检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Yao, Chaoyu Fan, Wanshu Wang, Cong Yang, Xin Wei, Xiaopeng Zhou, Dongsheng School of Computer Science and Technology Dalian University of Technology China National and Local Joint Engineering Laboratory of Computer Aided Design School of Software Engineering Dalian University China Department of Computing The Hong Kong Polytechnic University China

In Unsupervised Domain Adaptation Semantic Segmentation (UDASS), while self-training techniques have become one of the most effective methods to date, the absence of target labels makes models susceptible to overfitting. To address this problem, consistency techniques regularize the model by perturbing the target domain and enforcing consistent pixel predictions before and after the perturbation. Consistency techniques are based on the idea that a well-performing model should yield consistent predictions for the target data stream before and after perturbations. In this paper, we introduce the Feature Dropout Consistency (FDC) module, a novel approach that instigates perturbations at the feature level. This is achieved by implementing random feature dropout between the encoder and decoder of the student network. Subsequently, a feature dropout consistency loss is applied to minimize discrepancies in predictions between the perturbed student network and its teacher counterpart. Notably, FDC seamlessly integrates with existing self-training methodologies and, when paired with input perturbation, delves into an even broader perturbation spectrum. Through rigorous experimentation in standard UDA settings—training with synthetic labeled and real unlabeled data—FDC consistently outperforms baseline models, establishing new benchmarks on GTAV →Cityscapes and SYNTHIA →Cityscapes. © 2024, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Feature Correlation-Guided Knowledge Transfer for Federated Self-Supervised Learning

引用

IEEE Transactions on Neural Networks and Learning Systems 2025年第6期36卷 10544-10557页

作者： Yi Liu Song Guo Jie Zhang Yufeng Zhan Qihua Zhou Yingchun Wang Department of Computing The Hong Kong Polytechnic University Hong Kong SAR China Department of Computer Science and Engineering Hong Kong University of Science and Technology Hong Kong SAR China Beijing Institute of Technology Beijing China College of Computer Science and Software Engineering Shenzhen University Shenzhen China Huawei Technologies Company Ltd. Hangzhou China

Extensive attention has been paid to the application of self-supervised learning (SSL) approaches on federated learning (FL) to tackle the label scarcity problem. Previous works on federated SSL (FedSSL) generally fall into two categories: parameter-based model aggregation or data-based feature sharing to achieve knowledge transfer among multiple unlabeled clients. Despite the progress, they inevitably rely on some assumptions, such as homogeneous models or the existence of an additional public dataset, which hinder the universality of the training frameworks for more general scenarios (e.g., unlabeled clients with heterogeneous models). Therefore, in this article, we propose a novel and general method named federated self-supervised learning with feature-c orrelation-based aggregation (FedFoA) to tackle the above limitations. By exchanging feature correlation instead of model parameters or feature mappings, our approach reduces the discrepancies of local representations learning processes, thus promoting collaboration between heterogeneous clients. A factorization-based method is designed to extract the cross-feature relation matrix from local representations, which serves as a knowledge medium for the aggregation phase. We demonstrate that FedFoA is a heterogeneity-supportive and privacy-preserving training framework and can be easily compatible with state-of-the-art FedSSL methods. Extensive empirical experiments demonstrate our proposed approach outperforms the state-of-the-art methods by a significant margin.

关键词： Training Feature extraction Matrix decomposition Correlation Self-supervised learning Knowledge transfer Data models Vectors Electronic mail Collaboration

来源：评论

学校读者我要写书评

暂无评论

Social Media and Stock Market Prediction: A Big Data Approach

引用

Computers, Materials & Continua 2021年第5期67卷 2569-2583页

作者： Mazhar Javed Awan Mohd Shafry Mohd Rahim Haitham Nobanee Ashna Munawar Awais Yasin Azlan Mohd Zain Azlanmz School of Computing Faculty of EngineeringUniversity Teknologi MalaysiaJohorMalaysia Department of Software Engineering University of Management and TechnologyLahorePakistan Collage of Business Abu Dhabi UniversityAbu DhabiUnited Arab Emirates Oxford Center for Islamic Studies The University of OxfordOxfordUK The University of Liverpool Management School The University of LiverpoolLiverpoolUK Department of Computer Engineering National University of TechnologyIslamabadPakistan School of Computing UTM Big Data CentreUniversiti Teknologi MalaysiaJohorMalaysia

