检索结果-内蒙古大学图书馆

3rd International Conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2025

作者： Srinivasa Gupta, N. Valarmathi, B. Hiwanj, Ishaan Department of Manufacturing Engineering School of Mechanical Engineering Vellore Institute of Technology Tamil Nadu Vellore India Department of Software and Systems Engineering School of Computer Science Engineering and Information Systems Vellore Institute of Technology Tamil Nadu Vellore India Department of Information Technology School of Computer Science Engineering and Information Systems Vellore Institute of Technology Tamil Nadu Vellore India

ISBN: (纸本)9798331527549

Predicting water quality is essential to preserving human health and environmental sustainability. Traditional water quality assessment methods often face scalability and real-time monitoring limitations. With accuracies of 62%, 72%, 83%, 69%, 63%, 66%, 71%, 63%, and 64%, respectively, the current techniques utilized were Logistic Regression, Decision Trees, Random Forest Regressor, Extreme Gradient Boosting, Naive Bayes, K-nearest neighbors, Support Vector Machine, AdaBoost, and Bagging [9]. This study addresses these challenges by leveraging Adaptive Synthetic Sampling (ADASYN) to balance the dataset and evaluating model performance on datasets of 5,000 and 10,000 entries per class. A robust dataset obtained from Kaggle was used, with five models – Long Short-Term Memory (LSTM), Feed Forward Neural Network (FFNN), Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost), and Random Forest – evaluated and compared. The proposed methods demonstrate significant improvements in accuracy, with XGBoost achieving the highest accuracy of 95.53%, followed by Random Forest at 93.98%. This work underscores the importance of advanced machine learning techniques in addressing the limitations of traditional methods, enhancing accuracy, scalability, and adaptability in water quality prediction. These findings contribute to advancing environmental monitoring and management practices with reliable, data-driven insights. © 2025 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

MoreGCN: Distributed IoT Service Recommendation Considering Temporal User Interest Dynamics

引用

IEEE Internet of Things Journal 2025年第10期12卷 13657-13670页

作者： Zhou, Yu Hu, Chunqiang Liu, Zewei Xing, Xiaoshuang Li, Xingwang Chongqing University School of Big Data and Software Engineering Chongqing401331 China China Southern Power Grid Joint Laboratory on Cyberspace Security Guangdong510663 China Changshu Institute of Technology School of Computer Science and Engineering Suzhou215556 China Henan Polytechnic University College of Physics and Electronic Information Engineering Jiaozuo454000 China

With the continuous development of Internet of Things (IoT), significant value has been generated, but numerous challenges remain. Recommender systems, as an effective tool to optimize IoT services, can significantly enhance user experience. However, the IoT’s demands for low latency and high-computational load make it difficult for traditional recommender systems to adapt. Moreover, traditional approaches often overlook the dynamic nature of user preferences, which are crucial for determining user satisfaction with IoT services. To address these issues, we propose a novel method called MoreGCN, which quantifies the temporal evolution of user preferences and integrates user interest modeling to accurately match similar users. This approach guides the learning of convolutional networks during the recommendation process. Deployed within a distributed computing framework and combined with meta computing, MoreGCN significantly improves computational efficiency and recommendation accuracy. Experimental results demonstrate that, across three benchmark datasets, MoreGCN consistently outperforms several existing state-of-the-art methods in terms of performance. © 2014 IEEE.

