检索结果-内蒙古大学图书馆

Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems

作者： Michel Hohendanner Chiara Ullstein Bukola Abimbola Onyekwelu Amelia Katirai Jun Kuribayashi Olusola Babalola Arisa Ema Jens Grossklags Department of Computer Science Technical University of Munich Munich Germany and Munich Center for Digital Sciences and AI Hochschule München University of Applied Sciences Munich Germany Department of Computer Science Technical University of Munich Munich Germany Department of Mathematics and Computer Science Elizade University Ilara-Mokin Ondo State Nigeria Research Center on Ethical Legal and Social Issues Osaka University Osaka Japan Tokyo College The University of Tokyo Tokyo Japan Department of Mathematics and Computer Science Elizade University Ilara-Mokin Nigeria

来源：评论

学校读者我要写书评

暂无评论

Reading the Readers Mind through Eye Tracking: Can AI Generated Texts Match Human Authors? 25

Reading the Readers Mind through Eye Tracking: Can AI Genera...

引用

Proceedings of the 2025 Symposium on Eye Tracking Research and Applications

作者： Chaudhary Muhammad Aqdus Ilyas Sifat-E Noor Ashkan Tashk Bart Cooreman Sofie Beier Per Bækgaard Department of Applied Mathematics and Computer Science Section for Cognitive Systems Technical University of Denmark Kgs. Lyngby Denmark Technical University of Denmark Kgs. Lyngby Denmark Department of Applied Mathematics and Computer Science Section for Cognitive Systems Denmark Technical University Kgs. Lyngby Denmark Centre for Visibility Design Royal Danish Academy Copenhagen Denmark Royal Danish Academy Copenhagen Denmark

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Based Prediction of Heavy Rainfall Days in the Northern Himalayas: A Case Study of Shimla, Srinagar and Kargil

Machine Learning-Based Prediction of Heavy Rainfall Days in ...

引用

Cognitive Computing in Engineering, Communications, sciences and Biomedical Health Informatics (IC3ECSBHI), International Conference on

作者： Harshita Singh Bhumika Rana Sushil Kumar Anurag Singh Baghel Jagabandhu Panda Nikhil Singh Department of Computer Science Gautam Buddha University Greater Noida India Department of Applied Mathematics Gautam Buddha University Greater Noida India Department of Earth and Atmospheric Sciences National Institute of Technology Rourkela India

ISBN: (数字)9798331518523

ISBN: (纸本)9798331518530

In regions vulnerable to heavy rainfall, such as the north Himalayan region of India, which includes the states of Himachal Pradesh, Jammu & Kashmir and Union Territory of Ladakh, an effective disaster management and ecosystem protection are especially important. Forecasting days with heavy rainfall accurately is crucial because these occurrences not only interfere with agricultural operations, but also cause landslides and flash floods, which cause serious damage to the environment and a large loss of life and property. Although traditional and conventional numerical weather prediction (NWP) models are widely used, problems with accuracy and operational efficiency are common with such models. The present study incorporates a machine learning technique known as the Long Short-Term Memory (LSTM) networks, to analyze long-term nonlinear time series rainfall data and to encapsulate temporal dependencies in the data. The study uses daily climatological rainfall data from the Indian Meteorological department in Pune for a period of 123 years (1901 to 2023). The extensive data set has enabled the LSTM model to be trained to forecast rainfall categories such as Rather Heavy Rain and Heavy Rain, as specified by the India Meteorological department (IMD). This study also investigates the model's ability to forecast days with relatively rather heavy and heavy rainfall in the districts of Shimla, Srinagar and Kargil of Himachal Pradesh, Jammu & Kashmir and Ladakh respectively. With an emphasis on the model's ability to predict days with rather heavy and heavy rainfall, this study also evaluates the model's performance using a variety of statistical measures including Mean Absolute Error(MAE), Root Mean Square Error(RMSE) and Coefficient of determination(R 2 ) error. The encouraging findings of the present study implies that LSTM may prove to be a good substitute for traditional techniques in the North Himalayan region when it comes to rainfall forecasting. The present study add

关键词： Adaptation models Rain Biological system modeling Weather forecasting Machine learning Predictive models Market research Numerical models Jamming Long short term memory

来源：评论

学校读者我要写书评

暂无评论

InfoGain Wavelets: Furthering the Design of Diffusion Wavelets for Graph-Structured Data

arXiv

引用

arXiv 2025年

作者： Johnson, David R. Krishnaswamy, Smita Perlmutter, Michael Program in Computing Boise State University BoiseID United States Department of Computer Science Department of Genetics Yale University New HavenCT United States Department of Mathematics Program in Computing Boise State University BoiseID United States

Diffusion wavelets extract information from graph signals at different scales of resolution by utilizing graph diffusion operators raised to various powers, known as diffusion scales. Traditionally, the diffusion scales are chosen to be dyadic integers, 2j. Here, we propose a novel, unsupervised method for selecting the diffusion scales based on ideas from information theory. We then show that our method can be incorporated into wavelet-based GNNs via graph classification experiments. Copyright © 2025, The Authors. All rights reserved.