Big data is the collection of large datasets from traditional and digital sources to identify trends and *** quantity and variety of computer data are growing exponentially for many *** example,retailers are building vast databases of customer sales *** are working on logistics financial services,and public social media are sharing a vast quantity of sentiments related to sales price and *** of big data include volume and variety in both structured and unstructured *** this paper,we implemented several machine learning models through Spark MLlib using PySpark,which is scalable,fast,easily integrated with other tools,and has better performance than the traditional *** studied the stocks of 10 top companies,whose data include historical stock prices,with MLlib models such as linear regression,generalized linear regression,random forest,and decision *** implemented naive Bayes and logistic regression classification *** results suggest that linear regression,random forest,and generalized linear regression provide an accuracy of 80%-98%.The experimental results of the decision tree did not well predict share price movements in the stock market.

关键词： Big data analytics artificial intelligence machine learning stock market social media business analytics

来源：评论

学校读者我要写书评

暂无评论

How Does Forecasting Affect the Convergence of DRL Techniques in O-RAN Slicing?

arXiv

引用

arXiv 2023年

作者： Nagib, Ahmad M. Abou-Zeid, Hatem Hassanein, Hossam S. School of Computing Queen’s University Canada Department of Electrical and Software Engineering University of Calgary Canada Faculty of Computers and Artificial Intelligence Cairo University Egypt

The success of immersive applications such as virtual reality (VR) gaming and metaverse services depends on low latency and reliable connectivity. To provide seamless user experiences, the open radio access network (O-RAN) architecture and 6G networks are expected to play a crucial role. RAN slicing, a critical component of the O-RAN paradigm, enables network resources to be allocated based on the needs of immersive services, creating multiple virtual networks on a single physical infrastructure. In the O-RAN literature, deep reinforcement learning (DRL) algorithms are commonly used to optimize resource allocation. However, the practical adoption of DRL in live deployments has been sluggish. This is primarily due to the slow convergence and performance instabilities suffered by the DRL agents both upon initial deployment and when there are significant changes in network conditions. In this paper, we investigate the impact of time series forecasting of traffic demands on the convergence of the DRL-based slicing agents. For that, we conduct an exhaustive experiment that supports multiple services including real VR gaming traffic. We then propose a novel forecasting-aided DRL approach and its respective O-RAN practical deployment workflow to enhance DRL convergence. Our approach shows up to 22.8%, 86.3%, and 300% improvements in the average initial reward value, convergence rate, and number of converged scenarios respectively, enhancing the generalizability of the DRL agents compared with the implemented baselines. The results also indicate that our approach is robust against forecasting errors and that forecasting models do not have to be ideal. Copyright © 2023, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A First Look at Information Highlighting in Stack Overflow Answers

SSRN

引用

SSRN 2023年

作者： Ahmed, Shahla Shaan Wang, Shaowei Tian, Yuan Chen, Tse-Hsun Zhang, Haoxiang Department of Computer Science University of Manitoba Canada School of Computing Queen’s University Canada Department of Computer Science and Software Engineering Concordia University Canada Huawei Canada Canada

Navigating the knowledge on Stack Overflow (SO) remains challenging. To make the posts vivid to users, SO allows users to write and edit posts with Markdown or HTML, so that users can leverage various formatting styles (e.g., bold, italic, and code) to highlight the important information. We carry out the first large-scale exploratory study on the information highlighting in SO answers in our recent study. To extend our previous study, we develop approaches to recommend highlighted content automatically by using neural network architectures initially designed for Named Entity Recognition task. We train models for each type of formatting (i.e., Bold, Italic, Code, and Heading) using the information highlighting dataset we collected from SO answers. Our models achieve precision values ranging from 0.71 to 0.82. While the recall values are much lower than precision values, the model for automatic code content highlighting achieves a recall of 0.73 and an F1 score of 0.75, outperforming the others. Our analysis of failure cases indicates that the majority of the failure cases are missing identification (i.e., the model misses the content that is supposed to highlight) due to that models tend to learn the frequent highlighted words while struggling to learn less frequent words. © 2023, The Authors. All rights reserved.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Container lifecycle-aware scheduling for serverless computing

Container lifecycle-aware scheduling for serverless computin...

