检索结果-内蒙古大学图书馆

18th International Conference on Wireless Artificial intelligent computing Systems and Applications, WASA 2024

作者： Ou, Pingjie Chen, Ningjiang Huang, Zizhan School of Computer and Electronic Information Guangxi University Guangxi Nanning530004 China Key Laboratory of Parallel Distributed and Intelligent Computing Education Department of Guangxi Zhuang Autonomous Region Guangxi University Guangxi Nanning530004 China

ISBN: (纸本)9783031714696

Due to the limited resources of edge networks, the heterogeneity of user content requests, high-cost caching from direct resource hits, and redundancy in resource retention time hinder system performance. Traditional caching methods based on learning and popularity tend to encounter data skew when handling vehicle requests, leading to untimely cache data updates. To address these issues, this paper proposes an efficient edge cooperative caching strategy with recommendation awareness. Firstly, a time series prediction model based on a multi-dense multilayer perceptron is designed. This model segments historical request information into sub-sequence level slivers to enhance positional correlations between information, predicting user requests through historical time series and covariates. Secondly, to optimize cache utilization, comprehensively considers user mobility and contribution to enable recommendation awareness. Additionally, an optimized lifecycle is introduced to design a cache replacement strategy. Simulation experiments demonstrate that, compared to caching schemes based on content popularity and deep reinforcement learning, the proposed method achieves lower caching costs and higher cache hit rates. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Deep anomaly detection with partition contrastive learning for tabular data

引用

Data Mining and Knowledge Discovery 2025年第4期39卷 1-38页

作者： Li, Yizhou Wang, Yijie Xu, Hongzuo Li, Bin Zhou, Xiaohui National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology Changsha China Intelligent Game and Decision Lab (IGDL) Beijing China

Self-supervised anomaly detection (AD) methods define transformations and surrogate tasks to deeply learn data “normality”, presenting superior performance. Different from most existing work designed for images, this paper considers self-supervised AD for tabular data, which has two desiderata: (i) transformation operations need to generate diverse transformed samples for downstream contrastive learning; (ii) the learning of surrogate tasks is required to perceive the semantics of tabular data, i.e., local–global patterns including intra-instance regularities and inter-instance structures. Related studies devise applicable transformations, but their surrogate tasks often neglect the inter-instance structures, failing to describe comprehensive data “normality" accurately. To fill these gaps, this paper proposes a novel partition contrastive learning-based anomaly detection method. We first devise a new transformation. Transformed samples are created by representing sub-vectors generated from different partitions of the whole feature space, during which different feature couplings are embedded in the transformed samples to make them sufficiently diverse. To capture intra-instance regularities, our approach learns a representation space where the transformed samples repel each other while concurrently resembling their corresponding spatial centers. A constraint is posed on the frames of these spatial centers to preserve their similarities between the corresponding original instances, facilitating the learning of inter-instance structures. The synergy of these two learning objectives encourages the modeling of tabular data semantics, thereby comprehensively modeling data “normality”. Abnormal degrees of the testing data are obtained by evaluating whether they conform to these learned patterns shared in the majority of data. Extensive experiments show that our approach achieves significant improvements (13% AUC-ROC and 30% AUC-PR) over state-of-the-art AD methods.

关键词： Anomaly detection Contrastive learning Data mining Deep learning Local–global contrastive loss Self-supervised learning Subspace partition Tabular data Transformation Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

HyperGAC: Information Entropy-Driven Resource Allocation for Cloud-Edge-End computing

引用

Journal of Grid computing 2025年第2期23卷 1-23页

作者： Chen, Juan Bai, Lin Chen, Ningjiang Yang, Yubin School of Computer and Electronic Information Guangxi University Nanning China Key Laboratory of Parallel Distributed and Intelligent Computing (Guangxi University) Education Department of Guangxi Zhuang Autonomous Region Nanning China Guangxi Center of Technology Innovation for Intelligent Digital Services Nanning China

