检索结果-内蒙古大学图书馆

3rd International Conference on Algorithms, Data Mining, and Information Technology, ADMIT 2024

作者： Chen, Zhouguo Chen, Hongda Liu, Shuai Liu, Yongdong Schoolof Computer Science and Engineering Southeast University Jiangsu Nanjing China The 30th Research Institute of China Electronics Technology Group Corporation Sichuan Chengdu China School of Software Engineering Xian Jiaotong University Shaanxi Xian China China Industrial Control Systems Cyber Emergency Response Team Beijing China

ISBN: (纸本)9798400718120

Federated Learning (FL) has gained considerable attention for collaborative training in big data analysis, particularly in terms of privacy and communication constraints. Despite its promising advantages, FL faces the challenge of dealing with imbalanced label distribution on edge devices, which complicates the learning task. In literature, numerous studies have attempted to tackle this issue by introducing additional hyperparameters, which can complicate and prolong the process of model training and tuning. This complexity may potentially lead to overfitting and limited performance on new datasets. In this paper, we present a new Federated Learning framework called FedVC, which introduces virtual clients to minimize hyperparameter reliance. FedVC aims to bridge the gap between balanced and imbalanced label distribution. Compared with traditional federated learning frameworks, the proposed FedVC model designs a dynamically balanced dataset by utilizing virtual clients, which is proposed for the first time in this work. These virtual clients determine which part of the data should be used in one single training round and manage backpropagation timing for local real clients. The balanced dataset promotes greater stability in local models, thereby enhancing the overall stability of the training process. The experimental results show that FedVC outperforms classical FedAvg and the compared baseline methods for the data with imbalanced label distribution. Especially, FedVC exhibits more stable and accurate results than alternative approaches in highly heterogeneous settings. The source code is available at https://***/aikedaerC/FedVC. © 2024 Copyright held by the owner/author(s).

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

TMRE:Novel Algorithm for Computing Daily Reference Evapotranspiration Using Transformer-Based Models

引用

computers, Materials & Continua 2025年第5期83卷 2851-2864页

作者： Bushra Tayyaba Muhammad Usman Ghani Khan Talha Waheed Shaha Al-Otaibi Tanzila Saba National Center of Artificial Intelligence(NCAI) Al Khwarizmi Institute of Computer Sciences(KICS)Lahore54890Pakistan Department of Computer Science University of Engineering and Technology LahoreLahore54890Pakistan Artificial Intelligence&Data Analytics Lab Prince Sultan UniversityRiyadh11586Saudi Arabia Department of Information Systems College of Computer and Information SciencesPrincess Nourah bint Abdulrahman UniversityRiyadh11671Saudi Arabia

Reference Evapotranspiration(ETo)iswidely used to assess totalwater loss between land and atmosphere due to its importance in maintaining the atmospheric water balance,especially in agricultural and environmental *** estimation of ETo is challenging due to its dependency onmultiple climatic variables,including temperature,humidity,and solar radiation,making it a complexmultivariate time-series *** machine learning and deep learning models have been applied to forecast ETo,achieving moderate ***,the introduction of transformer-based architectures in time-series forecasting has opened new possibilities formore precise ETo *** this study,a novel algorithm for ETo forecasting is proposed,focusing on four transformer-based models:Vanilla Transformer,Informer,Autoformer,and FEDformer(Frequency Enhanced Decomposed Transformer),applied to an ETo dataset from the Andalusian *** novelty of the proposed algorithm lies in determining optimized window sizes based on seasonal trends and variations,which were then used with each model to enhance prediction *** custom window-sizing method allows the models to capture ETo’s unique seasonal patterns more ***,results demonstrate that the Informer model outperformed other transformer-based models,achievingmean square error(MSE)values of 0.1404 and 0.1445 for forecast windows(15,7)and(30,15),*** Vanilla Transformer also showed strong performance,closely following the *** findings suggest that the proposed optimized window-sizing approach,combined with transformer-based architectures,is highly effective for ETo *** novel strategy has the potential to be adapted in othermultivariate time-series forecasting tasks that require seasonality-sensitive approaches.