关键词： Wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Improved Compact Genetic Algorithms with Efficient Caching

arXiv

引用

arXiv 2025年

作者： Dutta, Prasanta Mukhopadhyay, Anirban Physics and Applied Mathematics Unit Indian Statistical Institute West Bengal Kolkata700108 India Department of Computer Science and Engineering University of Kalyani West Bengal Kalyani741235 India

Compact Genetic Algorithms (cGAs) are condensed variants of classical Genetic Algorithms (GAs) that use a probability vector representation of the population instead of the complete population. cGAs have been shown to significantly reduce the number of function evaluations required while producing outcomes similar to those of classical GAs. However, cGAs have a tendency to repeatedly generate the same chromosomes as they approach convergence, resulting in unnecessary evaluations of identical chromosomes. This article introduces the concept of caching in cGAs as a means of avoiding redundant evaluations of the same chromosomes. Our proposed approach operates equivalently to cGAs, but enhances the algorithm’s time efficiency by reducing the number of function evaluations. We also present a data structure for efficient cache maintenance to ensure low overhead. The proposed caching approach has an asymptotically constant time complexity on average. The proposed method further generalizes the caching mechanism with higher selection pressure for elitism-based cGAs. We conduct a rigorous analysis based on experiments on benchmark optimization problems using two well-known cache replacement strategies. The results demonstrate that caching significantly reduces the number of function evaluations required while maintaining the same level of performance accuracy. Copyright © 2025, The Authors. All rights reserved.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Extracting Commuters from Automated Road Traffic Counters: A Gaussian Mixture Approach

引用

Data science for Transportation 2025年第2期7卷 1-28页

作者： Brustad, Hilde Kjelgaard Midtbø, Jørgen E. Tomba, Gianpaolo Scalia Kivelä, Mikko Gregersen, Fredrik Alexander Alessandretti, Laura Oslo Center for Biostatistics and Epidemiology Oslo University Hospital Oslo Norway Department of Method Development and Analytics Norwegian Institute of Public Health Oslo Norway Department of Mathematics Stockholm University Stockholm Sweden Universita’ di Roma Tor Vergata Rome Italy Department of Computer Science Aalto University Espoo Finland Institute of Transportation economics Gaustadalleén 21 Oslo 0349 Norway Department of Applied Mathematics and Computer Science Technical University of Denmark Kongens Lyngby 2800 Denmark

Assessing traffic patterns is important for many applications such as rush hour traffic management, cross-border commuting statistics, transportation disruption assessment, and crisis management. We present a method for detecting commuting patterns from time-detailed traffic sensor data. Our method uses Gaussian mixture models to identify morning peaks that also exhibit expected variation patterns over weekends and holidays as corresponding to commuting. We apply the method to detect the variation in commuting between countries in the Nordics during the disruptions caused by the COVID-19 pandemic. Results show that the commuting traffic experienced a smaller decrease (42–71%) than the total traffic (87–92%) during the pandemic. For Finland and Sweden, both types of traffic have in 2023 returned to approximately the same level as before the pandemic, while the traffic between Norway and Sweden has only recovered to about 73% of the pre-pandemic level. Our methods can be applied in real-time to provide useful information for applications. © The Author(s) 2025.

关键词： Automated traffic counters Commuting patterns COVID-19 Cross-border commuting Gaussian mixture model Nordic countries Road traffic data

来源：评论

学校读者我要写书评

暂无评论

Solving Inverse Problems with Deep Linear Neural Networks: Global Convergence Guarantees for Gradient Descent with Weight Decay

arXiv

引用

arXiv 2025年

作者： Laus, Hannah Parkinson, Suzanna Charisopoulos, Vasileios Krahmer, Felix Willett, Rebecca Department of Mathematics Technical University of Munich Germany Germany Committee on Computational and Applied Mathematics University of Chicago United States Data Science Institute University of Chicago United States Technical University of Munich Germany Department of Computer Science University of Chicago United States Department of Statistics University of Chicago United States