Cloud-Edge-End computing provides favorable support for emerging applications by virtue of its low-latency and high reliability. Resource allocation strategies significantly impact the quality of service and response time of cloud-edge-end systems. Due to the decentralized storage of resources, large fluctuations in task demands, and the complexity of cloud-edge-end environments, achieving effective resource allocation remains challenging. Existing resource allocation strategies rely on static models or predefined rules, and cannot flexibly respond to real-time network conditions and changes in task requirements. Consequently, they are limited in achieving optimal resource utilization and effective load balancing. Therefore, this paper proposes an Information Entropy (IE)-driven cloud-edge-end resource allocation mechanism called HyperGAC. First, IE is employed to quantify resource fluctuations, so as to accurately capture the dynamic changes in resource demands. And an Entropy-based Correlation Recurrent Unit (ECRU) is designed to predict server resource utilization. Second, the hypergraph is used to model the complex relationships between resources and tasks, breaking through the limitations of traditional graph structures in multi-dimensional relationships modeling. Finally, the Hypergraph Neural Network (HGNN) is integrated into the actor-critic framework to extract the features from the hypergraph association model, while dynamically optimizing resource allocation strategy based on the prediction results. Experimental results demonstrate that the HyperGAC mechanism significantly improves resource utilization and load balancing performance while effectively reducing the task rejection rate.

关键词： Cloud-edge-end computing Deep reinforcement learning Information entropy Resource allocation Resource utilization prediction

来源：评论

学校读者我要写书评

暂无评论

Deep Time Series Anomaly Detection with Local Temporal Pattern Learning

Deep Time Series Anomaly Detection with Local Temporal Patte...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Yizhou Li Yijie Wang Hongzuo Xu Xiaohui Zhou National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology Changsha China Intelligent Game and Decision Lab (IGDL) Beijing China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Self-supervised time series anomaly detection (TSAD) demonstrates remarkable performance improvement by extracting high-level data semantics through proxy tasks. Nonetheless, most existing self-supervised TSAD techniques rely on manual- or neural-based transformations when designing proxy tasks, overlooking the intrinsic temporal patterns of time series. This paper proposes a local temporal pattern learning-based time series anomaly detection (LTPAD). LTPAD first generates sub-sequences. Pairwise sub-sequences naturally manifest proximity relationships along the time axis, and such correlations can be used to construct supervision and train neural networks to facilitate the learning of temporal patterns. Time intervals between two sub-sequences serve as labels for sub-sequence pairs. By classifying these labeled data pairs, our model captures the local temporal patterns of time series, thereby modeling the temporal pattern-aware "normality". Abnormal scores of testing data are acquired by evaluating their conformity to these learned patterns shared in training data. Extensive experiments show that LTPAD significantly outperforms state-of-the-art competitors.

关键词： Time series analysis Semantics Neural networks Training data Manuals Signal processing Data models Speech processing Anomaly detection Testing

来源：评论

学校读者我要写书评

暂无评论

An End-to-End Classification Network Model Based on Hybrid Supervision for Industrial Surface Defect Detection

引用

International Journal of Pattern Recognition and Artificial Intelligence 2025年第5期39卷

作者： Qin, Runbing Chen, Ningjiang Gan, Shukun Guangxi University School of Computer Electronics and Information Nanning530004 China Education Department of Guangxi Zhuang Autonomous Region Key Laboratory of Parallel Distributed and Intelligent Computing Guangxi University Nanning530004 China Guangxi Intelligent Digital Services Research Center of Engineering Technology Nanning530004 China

Industrial part surface defect detection aims to precisely locate defects in images, which is crucial for quality control in manufacturing. The traditional method needs to be designed in advance, but it has shortcomings in terms of generalization ability. Deep neural network-based defect detection faces issues such as low resolution, data scarcity, labeling costs, and high computation. Therefore, an improved solution is necessary to enhance industrial product quality control efficiency and accuracy. A new method is proposed that utilizes end-to-end training of a two-stage neural network based on segmentation with an extended training process. The gradient flow from the classification to the segmentation network is adjusted to prevent unstable features from interfering with learning. To address the problem of image oversampling and undersampling during training, a frequency sampling scheme for negative samples is introduced. Additionally, positive pixels in the region-based segmentation mask are weighted using the distance transform algorithm so that regions with a high probability of defects can be detected without detailed annotation. Experiments are conducted across three distinct defect datasets by applying hybrid supervision encompassing diverse conditions. In the optimal case, the AP rate of the proposed model on the DAGM and KolektorSDD datasets reaches 100%, and the AP rate on the Severstal Steel dataset reaches 98.99%. The experiments’ results show that the detection accuracy has improved greatly, showing how well the proposed method works in the industry. © 2025 World Scientific Publishing Company.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A Multi-holder Role and Strange Attractor-Based Data Possession Proof in Medical Clouds 24th

A Multi-holder Role and Strange Attractor-Based Data Possess...