引用

作者： Wu, Song Tao, Zhiheng Fan, Hao Huang, Zhuo Zhang, Xinmin Jin, Hai Yu, Chen Cao, Chun National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China State Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University Nanjing China

Elastic scaling in response to changes on demand is a main benefit of serverless computing. When bursty workloads arrive, a serverless platform launches many new containers and initializes function environments (known as cold starts), which incurs significant startup latency. To reduce cold starts, platforms usually pause a container after it serves a request, and reuse this container for subsequent requests. However, this reuse strategy cannot efficiently reduce cold starts because the schedulers are agnostic of container lifecycle. For example, it may ignore soon available containers or evict soon needed containers. We propose a container lifecycle-aware scheduling strategy for serverless computing, CAS. The key idea is to control distribution of requests and determine creation or eviction of containers according to different lifecycle phases of containers. We implement a prototype of CAS on OpenWhisk. Our evaluation shows that CAS reduces 81% cold starts and therefore brings a 63% reduction at 95th percentile latency compared with native scheduling strategy in OpenWhisk when there is worker contention between workloads, and does not add significant performance overhead. © 2021 John Wiley & Sons Ltd.

关键词： Containers

来源：评论

学校读者我要写书评

暂无评论

Enhancing Hyperspectral Image Classification: A Hybrid CNN Approach with Spatial and Channel Attention Mechanisms 20

Enhancing Hyperspectral Image Classification: A Hybrid CNN A...

引用

20th International Computer Conference on Wavelet Active Media Technology and Information Processing, ICCWAMTIP 2023

作者： Butt, Imtiaz Ahmed Bo, Li Butt, Muhammad Hassaan Farooq Tassew, Megabiaw Tewodros Bin Akhtar Khan, Muhammad Usama School of Electronics and Information Northwestern Polytechnical University Shaanxi Xi'an710072 China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu611731 China School of Software Northwestern Polytechnical University Shaanxi Xi'an710072 China Comsats University Islamabad Department of Computing Wah Campus Wah47040 Pakistan

ISBN: (纸本)9798350318982

In the domain of hyperspectral imaging classification (HSIC) for remote sensing, recent advancements in deep learning have proven transformative. Convolutional neural networks have demonstrated immense potential in resolving these issues of HSIC, including insufficient labeled data, redundant spatial and spectral features, and overfitting. Traditional convolutional neural networks (CNNs) have effectively extracted spectral and spatial features, but 2D CNNs are limited in spatial modeling, while 3D CNNs alone struggle to distinguish spectral and spatial characteristics. Given that the accuracy of HSIC hinges on both spatial and spectral information, we proposed a hybrid-CNN model designed to mitigate the constraints of 2D and 3D CNNs. Our approach involves leveraging hybrid CNNs with spatial and channel attention (CA) mechanisms to address challenges like overfitting and model complexity. The proposed framework also improves generalization performance compared to 2D or 3D CNNs alone. Experiments were conducted on open datasets from Pavia University, Indian Pines, and Salinas to validate the proposed approach. The results demonstrate the efficacy of the hybrid CNN model with spatial and CA in consistently producing excellent classification results through thorough analyses with different deep-learning models. © 2023 IEEE.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Correction: Harnessing AI for sustainable higher education: ethical considerations, operational efficiency, and future directions

引用

Discover Sustainability 2025年第1期6卷 1-2页

作者： Khan, Sunawar Mazhar, Tehseen Shahzad, Tariq Khan, Muhammad Amir Rehman, Ateeq Ur Saeed, Mamoon M. Hamam, Habib Department of Software Engineering Islamia University of Bahawalpur Bahawalnagar Pakistan School of Computer Science National College of Business Administration and Economics Lahore Pakistan Department of Computer Science School Education Department Government of Punjab Layyah Pakistan Department of Computer Engineering COMSATS University Islamabad Sahiwal Pakistan School of Computing Sciences College of Computing Informatics and Mathematics Universiti Teknologi MARA Shah Alam Malaysia School of Computing Gachon University Seongnam Republic of Korea Department of Communications and Electronics Engineering Faculty of Engineering University of Modern Sciences (UMS) Sana’a Yemen Faculty of Engineering Uni de Moncton Moncton Canada School of Electrical Engineering University of Johannesburg Johannesburg South Africa International Institute of Technology and Management (IITG) Av. Grandes Ecoles Libreville BP 1989 Gabon Bridges for Academic Excellence-Spectrum Tunis Tunisia

