检索结果-内蒙古大学图书馆

Improving Accuracy and Convergence in Group-Based Federated Learning on Non-IID Data

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 2023年第3期10卷 1389-1404页

作者： He, Ziqi Yang, Lei Lin, Wanyu Wu, Weigang South China Univ Technol Coll Software Engn Guangzhou 510006 Peoples R China Hong Kong Polytech Univ Dept Comp Hong Kong Peoples R China Sun Yat Sen Univ Coll Comp Sci & Engn Guangzhou 510006 Peoples R China

Federated learning (FL) enables a large number of edge devices to learn a shared model without data sharing collaboratively. However, the imbalanced data distribution among users poses challenges to the convergence performance of FL. Group-based FL is a novel framework to improve FL performance, which appropriately groups users and allows localized aggregations within the group before a global aggregation. Nevertheless, most existing Group-based FL methods are K-means-based approaches that need to explicitly specify the number of groups, which may severely reduce the efficacy and optimality of the proposed solutions. In this paper, we propose a grouping mechanism called Auto-Group, which can automatically group users without specifying the number of groups. Specifically, various grouping strategies with different numbers of groups are generated with our mechanism. In particular, equipped with an optimized Genetic algorithm, Auto-Group ensures that the data distribution of each group is similar to the global distribution, further reducing the communication delay. We conduct extensive experiments in various settings to evaluate Auto-Group. Experimental results show that, compared with the baselines, our mechanism can significantly improve the model accuracy while accelerating the training speed.

关键词： Delays Statistics Sociology Genetic algorithms Costs Approximation algorithms Data models Federated learning distributed machine learning grouping algorithm non-IID data

来源：评论

学校读者我要写书评

暂无评论

An algorithm to Form Balanced and Diverse Groups of Students

引用

COMPUTER APPLICATIONS IN ENGINEERING EDUCATION 2011年第3期19卷 582-590页

作者： Yeoh, Hak Koon Nor, Mohamad Iskandr Mohamad Univ Malaya Fac Engn Dept Chem Engn Kuala Lumpur 50603 Malaysia

We share the details of a simple algorithm which groups students based on diversity in gender and race while keeping the group average CGPA nearly equal. Application on simulated and actual class profiles gave multiple-feasible solutions within seconds, with a maximum absolute deviation in group average CGPA of 0.03/4.00 or lower. (C) 2009 Wiley Periodicals, Inc. Comput Appl Eng Educ 19: 582-590, 2011;View this article online at ***/journal/cae;DOI 10.1002/cae.20338

关键词： grouping algorithm student grouping balanced groups diversity

来源：评论

学校读者我要写书评

暂无评论

Minority Sub-Region Estimation-Based Oversampling for Imbalance Learning

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2022年第5期34卷 2324-2334页

作者： Sun, Yi Cai, Lijun Liao, Bo Zhu, Wen Hunan Univ Coll Comp Sci & Elect Engn Changsha 410082 Peoples R China Hainan Normal Univ Key Lab Computat Sci & Applicat Hainan Prov Haikou 571158 Hainan Peoples R China

Class imbalance problem that characterized with the skew distribution towards the majority arises as one challenge in recent years. Many oversampling techniques have been proposed to cope with this problem and some of them combine the oversampling procedure with the clustering algorithm which guaranteeing new synthetic samples being generated in clusters. However far-away samples but with the same minority sub-region are generally clustered into different groups owing to the characteristic of clustering algorithm itself. Therefore, the following oversampling procedure is mostly carried in incomplete minority sub-regions that synthetic samples not well cover the integral minority region. And to our best knowledge, none of existing algorithm is designed to directly estimate minority sub-regions for class imbalance problem. Thus, one new grouping algorithm, named Direction Distribution-based Minority Sub-region Estimation (DDMSE), is first proposed. The new algorithm exploits the intuitive observation, that the minority with the same sub-region almost distribute within the same direction when compared to other majority, to estimate minority sub-regions that tactfully ignoring negative impacts brought by the distance factor like in clustering algorithms. Finally, new synthetic samples are generated in those minority sub-regions. And experimental results on real-world datasets show the comparable performance with other state-of-the-art oversampling methods.