引用

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Guo, Jinyuan Sun, Lijuan Wu, Jingchen Chu, Chiawei Gao, Yutong Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education Beijing China School of Cybersecurity Beijing University of Posts and Telecommunications Beijing China Key Laboratory of Ethnic Language Intelligent Analysis and Security Governance Ministry of Education Minzu University Beijing China Intellectual Property Information Service Center BHU Beihang University Beijing China City University of Macau Taipa China Li’an International Education Innovation Pilot Zone Hainan International College of Minzu University of China Hainan Beijing China

ISBN: (纸本)9789819615278

Proof of Data Possession is a technique for ensuring the integrity of data stored in cloud storage. However, most audit schemes assume only one role for data owners, which is not suitable for complex Smart Healthcare Network with multiple roles. In such schemes addressing the risk of third-party auditors (TPA) colluding with other entities, blockchain is typically employed. However, this necessitates extensive computation, resulting in unacceptably high energy consumption and prolonged verification times. In this paper, we propose a Proof of Data Possession scheme tailored for medical network environments, which incorporates protocols designed for both healthcare professionals and patients, considering the diverse roles that generate data within medical network applications and the varying performance of the devices they use. During the generation of metadata for verification, the protocol incorporates the varying security risk levels associated with medical data. For subsequent metadata selection, an audit block extraction strategy based on strange attractors is employed, effectively mitigating the risk of collusion by third-party TPA, thereby enhancing the security and efficiency of the scheme. Additionally, we conducted experimental analysis and comparisons, demonstrating higher security and practical applicability. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Data integrity

来源：评论

学校读者我要写书评

暂无评论

Towards intelligent Transportation with Pedestrians and Vehicles In-the-Loop: A Surveillance Video-Assisted Federated Digital Twin Framework

arXiv

引用

arXiv 2025年

作者： Li, Xiaolong Wei, Jianhao Wang, Haidong Dong, Li Chen, Ruoyang Yi, Changyan Cai, Jun Niyato, Dusit Shen, Xuemin The Xiangjiang Laboratory Hunan Provincial General University Key Laboratory of IoT Intelligent Sensing and Distributed Collaborative Optimization Hunan University of Technology and Business Changsha410205 China The College of Computer Science and Technology Nanjing University of Aeronautics and Astronautics China The Department of Electrical and Computer Engineering University of Concordia Montreal Canada The College of Computing and Data Science Nanyang Technological University Singapore The Department of Electrical and Computer Engineering University of Waterloo Canada

In intelligent transportation systems (ITSs), incorporating pedestrians and vehicles in-the-loop is crucial for developing realistic and safe traffic management solutions. However, there is falls short of simulating complex real-world ITS scenarios, primarily due to the lack of a digital twin implementation framework for characterizing interactions between pedestrians and vehicles at different locations in different traffic environments. In this article, we propose a surveillance video assisted federated digital twin (SV-FDT) framework to empower ITSs with pedestrians and vehicles in-the-loop. Specifically, SVFDT builds comprehensive pedestrian-vehicle interaction models by leveraging multi-source traffic surveillance videos. Its architecture consists of three layers: (i) the end layer, which collects traffic surveillance videos from multiple sources;(ii) the edge layer, responsible for semantic segmentation-based visual understanding, twin agent-based interaction modeling, and local digital twin system (LDTS) creation in local regions;and (iii) the cloud layer, which integrates LDTSs across different regions to construct a global DT model in realtime. We analyze key design requirements and challenges and present core guidelines for SVFDT’s system implementation. A testbed evaluation demonstrates its effectiveness in optimizing traffic management. Comparisons with traditional terminal-server frameworks highlight SV-FDT’s advantages in mirroring delays, recognition accuracy, and subjective evaluation. Finally, we identify some open challenges and discuss future research directions. Copyright © 2025, 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

Type-2 fuzzy description logic

引用

Frontiers of Computer Science 2011年第2期5卷 205-215页

作者： Ruixuan LI Kunmei WEN Xiwu GU Yuhua LI Xiaolin SUN Bing LI Intelligent and Distributed Computing Laboratory School of Computer Science and Technology Huazhong University of Science and Technology Wuhan 430074 China State Key Laboratory of Software Engineering Wuhan University Wuhan 430072 China