关键词： Internet of Things Recommender systems Computational modeling Accuracy Load modeling Edge computing Feature extraction Collaborative filtering Vectors Servers

来源：评论

学校读者我要写书评

暂无评论

DocTTT: Test-Time Training for Handwritten Document Recognition Using Meta-Auxiliary Learning

arXiv

引用

arXiv 2025年

作者： Gu, Wenhao Gu, Li Wang, Ziqiang Suen, Ching Yee Wang, Yang Department of Computer Science and Software Engineering Concordia University Canada

Despite recent significant advancements in Handwritten Document Recognition (HDR), the efficient and accurate recognition of text against complex backgrounds, diverse handwriting styles, and varying document layouts remains a practical challenge. Moreover, this issue is seldom addressed in academic research, particularly in scenarios with minimal annotated data available. In this paper, we introduce the DocTTT framework to address these challenges. The key innovation of our approach is that it uses test-time training to adapt the model to each specific input during testing. We propose a novel Meta-Auxiliary learning approach that combines Meta-learning and self-supervised Masked Autoencoder (MAE). During testing, we adapt the visual representation parameters using a self-supervised MAE loss. During training, we learn the model parameters using a meta-learning framework, so that the model parameters are learned to adapt to a new input effectively. Experimental results show that our proposed method significantly outperforms existing state-of-the-art approaches on benchmark datasets. Copyright © 2025, The Authors. All rights reserved.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Machine Learning Approach for Prediction of Road Accident Severity 3

An Intelligent Machine Learning Approach for Prediction of R...

引用

3rd IEEE International Conference on Industrial Electronics: Developments and Applications, ICIDeA 2025

作者： Ragam, Prashanth Rahul, Namana Sri Balaji, Pandi Jyothi, Sankati Danaveni, Pavithra School of Computer Science and Engineering VIT-AP University Andhra Pradesh Amaravati India VIT-AP University Software Development Cell Andhra Pradesh Amaravati522237 India Gokaraju Rangaraju Institute of Engineering and Technology Department of Computer Science and Engineering Telangana Hyderabad500090 India

ISBN: (纸本)9798331533205

Road accidents are a primary global concern for public safety, with India having a very high death toll. This study presents an intelligent machine learning approach to predict the severity of road accidents, contributing to the development of more effective and preventive measures. We utilized a comprehensive dataset of road accidents in India, employing advanced machine learning algorithms, including Histogram-based Gradient Boosting (HGB), Random Forest (RF), Support Vector Machines (SVM), and K-Nearest Neighbors (KNN). Our methodology involved extensive data pre-processing, feature engineering, and model optimization. The Random Forest algorithm demonstrated superior performance, achieving an accuracy of 94.7% and an F1-score of 0.947. This research provides a robust predictive model and information on the key factors influencing the accident's severity. The novelty of our approach lies in the comprehensive feature engineering tailored to the Indian context and the application of ensemble learning techniques to capture complex, nonlinear relationships in accident data. The findings of this study have significant implications for policymakers, urban planners, and law enforcement agencies in their efforts to improve road safety. © 2025 IEEE.

关键词： Highway accidents

来源：评论

学校读者我要写书评

暂无评论

引用

25th International Conference on Web Information Systems engineering, WISE 2024

作者： Tu, Longchuan Meng, Shunmei Liu, Xiao Liu, Guanfeng Beheshti, Amin Zhang, Xuyun School of Computer Science and Engineering Nanjing University of Science and Technology Nanjin China State Key Laboratory for Novel Software Technology Nanjing University Nanjin China School of Computing Macquarie University Sydney Australia

ISBN: (纸本)9789819605699

Recently, graph neural networks(GNNs) have played a key crucial in many recommendation situations. In particular, contrastive learning-based hypergraph neural networks (HGNNs) are gradually becoming a research focus for addressing issues of data sparsity and noise. Despite many studies proving their outstanding performance, there are still some shortcomings: i) Most contrastive learning-based HGNNs primarily rely on cross-view contrastive learning, while neglecting contrastive learning on the interaction graph. ii) Utilizing node embedding for hypergraph structure learning is susceptible to the influence of low-quality representations, thereby constraining the learning capability of HGNNs. To address these issues, we offer a Semantic Similarity-based Graph Contrastive Learning framework (SSGCL), which aims to jointly learn representations with rich semantic information through within-view and cross-view contrastive learning. Specifically, we first introduce consistency contrastive learning, which captures self-supervised signals through semantic similarity between nodes and their neighborhoods. It better captures the unique features of nodes through the connections between nodes and their neighborhoods. Then, we utilize interaction graph learning hypergraph structure, promoting it to extract potential node connections and thus improving its ability to describe homogeneous node relationships. Experimental evaluations on three actual datasets show that SSGCL performs much better than the current baseline models. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