来源：评论

学校读者我要写书评

暂无评论

Runtime Backdoor Detection for Federated Learning via Representational Dissimilarity Analysis

arXiv

引用

arXiv 2025年

作者： Zhang, Xiyue Xue, Xiaoyong Du, Xiaoning Xie, Xiaofei Liu, Yang Sun, Meng School of Computer Science University of Bristol United Kingdom Department of Information and Computing Science School of Mathematical Science Peking University China Department of Software Systems and Cybersecurity Monash University Australia School of Computing and Information Systems Singapore Management University Singapore School of Computer Science and Engineering Nanyang Technological University Singapore

Federated learning (FL), as a powerful learning paradigm, trains a shared model by aggregating model updates from distributed clients. However, the decoupling of model learning from local data makes FL highly vulnerable to backdoor attacks, where a single compromised client can poison the shared model. While recent progress has been made in backdoor detection, existing methods face challenges with detection accuracy and runtime effectiveness, particularly when dealing with complex model architectures. In this work, we propose a novel approach to detecting malicious clients in an accurate, stable, and efficient manner. Our method utilizes a sampling-based network representation method to quantify dissimilarities between clients, identifying model deviations caused by backdoor injections. We also propose an iterative algorithm to progressively detect and exclude malicious clients as outliers based on these dissimilarity measurements. Evaluations across a range of benchmark tasks demonstrate that our approach outperforms state-of-the-art methods in detection accuracy and defense effectiveness. When deployed for runtime protection, our approach effectively eliminates backdoor injections with marginal overheads. © 2025, CC BY.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Gene selection based on adaptive neighborhood-preserving multi-objective particle swarm optimization

引用

PeerJ Computer Science 2025年 11卷 1-35页

作者： Mehta, Sumet Sohail, Muhammad Twala, Bhekisipho Ullah, Asad Ullah, Fasee Khan, Arfat Ahmad Ling, Qinghua School of Computer Science & Communication Engineering Jiangsu University Jiangsu Zhenjiang China Department of Computer Software Engineering Military College of Signals NUST Islamabad Islamabad Pakistan Faculty of Information and Communication Technology Tshwane University of Technology Pretoria West Pretoria South Africa The Department of Computing Universiti Teknologi PETRONAS Perak Darul Ridzuan Seri Iskandar Malaysia Department of Computer Science College of Computing Khon Kaen University Khon Kaen Khon Kaen Thailand School of Computer Science and Engineering Jiangsu University of Science & Technology Jiangsu Zhenjiang China

The analysis of high-dimensional microarray gene expression data presents critical challenges, including excessive dimensionality, increased computational burden, and sensitivity to random initialization. Traditional optimization algorithms often produce inconsistent and suboptimal results, while failing to preserve local data structures limiting both predictive accuracy and biological interpretability. To address these limitations, this study proposes an adaptive neighborhood-preserving multi-objective particle swarm optimization (ANPMOPSO) framework for gene selection. ANPMOPSO introduces four key innovations: (1) a weighted neighborhood-preserving ensemble embedding (WNPEE) technique for dimensionality reduction that retains local structure;(2) Sobol sequence (SS) initialization to enhance population diversity and convergence stability;(3) a differential evolution (DE)-based adaptive velocity update to dynamically balance exploration and exploitation;and (4) a novel ranking strategy that combines Pareto dominance with neighborhood preservation quality to prioritize biologically meaningful gene subsets. Experimental evaluations on six benchmark microarray datasets and eleven multi-modal test functions (MMFs) demonstrate that ANPMOPSO consistently outperforms state-of-the-art methods. For example, it achieves 100% classification accuracy on Leukemia and Small-Round-Blue-Cell Tumor (SRBCT) using only 3–5 genes, improving accuracy by 5–15% over competitors while reducing gene subsets by 40–60%. Additionally, on MMFs, ANPMOPSO attains superior hypervolume values (e.g., 1.0617 ± 0.2225 on MMF1, approximately 10–20% higher than competitors), confirming its robustness in balancing convergence and diversity. Although the method incurs higher training time due to its structural and adaptive components, it achieves a strong trade-off between computational cost and biological relevance, making it a promising tool for high-dimensional gene selection in bioinformatics. © 2025

关键词： Particle swarm optimization (PSO)

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：