

Machine Learning for Network Traffic in Home Automation IoT ...

引用

Communication, Computing and Energy Efficient Technologies (I3CEET), International Conference on

作者： S. P. Shingare D.Suresh Chetan K. Verma S. Lakshminarasimhan Raghunandan K R R. Navaneetha Krishnan Department of Chemical Engineering Bharati Vidyapeeth College of Engineering Navi Mumbai Department of ECE GMR Institute of Technology Andhra Pradesh India MBIT CVM University Anand Gujrat India Department Artificial Intelligence and Data Science Department of Computer Science and Engineering NMAM Institute of Technology Nitte (Deeemed to be a University) Department of Mathematics Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Chennai India

ISBN: (数字)9798331541583

ISBN: (纸本)9798331541590

IOT device security, which should protect the astute homeowner and family, has left hundreds of thousands, if not millions, of devices vulnerable to hacking. It is not practical to handle various sorts of gadgets under the same system monitoring approach since each of these devices has distinct requirements for Quality of Service (QoS). IoT Governance and Security: In combination with Device Categorization, IoT security and management may monitor device activity and automate network activities based on type or function to address concerns with rogue or susceptible devices. One of the most amazing applications of machine learning-based traffic analysis is this one, which can automatically de-anonymize devices and, as a result, reveal a variety of IoT data-hiding behaviors (as shown by an expanding body of work on the subject). This paper offers a detailed investigation into each approach to evaluate its strengths and limits. Initially, we provide a high-level procedure for classifying IoT devices and explain how various attack/fix combinations may be applied at each step of the workflow. Slides displaying specifics of the main components, like: Categorization of pubdata examination of IoT traffic data gathering situations and techniques as well. The literature on the extraction of features from IoT traffic is reviewed in Chapter 5 and includes information on related classes, comparisons to common feature types, and a brief description of a free tool that can be used to extract different types of OT or IT layered data using machine learning techniques for automatic device classification. These result in the taxonomies we have created, which enumerate patterns found in various literary corpora. Additionally, the study looks at and suggests several less-traveled paths that still need more research in this emerging sector.

关键词： Machine learning Telecommunication traffic Smart homes Quality of service Feature extraction Telecommunication computing Internet of Things Security Monitoring Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Word-Sequence Entropy: Towards Uncertainty Estimation in Free-Form Medical Question Answering Applications and Beyond

SSRN

引用

SSRN 2024年

作者： Wang, Zhiyuan Duan, Jinhao Yuan, Chenxi Chen, Qingyu Chen, Tianlong Yao, Huaxiu Zhang, Yue Wang, Ren Xu, Kaidi Shi, Xiaoshuang Center for Future Media School of Computer Science and Engineering University of Electronic Science and Technology of China Sichuan Chengdu611731 China Department of Computer Science Drexel University PhiladelphiaPA19104 United States Perelman School of Medicine University of Pennsylvania PhiladelphiaPA19104 United States Section of Biomedical Informatics & Data Science Yale School of Medicine Yale University New HavenCT06510 United States Department of Computer Science University of North Carolina at Chapel Hill Chapel HillNC27599 United States Department of Electrical and Computer Engineering Illinois Institute of Technology ChicagoIL60616 United States

Uncertainty estimation is crucial for the reliability of safety-critical human and artificial intelligence interaction systems, particularly in the healthcare engineering domain. However, a general method for quantifying the uncertainty of free-form answers has yet to be established in open-ended medical question-answering (QA) tasks, where irrelevant words and sequences with limited semantic information can be the primary source of uncertainty due to the presence of generative inequality. In this paper, we propose the Word-Sequence Entropy (WSE ), which calibrates the uncertainty proportion at both the word and sequence levels according to the semantic relevance, with greater emphasis placed on keywords and more relevant sequences when performing uncertainty quantification. We compare WSE with 6 baseline methods on 5 free-form medical QA datasets, utilizing 7 "off-the-shelf" large language models (LLMs), and show that WSE exhibits superior performance on accurate uncertainty measurement under two standard criteria for correctness evaluation. Additionally, in terms of the potential for real-world medical QA applications, we achieve a significant enhancement (e.g., +6.36% improvement of model accuracy on the COVID-QA dataset) in the performance of LLMs when employing sequences with lower uncertainty that are identified by WSE as final answers, without requiring additional task-specific fine-tuning or architectural modifications. © 2024, The Authors. All rights reserved.

关键词： Uncertainty analysis

来源：评论

学校读者我要写书评

暂无评论

The Quest for Fairness: A Comparative Study of Accuracy in AI Hiring Systems

The Quest for Fairness: A Comparative Study of Accuracy in A...

引用

Innovation in Technology (ASIANCON), Asian Conference on

作者： Jason D'souza Vedant Kadam Pranali Shinde Kumkum Saxena Information Technology Department Thadomal Shahani Engineering College Mumbai India Artificial Intelligence & Data Science Department Thadomal Shahani Engineering College Mumbai India Computer Engineering Department Thadomal Shahani Engineering College Mumbai India