When Generative AI Meets Semantic Communication: Optimizing Radio Map Construction and Distribution in Future Mobile Networks

引用

IEEE Network 2025年第3期39卷 47-55页

作者： Zhou, Li Deng, Xinfeng Ning, Zhaolong Zhao, Haitao Wei, Jibo Leung, Victor C. M. National University of Defense Technology College of Electronic Science and Technology Changsha410073 China Chongqing University of Posts and Telecommunications School of Communications and Information Engineering Chongqing400065 China Shenzhen MSU-BIT University AI Research Institute Shenzhen518000 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518000 China The University of British Columbia Department of Electrical and Computer Engineering VancouverBCV6T 1Z4 Canada

With the rapid development of the internet of things and smart cities, the demand for effective spectrum collaboration has grown *** maps play a crucial role in understanding the spatial radio environment, which is essential for wireless applications such as cell planning and radio resource management. However, generating radio maps is a resource-intensive task, as the construction process is complex and the distribution process consumes substantial bandwidth. This article proposes a method that combines semantic communication (SemCom) and generative artificial intelligence (GAI) to optimize the construction and distribution processes of radio maps. By incorporating SemCom, our method transmits only key semantic information during the distribution process, significantly reducing bandwidth usage while ensuring accurate and efficient information transfer. While the diverse generative capabilities of GAI introduce some instability, which could be a limitation in radio map construction, this article achieves precise content decoding through prompts based on multi-modal semantic information, enabling accurate construction and restoration of radio maps. By utilizing SemCom and GAI technologies, the cost of applications of radio maps can be significantly reduced, which is beneficial for stimulating the pace of intelligence evolution in future mobile networks. © 1986-2012 IEEE.

关键词： Bandwidth

来源：评论

学校读者我要写书评

暂无评论

Classifying white blood cells using combining different convolutional neural networks

引用

Multimedia Tools and Applications 2025年 1-24页

作者： Toptaş, Murat Toptaş, Buket Hanbay, Davut Software Eng. Dept. Engineering and Natural Science Faculty Bandırma Onyedi Eylül University Balıkesir Turkey Computer Eng. Dept. Engineering Faculty Inonu University Malatya44280 Turkey

White blood cells are warrior cells that protect the human body against external factors. Each of these warrior cells performs a distinct task, making every piece of information about them highly valuable in the medical field. In this article, a classification framework for the four known types of white blood cells is proposed. It is hoped that the classification of these types will contribute to the prediction of diseases such as AIDS, malaria, leukemia, and many others. In the proposed method, images of white blood cells from the Blood Cell Classification and Detection dataset were used as input to Convolutional Neural Networks. The feature vectors extracted using these Convolutional Neural Network architectures were combined into a single vector. A Minimum Redundancy Maximum Relevance algorithm was then employed to identify the most effective features within the feature vector. Experiments were conducted using these selected features, and the analysis of each experiment was reported in detail. The Support Vector Machines classifier achieved an accuracy of 98.63% in classifying white blood cell types by combining features from multiple deep learning architectures. The experimental results demonstrated that the features obtained from different layers of the Convolutional Neural Networks had varying impacts on the classification performance. © The Author(s) 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Based Open Set Recognition: A Survey

引用

Jisuanji Xuebao/Chinese Journal of computers 2025年第4期48卷 828-863页

作者： Zhang, Qin Liu, Zi-Qi Zhang, Xiao-Lin Zhang, Peng Liu, Han Chen, Xiao-Jun College of Computer Science and Software Engineering ShenZhen University Guangdong ShenZhen518060 China College of Electrical Engineering and Automation Shandong University of Science and Technology Shandong Qingdao266590 China College of Computer Science and Engineering Shandong University of Science and Technology Shandong Qingdao266590 China

