检索结果-内蒙古大学图书馆

Efficient algorithms for Rank-Regret Minimization

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2024年第8期36卷 3801-3816页

作者： Xiao, Xingxing Li, Jianzhong Harbin Inst Technol Dept Comp Sci & Technol Harbin 150001 Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Fac Comp Sci & Control Engn Shenzhen 518000 Peoples R China

Multi-criteria decision-making usually requires finding a small representative set from the database. A popular method, the regret minimization set (RMS) query, returns a size r subset S of the full dataset $D$D that minimizes the regret-ratio (the difference between the scores of top-1 in S and top-1 in D, for any utility function). RMS is not shift invariant, causing inconsistency in results. Further, the regret-ratio is often a "made up" number and users may mistake its absolute value. Instead, users do understand the notion of rank. Therefore, in this paper, we consider finding a fixed-size set S to minimize the maximum rank-regret (the rank of top-1 of S in the sorted list of D) over all possible utility functions, called the rank-regret minimization (RRM) problem, which is shift invariant. In 2D space, we design an exact algorithm 2DRRM for RRM. In HD space, we propose an approximate algorithm HDRRM with theoretical guarantees on rank-regret. It combines the ideas of space discretization and clustering. Extensive experiments verify the efficiency and effectiveness of our algorithms. In particular, HDRRM always has the best output quality in experiments.

关键词： Automobiles Databases Approximation algorithms Minimization clustering algorithms Logistics Decision making Multi-criteria decision-making rank-regret regret-ratio skyline top-k query

来源：评论

学校读者我要写书评

暂无评论

Transmit Power Minimization for IRS-Assisted Hierarchical Rate-Splitting Multiple Access Systems

引用

IEEE WIRELESS COMMUNICATIONS LETTERS 2025年第5期14卷 1386-1390页

作者： Qi, Ting Lyu, Bin You, Changsheng Nanjing Univ Posts & Telecommun Sch Commun & Informat Engn Nanjing 210003 Peoples R China Southern Univ Sci & Technol Dept Elect & Elect Engn Shenzhen 518055 Peoples R China

Rate-splitting multiple access (RSMA) and intelligent reflecting surface (IRS) are powerful techniques for satisfying diverse quality of service (QoS) requirements in future wireless networks. In this letter, we propose an efficient hierarchical rate-splitting (HRS) transmission scheme tailored for IRS-assisted multiple-input single-output systems. Specifically, we formulate an optimization problem to minimize the transmit power by jointly designing the user clustering, common rate distribution, transmit beamforming and IRS passive beamforming. In particular, we propose a low-complexity user clustering method by taking into account the QoS diversity and channel similarity of users. Then, we decompose the original non-convex problem into two subproblems and adopt the alternating optimization algorithm. The subproblems for optimizing the beamformer, common rate allocation, and IRS phase shift, efficiently solved by using successive convex approximation and semi-definite relaxation algorithms. Simulation results show that the proposed scheme can significantly reduce the transmit power compared with various benchmarks.

关键词： Streams Quality of service Vectors Optimization Interference Array signal processing clustering algorithms Resource management clustering methods Transmitting antennas Intelligent reflecting surface (IRS) hierarchical rate-splitting (HRS) rate-splitting multiple access (RSMA) user clustering alternating optimization

来源：评论

学校读者我要写书评

暂无评论

Joint NOMA clustering and Power Allocation in IoRT-Oriented Satellite Terrestrial Relay Networks

引用

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 2022年第10期71卷 11078-11088页

作者： Zhao, Bo Ren, Guangliang Dong, Xiaodai Northwestern Ploytech Univ Shenzhen Res & Dev Inst Shenzhen Peoples R China Xidian Univ State Key Lab Integrated Serv Networks Xian 710071 Peoples R China Univ Victoria Dept Elect & Comp Engn Victoria BC V8P 5C2 Canada

In this paper, a joint non-orthogonal multiple access (NOMA) clustering and power allocation problem is studied to maximize the uplink total sum rate in Internet of Remote Things (IoRT)-oriented satellite terrestrial relay networks (STRNs). The joint optimization problem is a mixed-integer programming (MIP) problem, which is non-convex and NP-hard. To solve this problem, we decompose it into two subproblems and propose staged algorithms to solve them. The first subproblem is an optimal NOMA clustering problem, which is still non-convex and NP-hard. To solve the subproblem efficiently, a reinforcement learning-based dynamic clustering algorithm (RL-DCA) is proposed. Using the RL-DCA, the users are able to gradually learn the optimal NOMA clustering policy in a distributed fashion. The RL-DCA has fast convergence and its computational complexity at each user remains fixed for any network size. The second subproblem is an intra-cluster optimal NOMA power allocation problem, which is still non-convex. To solve it, we firstly convert it into a convex problem by constrait approximation, then use the Karush-Kuhn-Tucker (KKT) conditions based algorithm to find the optimal power allocation policy. The second subproblem is solved in an offline fashion, which enables low computational complexity in transmission. Simulation results show that: 1) The RL-DCA has fast convergence and enables good clustering performance;2) The joint NOMA clustering and power allocation scheme greatly outperforms the orthogonal multiple access (OMA) scheme and other NOMA schemes in terms of total sum rate.