关键词： Reference evapotranspiration ETo transformer informer autoformer FEDformer timeseries forecasting self-attention

来源：评论

学校读者我要写书评

暂无评论

Multi-View Incongruity Learning for Multimodal Sarcasm Detection 31

Multi-View Incongruity Learning for Multimodal Sarcasm Detec...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Guo, Diandian Cao, Cong Yuan, Fangfang Liu, Yanbing Zeng, Guangjie Yu, Xiaoyan Peng, Hao Yu, Philip S. Institute of Information Engineering Chinese Academy of Sciences China School of Cyber Security University of Chinese Academy of Sciences China State Key Laboratory of Software Development Environment Beihang University China School of Computer Science and Technology Beijing Institute of Technology China Department of Computer Science University of Illinois Chicago United States

ISBN: (纸本)9798891761964

Multimodal sarcasm detection (MSD) is essential for various downstream tasks. Existing MSD methods tend to rely on spurious correlations. These methods often mistakenly prioritize non-essential features yet still make correct predictions, demonstrating poor generalizability beyond training environments. Regarding this phenomenon, this paper undertakes several initiatives. Firstly, we identify two primary causes that lead to the reliance of spurious correlations. Secondly, we address these challenges by proposing a novel method that integrate Multimodal Incongruities via Contrastive Learning (MICL) for multimodal sarcasm detection. Specifically, we first leverage incongruity to drive multi-view learning from three views: token-patch, entity-object, and sentiment. Then, we introduce extensive data augmentation to mitigate the biased learning of the textual modality. Additionally, we construct a test set, SPMSD, which consists potential spurious correlations to evaluate the the model's generalizability. Experimental results demonstrate the superiority of MICL on benchmark datasets, along with the analyses showcasing MICL's advancement in mitigating the effect of spurious correlation. © 2025 Association for Computational Linguistics.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

XMeta-OS: A Conceptual Framework for Unified Resource Management in Distributed Edge-Cloud Environments 3rd

XMeta-OS: A Conceptual Framework for Unified Resource Manag...

引用

3rd International Conference on Advances in Computing Research, ACR 2025

作者： Zacharioudakis, Eleftherios Kakoulli, Elena Evripidou, Salomi Intelligent Systems Laboratory Department of Computer Science Neapolis University Pafos Pafos Cyprus National Technical University of Ukraine Igor Sikorsky Kyiv Polytechnic Institute Kyiv Ukraine Electrical Engineering and Computer Engineering and Informatics Cyprus University of Technology Limassol Cyprus

ISBN: (纸本)9783031876462

This paper introduces XMeta-OS, a meta-operating system specially designed with Linux as its foundation to unify and optimize resource management for the distributed edge-cloud when dynamic use of GPU resources is possible. This approach offer a new GPU abstraction layer and dynamic edge-cloud allocation model that offer enhanced flexibility and resource utilization over existing systems like Kubernetes and OpenStack. By using likely sensitive data, the model applied cloud-pushed deployment model. XMeta-OS manages computer resources in real-time. Therefore, fulfilling high performance various needs. It allocates CPU and GPU in real-time to workloads requiring massive parallel processing, such as machine learning, scientific simulations, real-time data analytics, and transportation tasks. Fabrication in the laboratory of an edge-cloud environment can help to assess the scalability, fault tolerance and efficiency of the system. It showcases the overall architectural design of XMeta-OS. Changes to the kernel and security solutions for safe and effective distribution are suggested. This article presents the design of the XMeta-OS. The future of XMeta-OS entails usage and testing with which we see latency, energy, computation and other such measurements in labs. XMeta-OS is a solution to cost performance, emerging application, large-scale edge-cloud deployments, and distributed GPU-enabled infrastructures. Meta-Operating System in Edge-Cloud Continuum for dynamic resource management of GPU. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Cluster computing

来源：评论

学校读者我要写书评

暂无评论

Research on Entity Recognition Method for Quality Standards of Concrete Prefabricated Components 1

Research on Entity Recognition Method for Quality Standards ...

引用

1st International Conference on Intelligent Manufacturing and Cloud Computing, ICIMCC 2024

作者： Zhang, Tingshu Tu, Ming Wang, Gefan Chen, Bingyong Zhao, Yishi Shang, Jianga School of Computer Science China University of Geosciences Wuhang China Intelligent Construction System Research Institute China Construction Third Engineering Buerau Group Co.L Wuhang China Engineering Research Center of Natural Resource Information Management and Digital Twin Engineering Software Ministry of Education School of Computer Science China University of Geosciences Wuhang China