Description logics （DLs） are widely employed in recent semantic web application systems. However, classical description logics are limited when dealing with imprecise concepts and roles, thus providing the motivation for this work. In this paper, we present a type-2 fuzzy attributive concept language with complements （ALC） and provide its knowledge representation and reasoning algorithms. We also propose type-2 fuzzy web ontology language （OWL） to build a fuzzy ontology based on type- 2 fuzzy ALC and analyze the soundness, completeness, and complexity of the reasoning algorithms. Compared to type-1 fuzzy ALC, type-2 fuzzy ALC can describe imprecise knowledge more meticulously by using the membership degree interval. We implement a semantic search engine based on type-2 fuzzy ALC and carry out experiments on real data to test its performance. The results show that the type-2 fuzzy ALC can improve the precision and increase the number of relevant hits for imprecise information searches.

关键词： description logic (DL) type-2 fuzzy attributive concept language with complements (ALC) fuzzy ontology reasoning semantic search engine

来源：评论

学校读者我要写书评

暂无评论

Optimizing ranking method using social annotations based on language model

引用

Artificial Intelligence Review 2014年第1期41卷 81-96页

作者： Wen, Kunmei Li, Ruixuan Xia, Jing Gu, Xiwu Intelligent and Distributed Computing laboratory of theHuaZhong University of Science and Technology WuHan China 709th Research Institute China Shipbuilding Industry Corporation WuHan China

Recent research has shown that more and more web users utilize social annotations to manage and organize their interested resources. Therefore, with the growing popularity of social annotations, it is becoming more and more important to utilize such social annotations to achieve effective web search. However, using a statistical model, there are no previous studies that examine the relationships between queries and social annotations. Motivated by this observation, we use social annotations to re-rank search *** intend to optimize retrieval ranking method by using the ranking strategy of integrating the queryannotation similarity into query-document similarity. Specifically, we calculate the queryannotation similarity by using a statistical language model, which in a shorter form we call simply a language model. Then the initial search results are re-ranked according to the computational weighted score of the query-document similarity score and the query-annotation similarity score. Experimental results showthat the proposed method can improve theNDCG score by 8.13%. We further conduct an empirical evaluation of the method by using a query set including about 300 popular social annotations and constructed phrases. More generally, the optimized results with social annotations based on a language model can be of significant benefit to web search.©Springer Science+Business Media B.V. 2011.

关键词： Query processing

来源：评论

学校读者我要写书评

暂无评论

ESFMTO: A reliable task offloading strategy based on edge server failure model in IIoT

引用

Ad Hoc Networks 2025年 176卷

作者： Yubin Yang Yan Chen Ningjiang Chen Juan Chen School of Computer and Electronic Information Guangxi University Nanning 530004 China Key Laboratory of Parallel Distributed and Intelligent Computing (Guangxi University) Education Department of Guangxi Zhuang Autonomous Region Nanning 530004 China Guangxi Center of Technology Innovation for Intelligent Digital Services Nanning 530004 China

The extensive use of automation equipment and sensors in the Industrial Internet of Things (IIoT) has led to a significant increase in data volume, which has placed higher demands on the real-time processing capability. Edge computing enables real-time response and rapid decision-making by offloading data processing to the edge of the network. However, the complexity of the industrial production environment leads to edge server failures, which seriously affects the system stability and security. To address this issue, this paper develops an edge server failure model for IIoT, analyzing the interaction between the hardware failure and the container failure during the failure occurrence and recovery. Further, based on the edge server failure model, a task offloading strategy named ESFMTO is proposed, which employs the SAC-BNN (Soft Actor-Critic with Bayesian Neural Network) algorithm. The probability distribution of the task completion time is updated through the Bayesian Neural Network (BNN), accurately evaluating the Conditional Value at Risk (CVaR). During the training process of Deep Reinforcement Learning (DRL), a perturbation neural network is introduced to perturb the input state, which enhances the system robustness under uncertain failure conditions. Unlike existing approaches that often assume each edge server hosts a single container, the paper explicitly considers multi-container deployment in IIoT, bridging the gap between theoretical assumptions and real-world industrial requirements. Experimental results demonstrate that SAC-BNN outperforms existing methods in dealing with the edge server failure, reducing CVaR by at least around 0.9 and improving task completion rates by at least 13.86% compared to the baseline algorithms.

关键词： Normal distribution

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：