关键词： NOMA Relays Uplink Resource management Satellites System performance clustering algorithms Satellite terrestrial relay networks Internet of Remote Things NOMA clustering power allocation joint optimization

来源：评论

学校读者我要写书评

暂无评论

Pavement Crack Classification and Recognition Algorithm Combined With Tensor Voting and RANSAC

引用

IEEE ACCESS 2024年 12卷 72117-72130页

作者： Wang, Haiying Tao, Jianyu Li, Hansheng Changan Univ Key Lab Rd Construct Technol & Equipment MOE Xian 710064 Peoples R China Shaanxi Transportat Holding Grp Co Ltd Xian 710075 Peoples R China

The existing methods for pavement crack classification and identification solely offer information about the crack type, neglecting size and direction details, which are essential for guiding repair efforts and forming the engineer digital information data. In response to the challenges posed by insufficient crack information, prolonged training time and intricate parameter adjustment inherent in employing deep learning algorithms for pavement crack classification and recognition, we propose an integrated approach combining tensor voting with the random sample consensus for pavement crack classification and recognition. The method involves pre-processing road images using gray value transformation and the K-Means clustering algorithm. Subsequently, the tensor voting algorithm is applied to enhance the linear features, resulting in the generation of linear saliency maps of cracks along with crack junction information. Furthermore, a non-maximum suppression method and the RANSAC algorithm are employed to refine and fit the crack skeleton curves respectively, accomplishing the crack classification and recognition. The outcomes demonstrate that the proposed integrated approach in the crack skeleton segmentation algorithm yields an average F1-score of 0.7879, outperforming traditional non-maximum suppression methods. The accuracy of crack classification and recognition reaches 96%, outperforming other crack classification and recognition algorithms grounded in digital image processing methods. Compared with the neural networks employed for classification and recognition, the proposed algorithm is able to capture direction and size details of cracks, which can provide guidance for intelligent crack repair. This additional information can offer valuable guidance for intelligent crack repair processes.

关键词： Classification algorithms clustering algorithms Tensors Image segmentation Roads Interference Surface cracks Random sequences Crack classification and recognition non-maximum suppression random sample consensus (RANSAC) tensor voting

来源：评论

学校读者我要写书评

暂无评论

Hybrid-Algorithm-Based Full Coverage Search Approach With Multiple AUVs to Unknown Environments in Internet of Underwater Things

引用

IEEE INTERNET OF THINGS JOURNAL 2024年第6期11卷 11058-11072页

作者： Han, Guangjie Lai, Weizhe Wang, Hao Zhu, Shengchao Hohai Univ Dept Internet Things Engn Changzhou 213022 Peoples R China Hohai Univ Dept Commun & Informat Syst Changzhou 213022 Peoples R China Hohai Univ Coll Harbour Coastal & Offshore Engn Nanjing 210024 Peoples R China Chinese Acad Sci Inst Acoust State Key Lab Acoust Beijing 100190 Peoples R China

In the development of Internet of Underwater Things (IoUT), the unknown nature of the underwater environment is a challenging issue. In various domains related to IoUT, utilizing autonomous underwater vehicles (AUVs) for unmanned and autonomous missions has become an inevitable trend. Considering the particularity of underwater environments, this study proposes a hybrid-algorithm-based full coverage search approach to searching moving targets in unknown underwater environments. This approach combines the improved Voronoi clustering strategy, the improved artificial bee colony (ABC) algorithm, the improved line-of-sight (LOS) technique, and the artificial potential field (APF) method to enhance the efficiency of underwater full coverage search (FCS). First, the improved Voronoi clustering strategy is employed to partition the entire underwater region and allocate each part to an AUV. Second, to enhance the search capability of AUVs, a full-dimensional ABC algorithm with adaptive factor is designed to plan global paths for AUVs to search for targets, and the paths are further smoothed using the improved acrlong SLOS technique. During the navigation of the AUVs along the global paths, obstacles may be detected;thus, the APF method is utilized to dynamically plan local paths for AUVs to avoid obstacles. Experimental results demonstrate that the proposed approach significantly improves the efficiency of underwater FCS.