关键词： Clustering algorithms Estimation Silicon Sun Machine learning algorithms Probabilistic logic Indexes Imbalance learning minority sub-region grouping algorithm oversampling

来源：评论

学校读者我要写书评

暂无评论

A 9-b 40-MSample/s reconfigurable pipeline analog-to-digital converter

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS 2002年第7期49卷 449-456页

作者： Liu, H Hassoun, M Conexant Syst Inc San Diego CA 92121 USA

In this paper, a reconfigurable pipeline analog-to-digital converter (ADC) architecture is proposed. Based on dynamic performance measurements, the best performance configuration will be chosen from a collection of possible configurations. A 40-MSample/s 9-b reconfigurable pipeline ADC is designed and implemented in Taiwan Semiconductor Manufacturing Corporation's (TSMC's) 0.25-mum single-poly CMOS digital process. The chip is measured for all the configurations under different temperatures to prove the reconfiguration will provide significant effective number of bits (ENOB) improvement among the set of configurations. The active area of the design is 5.9 mm(2).. The power consumption is 425 mW.

关键词： analog-to-digital converter (ADC) dynamic performance measurement grouping algorithm multiplying digital-to-analog converter (DAC) pipeline reconfiguration

来源：评论

学校读者我要写书评

暂无评论

Plant-wide oscillation detection using multivariate empirical mode decomposition

引用

COMPUTERS & CHEMICAL ENGINEERING 2018年 117卷 320-330页

作者： Aftab, Muhammad Faisal Hovd, Morten Sivalingam, Selvanathan NTNU Dept Engn Cybernet Trondheim Norway Siemens AS Trondheim Norway

Plant-wide oscillation detection is an important task in the maintenance of large-scale industrial control systems, owing to the fact that in an interactive multi-loop environment oscillation generated in one loop may propagate to the different parts of the plant. In such a scenario, it is required that different loops oscillating due to a common cause and hence similar frequency may be grouped together. In this paper an adaptive method for plant-wide oscillation detection based on multivariate empirical mode decomposition (MEMD) along with a grouping algorithm is proposed. The method can identify multiple oscillation groups among different variables as well as variables with random noise only. The proposed method is also applicable to both non-linear and non-stationary time series where the techniques based on the conventional Fourier analysis are prone to errors. Within each group that oscillate due to a common cause, the method can also indicate the location of the probable root cause of oscillations. The efficacy of the proposed method is established with the help of both simulation and industrial case studies. (C) 2018 Elsevier Ltd. All rights reserved.

关键词： Multivariate empirical mode decomposition Mode alignment grouping algorithm Dyadic filter bank property

来源：评论

学校读者我要写书评

暂无评论

On the Achievable Degrees of Freedom of Two-Cell Multiuser MIMO Interference Networks

引用

IEEE TRANSACTIONS ON COMMUNICATIONS 2014年第8期62卷 2880-2893页

作者： Chou, Hsin-Jui Chou, Che-Chen Wu, Jen-Ming Chang, Ronald Y. Natl Tsing Hua Univ Inst Commun Engn Hsinchu 30013 Taiwan Acad Sinica Res Ctr Informat Technol Innovat Taipei 115 Taiwan

In this paper, we study the sum degrees of freedom (DoF) of an uplink two-cell multiuser MIMO interference network with asymmetric number of users in the cells. The achievable DoF is devised based on a two-dimensional space-time spreading code framework with linear precoding/decoding design and finite channel extension. The derivation of the achievable DoF is shown related to a rank minimization problem, which corresponds to the minimization of the dimension of the interference subspace. The problem is solved by the proposed grouping algorithm (GA) based on aligning interfering signals into a low-dimensional subspace as a group and attaining the minimum number of groups. The achievable sum DoF derived based on the proposed GA is shown to be greater than prior arts and achieves the theoretic upper bound in several cases. We also give a closed-form expression of the maximum achievable sum DoF when there is the maximum number of admissible users in the considered finite diversity environment.

关键词： Degrees of freedom interference alignment multiuser MIMO cellular networks grouping algorithm

来源：评论

学校读者我要写书评

暂无评论

Patient classification and hospital reimbursement for inguinal hernia repair: a comparison across 11 European countries

引用

HERNIA 2014年第2期18卷 273-281页

作者： Serden, L. O'Reilly, J. Natl Board Hlth & Welf Dept Stat & Comparisons S-10630 Stockholm Sweden Econ & Social Res Inst Dublin Ireland Univ Dublin Trinity Coll Dublin 2 Ireland