In recent years, machine learning research has continuously made breakthroughs, which has led to the maturity and implementation of a large number of intelligent systems. However, the current machine learning paradigm overly relies on complete prior knowledge, and its application scenarios are limited to static closed specialized systems. The emergence of new categories is one of the main challenges in the research of dynamic open environments, to which the conventional closed-world machine learning methods can not handle it well. In order to break the excessive dependence of existing machine learning on complete category information, the research on open set recognition problem has become a new trend, which extends the traditional closed-world classification methods to open-world applications. Open set recognition problem requires the classification models classify the learned known classes correctly, and effectively identify unknown class samples that appear in the testing phase simultaneously, to avoid the large number of misclassifications. This article systematically surveys the research on open set recognition in recent years, focusing on deep learning based open set recognition methods. It introduces the classic models by sorting out them into six categories based on the main ideas they employ, and horizontally compares various research achievements, and give the comparison details of their performances. Related paper and code links are collected and available online. © 2025 science Press. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Improved Speaker-Emotion Voice Conversion with Limited Corpus based on Large Language Model and Pre-trained Model 6

Improved Speaker-Emotion Voice Conversion with Limited Corpu...

引用

6th International Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2025

作者： Fanchang TangYaLing Yadav, Amit Sah, Abhishek Pratap Khan, Asif Department of Software Engineering Chengdu Neusoft University Chengdu China College of Engineering IT and Environment Charles Darwin University NT Australia School of Management Science and Engineering University of Electronic Science and Technology of China Chengdu China Department of Computer Application Integral University Lucknow India

ISBN: (纸本)9798331522667

Existing research on speaker and emotional voice conversion often focuses on separate tasks, neglecting their joint exploration. Furthermore, the limited availability of emotional corpora for target speakers poses a significant challenge for training robustness and generalized models. This paper proposes an improved scheme for speaker-emotion voice conversion with limited target speaker's emotional corpus, integrating a large language model and a pre-trained emotional speech synthesis model. It introduces several enhancements to enhance the quality of converted speech in terms of speaker similarity and emotional expressiveness. First, emotionally tagged text is generated using a large language model and emotional speech is synthesized from this text using a fine-tuned pre-trained emotional speech synthesis model. Then, a speaker-emotion voice conversion model is co-trained with both synthesized and real target emotional speech. Finally, the model is fine-tuned with the real target emotional speech to further boost the speaker and emotional similarity. © 2025 IEEE.

关键词： Speech enhancement

来源：评论

学校读者我要写书评

暂无评论

LLM-based Methodology for Expanding 3D Metadata in Point Cloud 27

LLM-based Methodology for Expanding 3D Metadata in Point Clo...

引用

27th International Conference on Advanced Communications Technology, ICACT 2025

作者： Kim, Geon-Woo Kim, Joo-Chang Department of AI Computer Science and Engineering Kyonggi University Suwon16227 Korea Republic of Contents Convergence Software Research Institute Kyonggi University Suwon16227 Korea Republic of

ISBN: (纸本)9791188428137

This paper proposes a method for expanding the metadata of three-dimensional point cloud data using Large Language Models (LLMs). Currently, point cloud data plays a crucial role in various fields such as autonomous driving and medical image reconstruction, necessitating the expansion of metadata for efficient processing. Traditionally, metadata construction has relied on manual input, which is prone to errors. In this study, we propose a method that utilizes LLMs, particularly the Llama 3.1 model, to extract the center points of each class in the point cloud data and expand the metadata by adding these center points to the annotation files. By using center points, computational costs are reduced, and the performance of segmentation and detection models based on this data is improved. Copyright 2025 Global IT Research Institute (GIRI). All rights reserved.

关键词： Data assimilation

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：