关键词： Task analysis Path planning Resource management clustering algorithms Heuristic algorithms Internet of Things Smoothing methods Artificial bee colony (ABC) autonomous under water vehicle (AUV) swarm full coverage search (FCS) line of sight (LOS) path planning

来源：评论

学校读者我要写书评

暂无评论

Overlapping Community Detection Based on Weak Equiconcept

引用

IEEE ACCESS 2024年 12卷 42147-42162页

作者： Shi, Sunqian Yan, Mengyu Li, Jinhai Kunming Univ Sci & Technol Fac Sci Kunming 650500 Peoples R China Kunming Univ Sci & Technol Data Sci Res Ctr Kunming 650600 Peoples R China

Community discovery refers to the process of searching for clusters in a network that are sparsely connected to other nodes and formed dense connections internally. In many real networks, some communities often overlap with each other, meaning that a node may belong to multiple communities simultaneously. Revealing these (overlapping) community structures is an important issue in complex network analysis, as it helps to better analyze the characteristics and organizational structure of the network. Community expansion methods are very important in the study of community detection problems. However, one of the key issues in developing effective community expansion methods is that the position of seed nodes greatly affects the performance of these algorithms, resulting in low robustness of these algorithms. Meanwhile, it is also difficult for them to provide high-quality results for community detection task. To solve the above problem, this paper proposes a seed selection method based on the weak equiconcepts in a network formal context, which integrates the attribute information of nodes during random walk to detect overlapping communities. Specifically, the weak equiconcepts are constructed by establishing a network formal context to obtain seed sets, and an improved PageRank clustering algorithm is used to expand these seed sets to better reveal the overlapping community structure in the network. Experiments show that seed selection is helpful to improving the performance of overlapping community detection algorithms.

关键词： clustering algorithms Complex networks Task analysis Optimization Loss measurement Knowledge discovery Information entropy Detection algorithms Overlapping community discovery network formal context weak equiconcept random walk

来源：评论

学校读者我要写书评

暂无评论

Recovering Static and Time-Varying Communities Using Persistent Edges

引用

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 2024年第2期11卷 2087-2099页

作者： Avrachenkov, Konstantin Dreveton, Maximilien Leskela, Lasse Inria Sophia Antipolis F-06902 Valbonne France Ecole Polytech Fed Lausanne Sch Comp & Commun Sci CH-1015 Lausanne Switzerland Aalto Univ Dept Math & Syst Anal Espoo 02150 Finland

This article focuses on spectral methods for recovering communities in temporal networks. In the case of fixed communities, spectral clustering on the simple time-aggregated graph (i.e., the weighted graph formed by the sum of the interactions over all temporal snapshots) does not always produce satisfying results. To utilise information carried by temporal correlations, we propose to employ different weights on freshly appearing and persistent edges. We show that spectral clustering on such weighted graphs can be explained as a relaxation of the maximum likelihood estimator of an extension of the degree-corrected stochastic block model with Markov interactions. We also study the setting of evolving communities, for which we use the prediction at time t-1 as an oracle for inferring the community labels at time t. We demonstrate the accuracy of the proposed methods on synthetic and real data sets.

关键词： Markov processes Maximum likelihood estimation clustering algorithms Biological system modeling Image edge detection Partitioning algorithms Optimization Graph clustering temporal networks spectral methods stochastic block model

来源：评论

学校读者我要写书评

暂无评论

Channel Pruning of Transfer Learning Models Using Novel Techniques

引用

IEEE ACCESS 2024年 12卷 94914-94925页

作者： Thaker, Pragnesh Mohan, Biju R. Natl Inst Technol Karnataka Srinivasnagar Surathkal 575025 Karnataka India

This research paper delves into the challenges associated with deep learning models, specifically focusing on transfer learning. Despite the effectiveness of widely used models such as VGGNet, ResNet, and GoogLeNet, their deployment on resource-constrained devices is impeded by high memory bandwidth and computational costs, and to overcome these limitations, the study proposes pruning as a viable solution. Numerous parameters, particularly in fully connected layers, contribute minimally to computational costs, so we focus on convolution layers' pruning. The research explores and evaluates three innovative pruning methods: the Max3 Saliency pruning method, the K-Means clustering algorithm, and the Singular Value Decomposition (SVD) approach. The Max3 Saliency pruning method introduces a slight variation by using the three maximum values of the kernel instead of all nine to compute the saliency score. This method is the most effective, substantially reducing parameter and Floating Point Operations (FLOPs) for both VGG16 and ResNet56 models. Notably, VGG16 demonstrates a remarkable 46.19% reduction in parameters and a 61.91% reduction in FLOPs. Using the Max3 Saliency pruning method, ResNet56 shows a 35.15% reduction in parameters and FLOPs. The K-Means pruning algorithm is also successful, resulting in a 40.00% reduction in parameters for VGG16 and a 49.20% reduction in FLOPs. In the case of ResNet56, the K-Means algorithm achieved a 31.01% reduction in both parameters and FLOPs. While the Singular Value Decomposition (SVD) approach provides a new set of values for condensed channels, its overall pruning ratio is smaller than the Max3 Saliency and K-Means methods. The SVD pruning method prunes 20.07% parameter reduction and a 24.64% reduction in FLOPs achieved for VGG16, along with a 16.94% reduction in both FLOPs and parameters for ResNet56. Compared with the state-of-the-art methods, the Max3 Saliency and K-Means pruning methods performed better in Flops reduction m