Machine learning methods are commonly used to solve inverse problems, wherein an unknown signal must be estimated from few measurements generated via a known acquisition procedure. In particular, neural networks perform well empirically but have limited theoretical guarantees. In this work, we study an underdetermined linear inverse problem that admits several possible solution mappings. A standard remedy (e.g., in compressed sensing) establishing uniqueness of the solution mapping is to assume knowledge of latent low-dimensional structure in the source signal. We ask the following question: do deep neural networks adapt to this low-dimensional structure when trained by gradient descent with weight decay regularization? We prove that mildly overparameterized deep linear networks trained in this manner converge to an approximate solution that accurately solves the inverse problem while implicitly encoding latent subspace structure. To our knowledge, this is the first result to rigorously show that deep linear networks trained with weight decay automatically adapt to latent subspace structure in the data under practical stepsize and weight initialization schemes. Our work highlights that regularization and overparameterization improve generalization, while overparameterization also accelerates convergence during training. © 2025, CC BY.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Blockchain Empowered Secure Federated Learning for Consumer IoT Applications in Cloud-Edge Collaborative Environment

引用

IEEE Transactions on Consumer Electronics 2025年

作者： Kumar, Mohit Samriya, Jitendra Kumar Walia, Guneet Kaur Verma, Prabal Wu, Huaming Gill, Sukhpal Singh National Institute of Technology Department of Information Technology Jalandhar India Indian Institute of Information Technology Computer Science and Engineering Sonepat India National Institute of Technology Department of Information Technology Srinagar India Tianjin University Center for Applied Mathematics Tianjin China Queen Mary University of London School of Electronic Engineering and Computer Science United Kingdom

The growing number of consumer Internet of Things (IoT) gadgets, including smart homes, fitness trackers, connected appliances, and home security systems, is transforming the way we live our daily lives. This has led to the emergence of a collaborative cloud-edge paradigm to leverage resources and services near the end-user, thereby providing prompt response to delay-sensitive real-time applications. Nevertheless, the tremendous amount of data generated by various IoT devices and sent over the network is always an open security challenge. The introduction of Federated Learning (FL) addresses the security and data privacy shortcomings of traditional centralised machine learning. Despite FL's use for data privacy, it must overcome a number of significant challenges, such as privacy concerns, communication overhead, stragglers, and heterogeneity. To solve these challenges, this paper proposes a novel technique for enhancing security in IoT-enabled edge cloud computing networks, utilising blockchain-driven FL and Gaussian Bayesian transfer convolutional neural network architectures for data analysis. Blockchain-driven FL ensures the security and privacy of consumer IoT applications. In comparison to state-of-the-art works, the experimental results achieved throughput of up to 89%, latency of 71%, training accuracy of 91%, validation accuracy of 96%, and network security of 92%. © 1975-2011 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Symmetry Resolved Entanglement with Up1q Symmetry: Some Closed Formulae for Excited States

arXiv

引用

arXiv 2025年

作者： Castro-Alvaredo, Olalla A. Santamaría-Sanz, Lucía Department of Mathematics City University of London 10 Northampton Square EC1V 0HB United Kingdom Department of Applied Mathematics and Computer Science University of Burgos Hospital del Rey s/n Burgos09001 Spain

In this work, we revisit a problem we addressed in previous publications with various collaborators, that is, the computation of the symmetry resolved entanglement entropies of zero-density excited states in infinite volume. The universal nature of the charged moments of these states has already been noted previously. Here, we investigate this problem further, by writing general formulae for the entropies of excited states consisting of an arbitrary number of subsets of identical excitations. When the initial state is written in terms of qubits with appropriate probabilistic coefficients, we find the final formulae to be of a combinatorial nature too. We analyse some of their features numerically and analytically and find that for qubit states consisting of particles of the same charge, the symmetry resolved entropies are independent of region size relative to system size, even if the number and configuration entropies are not. Copyright © 2025, The Authors. All rights reserved.

关键词： Qubits

来源：评论

学校读者我要写书评

暂无评论

Uncertainty-Aware Graph Self-Training with Expectation-Maximization Regularization

arXiv

引用

arXiv 2025年

作者： Wang, Emily Chen, Michael Li, Chao School of Information Technology Lakeside College Ohio United States Department of Applied Mathematics North River University Manchester United Kingdom Department of Computer Science Eastern Asia Institute of Technology Beijing China

In this paper, we propose a novel uncertainty-aware graph self-training approach for semi-supervised node classification. Our method introduces an Expectation-Maximization (EM) regularization scheme to incorporate an uncertainty mechanism during pseudo-label generation and model retraining. Unlike conventional graph self-training pipelines that rely on fixed pseudo-labels, our approach iteratively refines label confidences with an EM-inspired uncertainty measure. This ensures that the predictive model focuses on reliable graph regions while gradually incorporating ambiguous nodes. Inspired by prior work on uncertainty-aware self-training techniques [5], our framework is designed to handle noisy graph structures and feature spaces more effectively. Through extensive experiments on several benchmark graph datasets, we demonstrate that our method outperforms strong baselines by a margin of up to 2.5% in accuracy while maintaining lower variance in performance across multiple runs. Copyright © 2025, The Authors. All rights reserved.

关键词： Expectation maximization algorithm

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：