ISBN: (纸本)9781643685779

In recent years, the country has released a large number of standard documents related to prefabricated concrete components. Due to the dispersion and complexity of these standards, it is difficult for industry managers to implement them. Therefore, it is imperative to establish a knowledge graph in the field of quality standards for precast concrete components to achieve knowledge management. The construction of knowledge graph first requires named entity recognition, but due to the specificity of the construction industry, entity recognition of the text in this field still suffers from the problems of low recognition accuracy and high resource consumption. Aiming at the low recognition accuracy and high resource consumption in named entity recognition due to the specificity of the industry, this paper proposes a named entity recognition model ALBERT-BiLSTM-SA-CRF for the field of precast concrete component quality standard, which can capture the full text and local feature information more effectively through the combination of BiLSTM and self-attention mechanism, and accurately recognize entities in the standard. Meanwhile, using ALBERT as the word embedding layer can play the role of reducing resource consumption and improving processing efficiency. © 2025 The Authors.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Generating Targeted Universal Adversarial Perturbation against Automatic Speech Recognition via Phoneme Tailoring

Generating Targeted Universal Adversarial Perturbation again...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Zhang, Yujun Chen, Yanqu Wang, Jiakai Hu, Jin Tao, Renshuai Liu, Xianglong State Key Laboratory of Complex & Critical Software Environment Beihang University China School of Computer Science and Engineering Beihang University China College of Computer Science Beijing University of Technology China Zhongguancun Laboratory China School of Computer and Information Technology Beijing Jiaotong University China Institute of Data Space Hefei Comprehensive National Science Center China

ISBN: (纸本)9798350368741

There is a growing concern about adversarial attacks against automatic speech recognition (ASR) systems. Although research into targeted universal adversarial examples (AEs) has progressed, current methods are constrained by inefficient exploitation of audio features, demonstrating insufficient attack ability and robustness in the physical world. To solve this problem, we propose a phoneme-tailored attack (PTA) to improve the quality of the generated AEs. Specifically, to improve attack ability, we propose a Diverse Audio Composition Enrichment method, which enhances the utilization of audio features through phoneme-level slicing and recombination. To adapt AEs to complex environments, we propose a Natural Noise Pattern Guidance method to align AEs with natural noise patterns to improve their robustness. Experiments show that our method achieves an average accuracy of more than 72.34% and 98% with and without a norm constraint, and also demonstrates excellent performance in terms of generalization across datasets and resilience to MP3 compression. © 2025 IEEE.

关键词： Automatic speech recognition natural noise pattern phoneme-level targeted universal adversarial examples

来源：评论

学校读者我要写书评

暂无评论

A Comparison of Deep Learning and Machine Learning Methods for Water Quality Prediction

A Comparison of Deep Learning and Machine Learning Methods f...

引用

Intelligent Data Communication Technologies and Internet of Things (IDCIoT), International Conference on

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

ISBN: (数字)9798331527549

ISBN: (纸本)9798331527556

Predicting water quality is essential to preserving human health and environmental sustain ability. 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.

关键词： Support vector machines Adaptation models Accuracy Scalability Water quality Boosting Reliability Random forests Long short term memory Regression tree analysis

来源：评论

学校读者我要写书评

暂无评论

Universal Construction for Linearizable but Not Strongly Linearizable Concurrent Objects 10th

Universal Construction for Linearizable but Not Strongly L...

引用

10th International Symposium on Dependable software engineering: Theories, Tools and Applications, SETTA 2024

作者： Wang, Chao Wu, Peng Petri, Gustavo Jia, Qiaowen He, Youlin Lv, Yi Liu, Zhiming Centre for Research and Innovation in Software Engineering College of Computer and Information Science Southwest University Chongqing China and State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China Arm Research Cambridge United Kingdom China Great Wall Technology Group Co. Ltd. Shenzhen China University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9789819606016

Strong linearizability is a variant of linearizability and is more suitable for verification. In this paper we investigate the following two problems: (1) for which deterministic sequential specifications there exist linearizable but not strongly linearizable objects;(2) can we capture any violation scheme of strong linearizability for practical objects that are linearizable but not strongly linearizable. To deal with the first problem, we propose two classes of deterministic sequential specifications called inclusively-permutative sequential specifications and exclusively-permutative sequential specifications. The first class of sequential specifications contains sequential specifications of counter, register, max-register, snapshot, set, queue, stack and priority queue. The second class of sequential specifications contains sequential specifications of blocking queue, blocking stack and blocking priority queue. To establish these results, we propose a uniform methodology to construct a wait-free, linearizable but not strongly linearizable object for these two classes of sequential specifications. Our universal construction is based on the classical universal construction, and it additionally permits each operation to change the "take-effect-order" of previous operations. To deal with the second problem, we use bi-branch distinguishable executions to represent violations of strong linearizability. We investigate eight objects that are linearizable but not strongly linearizable, and we show that for each object, it has a pair of bi-branch distinguishable executions such that our universal construction can generate another pair of bi-branch distinguishable executions with the same set of histories. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Generalized ChatGPT-Based Collaborative Multi-Objective Decision-Making Framework for Robust Vehicle Platoon Collision Avoidance