Artificial Intelligence and Machine Learning based models are increasingly being used in developing models for automating recruitment systems in various Multinational Companies as well as small-scale and developing companies and industries. With the advent of Generative Artificial Intelligence models, many companies are turning towards automatic and AI-based hiring software. J.P. Morgan uses HireVue, Deloitte incorporates Hiretual for talent sourcing and engagement, IBM incorporates *** for automated interview scheduling and many more. But in all these systems a major factor to consider is the biasness in the models and how to detect them and efficiently mitigate them. Some models unwillingly sometimes amplify the biases in the models and results in faulty and biased predictions which can lead to the inaccurate matching of profiles to the job description and result in wayward hiring processes. The biasness can be based on the parameters such as gender, address, etc., or biasness transferred from the programmer into the model while creating the model which is called unconscious bias. The models implemented in the research are Random Forest, Decision Tree, KNN Tree, AdaBoost, and XGBoost. The accuracies for the models Random Forest, Decision Tree, KNN Tree, AdaBoost, and XGBoost were 0.80, 0.81, 0.81, 0.78, and 0.84 respectively.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DCNN: Dual Cross-current Neural Networks Realized Using An Interactive Deep Learning Discriminator for Fine-grained Objects

arXiv

引用

arXiv 2024年

作者： Fu, Da Rong, Mingfei Kim, Eun-Hu Huang, Hao Pedrycz, Witold Shandong University Department of Computer Science and Technology Shandong Weihai264200 China Linyi University Department of Big Data Shandong Linyi276000 China Suwon University Department of Computer Science Suwon Gyeonggi-do Hwaseong-si18323 Korea Republic of Alberta University Department of Electrical & Computer Engineering EdmontonT6G 2H5 Canada

Accurate classification of fine-grained images remains a challenge in backbones based on convolutional operations or self-attention mechanisms. This study proposes novel dual-current neural networks (DCNN), which combine the advantages of convolutional operations and self-attention mechanisms to improve the accuracy of fine-grained image classification. The main novel design features for constructing a weakly supervised learning backbone model DCNN include (a) extracting heterogeneous data, (b) keeping the feature map resolution unchanged, (c) expanding the receptive field, and (d) fusing global representations and local features. Experimental results demonstrated that using DCNN as the backbone network for classifying certain fine-grained benchmark datasets achieved performance advantage improvements of 13.5-19.5% and 2.2-12.9%, respectively, compared to other advanced convolution or attention-based fine-grained backbones. © 2024, CC BY-NC-ND.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Efficiency ordering of stochastic gradient descent 22

Efficiency ordering of stochastic gradient descent

引用

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Jie Hu Vishwaraj Doshi Do Young Eun Department of Electrical and Computer Engineering North Carolina State University Data Science and Advanced Analytics IQVIA

ISBN: (纸本)9781713871088

We consider the stochastic gradient descent (SGD) algorithm driven by a general stochastic sequence, including i.i.d noise and random walk on an arbitrary graph, among others; and analyze it in the asymptotic sense. Specifically, we employ the notion of 'efficiency ordering', a well-analyzed tool for comparing the performance of Markov Chain Monte Carlo (MCMC) samplers, for SGD algorithms in the form of Loewner ordering of covariance matrices associated with the scaled iterate errors in the long term. Using this ordering, we show that input sequences that are more efficient for MCMC sampling also lead to smaller covariance of the errors for SGD algorithms in the limit. This also suggests that an arbitrarily weighted MSE of SGD iterates in the limit becomes smaller when driven by more efficient chains. Our finding is of particular interest in applications such as decentralized optimization and swarm learning, where SGD is implemented in a random walk fashion on the underlying communication graph for cost issues and/or data privacy. We demonstrate how certain non-Markovian processes, for which typical mixingtime based non-asymptotic bounds are intractable, can outperform their Markovian counterparts in the sense of efficiency ordering for SGD. We show the utility of our method by applying it to gradient descent with shuffling and mini-batch gradient descent, reaffirming key results from existing literature under a unified framework. Empirically, we also observe efficiency ordering for variants of SGD such as accelerated SGD and Adam, open up the possibility of extending our notion of efficiency ordering to a broader family of stochastic optimization algorithms.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep Learning and its Applications: A Novel Approach to Machine Learning with Encoding Algorithms

Deep Learning and its Applications: A Novel Approach to Mach...

引用

Intelligent & Innovative Practices in engineering & Management (IIPEM), International Conference on

作者： Kamalesh Thoppaen Suresh Babu Bhagyashree Ashok Tingare Vijayshri Khedkar Tarun Dhar Diwan P. William Abhishek Badholia Globallogic Inc Santa Clara CA USA Department of Artificial Intelligence and Data Science D Y Patil College of Engineering Akurdi Pune Department of Computer Science & Engineering Symbiosis Institute of Technology Symbiosis International (Deemed University) Pune India Assistant Professor and Controller of Examination (COE) Atal Bihari Vajpayee University Bilaspur India Deaprtment of Information Technology Sanjivani College of Engineering Kopargaon Amity University Dubai UAE Department of Data Science Shri Shankaracharya Institute of Professional Management and Technology Raipur

ISBN: (数字)9798350390049

ISBN: (纸本)9798350390056

Deep learning has revolutionized data processing by recognizing complicated patterns using multi-layered neural networks. Encoding complex data like hyperspectral photos may help deep learning models perform better. Encoding shapes raw data for compression, dimensionality reduction, and feature extraction. These methods matter. This research found that higher-dimensional encodings can improve deep learning model accuracy and convergence. Complex encoding algorithms like FE-16D and BLOSUM-20D outperform simpler ones in validation accuracy and AUC. To prove this, extensive testing and comparison were done. The study emphasizes the necessity to combine deep learning frameworks with improved encoding techniques to improve machine learning and large-scale data processing.

关键词： Deep learning Dimensionality reduction Machine learning algorithms Computational modeling Neural networks Feature extraction data processing Encoding data models Hyperspectral imaging

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：