关键词： Computational modeling Convolutional neural networks Kernel clustering algorithms Accuracy Transfer learning Filtering algorithms Deep compression of CNN channel pruning structured pruning neural network compression transfer learning

来源：评论

学校读者我要写书评

暂无评论

MITDBA: Mitigating Dynamic Backdoor Attacks in Federated Learning for IoT Applications

引用

IEEE INTERNET OF THINGS JOURNAL 2024年第6期11卷 10115-10132页

作者： Wang, Yongkang Zhai, Di-Hua Han, Dongyu Guan, Yuyin Xia, Yuanqing Beijing Inst Technol Sch Automat Beijing 100081 Peoples R China Yangtze Delta Reg Acad Beijing Inst Technol Jiaxing 314001 Peoples R China Beijing Bldg Mat Acad Sci Res Beijing 100041 Peoples R China

Federated learning (FL) is widely used in the Internet of Things (IoT) systems. However, FL is susceptible to backdoor attacks due to its inherently distributed and privacy-preserving nature. Existing studies assume that backdoor triggers on different malicious clients are universal, and most defense algorithms are designed to counter backdoor attacks based on this assumption. Recently, dynamic backdoor attacks have been proposed to undermine robust algorithms in centralized machine learning. We introduce dynamic backdoor attacks into the FL system and develop three types of dynamic backdoors named Aggregation, Single, and Continuous to target the FL system. To defend against such attacks, we propose a novel robust algorithm called MITDBA, which utilizes gramian information to capture high-order representations, then employs spectral signatures to detect and remove malicious clients, and finally utilizes clipping operations to filter the selected local models during the aggregation process. We conduct attack and defense experiments on MNIST, CIFAR-10, and GTSRB data sets. The experimental results demonstrate that our designed attack strategies can successfully insert dynamic backdoors into the global model, bypassing the existing state-of-the-art defenses, but these attacks can be effectively mitigated by MITDBA.

关键词： Heuristic algorithms Training Internet of Things Data models Automobiles Vehicle dynamics clustering algorithms Dynamic backdoor federated learning (FL) gram matrix robust spectral signature

来源：评论

学校读者我要写书评

暂无评论

Double-Layer K-Means plus plus clustering Method for Evaluation of Dispatchable Potential of Massive Regional 5G Base Stations

引用

IEEE ACCESS 2022年 10卷 82870-82882页

作者： Zhong, Lijun Zhu, Mengting Chen, Shuiyao Shi, Minda Zhang, Xianglong Chen, Ding State Grid Zhejiang Elect Power Co Jiaxing Power Supply Co Jiaxing 314000 Peoples R China Zhejiang Univ Coll Elect & Engn Hangzhou 310027 Peoples R China State Grid Zhejiang Elect Power Co Ltd Hangzhou 310000 Peoples R China State Grid Econ & Technol Res Inst Co Ltd Beijing 102209 Peoples R China

5G base stations (BSs), which are the essential parts of the 5G network, are important user-side flexible resources in demand response (DR) for electric power system. However, a 5G BS has little and difference dispatchable potential, how to make massive 5G BSs participate in DR conveniently is an urgent problem to be solved. clustering is an effective solution. Aiming at the special requirements of big data analysis and dispatching difficulties brought by the massive and ultra-dense distribution of 5G BSs, this paper proposes a double-layer K-means++ clustering method for regional 5G BSs considering the main characteristics of 5G BSs. The proposed clustering method considers both the geographical location and power consumption characteristics of the regional 5G BSs to partition the 5G BSs into appropriate number of clusters. Besides, the comprehensive charging and discharging dispatchable potential evaluation indices are established based on the clustering results to evaluate the dispatchable potential of each cluster. The effectiveness of the proposed method and the dispatchable potential evaluation indices are verified through a case of 5000 5G BSs in Jiaxing, China. The simulation findings reveal that the performance of the proposed clustering method is better than classical clustering algorithm and the dispatchable potential of the 5G BSs clusters can be accurately analyzed by the proposed comprehensive indices, which lay the foundation for friendly and fine interaction between 5G BSs and the power grid.

关键词： 5G mobile communication Power demand Batteries clustering methods Base stations clustering algorithms Indexes 5G base station K-means plus plus clustering dispatchable potential evaluation big data analysis demand response

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：