引用

IEEE Transactions on Vehicular Technology 2025年第5期74卷 7212-7225页

作者： Zhang, Hanwen Hou, Chenfei Chen, Jicheng Zhang, Hui Wang, Fei-Yue Beihang University Hangzhou Innovation Institute Hangzhou310056 China Beihang University School of Transportation Science and Engineering Beijing100191 China Hong Kong University of Science and Technology Department of Electronic and Computer Engineering Hong Kong999077 Hong Kong University of Chinese Academy of Sciences School of Artificial Intelligence Beijing100190 China Obuda University Physiological Controls Research Center University Research and Innovation Center Budapest1034 Hungary Macau University of Science and Technology Macau Institute of Systems Engineering Macau999078 China

In this work, we propose a ChatGPT-based framework that collaborates with human expert instructions for safe and efficient vehicle platoon control. The collaborative framework integrates multi-objective decision-making to address challenges such as robust collision avoidance in complex traffic environments. To incorporate human- like reasoning, ChatGPT is firstly used to interpret human expert driving instructions, which often involve linguistic variables. It generates behavioral decisions and computes feasible trajectories for each vehicle to ensure driving safety. A distributed model predictive controller is then utilized for accurate, real-time path tracking. To efficiently handle ChatGPT response failures, a receding instruction utilization strategy is proposed. For critical response failures, hindsight experience replay is used to supplement ChatGPT's knowledge database and improve response accuracy in complex scenarios. Simulation tests and experimental validations demonstrate the framework's effectiveness and generalizability across various traffic conditions, highlighting its potential for broader applications in platoon control. © 1967-2012 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Exploratory Data Analysis and Predicting Indian Crime Against Women Using Machine Learning Approaches

Exploratory Data Analysis and Predicting Indian Crime Agains...

引用

Interdisciplinary Approaches in Technology and Management for Social Innovation (IATMSI), IEEE Conference on

作者： N. Srinivasa Gupta B. Valarmathi K. Santhi T.Chellatamilan Sameeksha Nanda Department of Manufacturing School of Mechanical Engineering Vellore Institute of Technology Vellore Tamil Nadu India Department of Software and Systems Engineering School of Computer Science Engineering and Information Systems Vellore Institute of Technology Vellore Tamil Nadu India Department of Analytics School of Computer Science and Engineering Vellore Institute of Technology Vellore Tamil Nadu India Department of Information Technology School of Computer Science Engineering and Information Systems Vellore Institute of Technology Vellore Tamil Nadu India

ISBN: (数字)9798331521691

ISBN: (纸本)9798331521707

Crime forecasting is a critical endeavor aimed at preventing future criminal actions, specifically the rising rate of violence against women in India. This study proposes the model to predict crimes committed against women using data from the National Crime Records Bureau's (NCRB) monthly crime report. The dataset encompasses district and state-wise crime activities from 2001 to 2014. The proposed work objectives are identifying factors influencing these crimes and developing predictive models using machine learning (ML). The study employs Exploratory Data Analysis (EDA), to enhance the predictive power of the models. The suggested study uses XGBoost, Random Forest (RF), and Neural Network (NN). The neural network algorithm outperforms the Random Forest and XGBoost techniques among these three. The Random Forest algorithm has the worst performance. After comparing the accuracies of various methods, such as XGBoost (92%), and Random Forest (89.84%) classifiers using imbalanced datasets (without employing Synthetic Minority Oversampling Technique (SMOTE) technique), it was found that the neural network algorithm (95.81%) delivers the highest accuracy. This study looks at model accuracies and finds that, after applying the SMOTE algorithm, neural network algorithms performed best, with an accuracy of 96.76%, outperforming the accuracies of the remaining models, which included the Random Forest (93.08%) and XG-Boost (95%).

关键词： Radio frequency Technological innovation Accuracy Machine learning algorithms Data analysis Predictive models Prediction algorithms Data models Classification algorithms Random forests

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：