This comparative study examines the categorisation of patients undergoing surgical repair of inguinal hernia in the diagnosis-related group (DRG) systems of 11 European countries (Austria, England, Estonia, Finland, France, Germany, Ireland, the Netherlands, Poland, Spain and Sweden). Understanding the design and operation of DRG systems for this common surgical procedure is important, given their increasing use internationally for hospital reimbursement and performance measurement. A common definition was used to identify inguinal hernia patients and the corresponding data were extracted from national databases. The analysis compared the variables and algorithms for classifying these patients to DRGs across the participating countries, as well as the number, composition and relative resource intensity of groups. An index case and six standardised vignettes were grouped using each country's DRG system and the associated quasi-prices were calculated. The number of groups to which inguinal hernia patients are assigned is typically three or four, but ranges from two in Poland to ten in France. In most systems, categorising patients is contingent on procedure, principal and secondary diagnoses, and age, with treatment setting (day case/inpatient) being less common. Added to these, the French system also incorporates length of stay and whether the patient died. More resource intensive DRGs generally contained patients who were older, treated as inpatients, did not die, had (more severe) complications and/or co-morbidities, and/or underwent laparoscopic repair. There are cross-country disparities in day case rates and the use of laparoscopic repairs. The categorisation of inguinal hernia patients varies across the 11 European DRG systems under study. By highlighting the main differences across these systems, this comparative analysis allows the relevant decision makers to assess the adequacy and specificity of their own DRG systems.

关键词： Diagnosis-related groups (DRGs) grouping algorithm Cost Length of stay Inguinal hernia repair

来源：评论

学校读者我要写书评

暂无评论

A new design of RF tag location and recognition

引用

International Conference on Frontiers of Manufacturing and Design Science (ICFMD2010)

作者： Liu, Meng Zhang, Ke Zhang, Wei Univ Elect Sci & Technol China Res Inst Elect Sci & Technol Chengdu 610054 Peoples R China

ISBN: (纸本)9783037850046

In RF applications, tag orientation and collision that occurred when reader identify many tags at the same time are the main problems difficult to be solved. Identification and positioning to tags will be the main application in the further too. To solve the problem, a variety of algorithms have been proposed and obtained many good results. This paper gave a new idea to solve the problem from the tag side rather than algorithm, the main method is to add some transistors to the tag circuit. The circuit can supply high accurate location information to tags and then assist the reader to know their location;in tag identification the circuit can help part the tags in several groups according their different position. In the groups using corresponding grouping algorithm can enhance the anti-collision efficiency. The simulations in this paper have improved the added circuit is helpful to the accuracy of tag orientation and efficiency of multi-tag identification. From the detail redesign scheme we can conclude there is almost no increase in costs for the improvement of tags and the production process is easy, having well economic and feasibility.

关键词： RF positioning RFID anti-collision tag redesign grouping algorithm

来源：评论

学校读者我要写书评

暂无评论

Pose Estimation for Non-Cooperative Targets using 3D Feature Correspondences Grouped via Local and Global Constraints 11

Pose Estimation for Non-Cooperative Targets using 3D Feature...

引用

11th International Symposium on Multispectral Image Processing and Pattern Recognition (MIPPR) - Pattern Recognition and Computer Vision

作者： Zhu, Angfan Yang, Jiaqi Cao, Zhiguo Wang, Li Gu, Yingying Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Natl Key Lab Sci & Technol Multispectral Informat Wuhan 430074 Peoples R China Northwestern Polytech Univ Sch Comp Sci Xian 710129 Peoples R China Natl Engn Lab Integrated AeroSp Ground Ocean Big Xian 710129 Peoples R China Beijing Inst Control Engn Beijing Peoples R China

ISBN: (纸本)9781510636385

The pose of a non-cooperative target represented by point cloud can be estimated through point cloud registration, which is generally performed by searching good correspondences. Seeking correspondences in the context of non-cooperative target pose estimation is a challenging task due to the low texture, noise and occlusion, resulting in a number of outliers in the initial correspondences. In order to gain a high quality set of feature correspondences, we employ a combination of local and global constraints to remove the outliers in initial correspondences. On a local scale, we use simple and low-level geometric invariants. On a global scale, we apply covariant constraints for finding compatible correspondences. In the experiments, we use four groups of different non-cooperative targets to evaluate our algorithm and the results verify that the quality of the correspondence set has been greatly improved by our method and the pose can be accurately estimated.

关键词： non-cooperative target correspondence grouping algorithm matching registration

来源：评论

学校读者我要写书评

暂无评论

An Exploratory Study on Group Formation Based on Learning Styles

An Exploratory Study on Group Formation Based on Learning St...

引用

13th IEEE Annual International Conference on Advanced Learning Technologies (ICALT)

作者： Liu, Shuangyan Joy, Mike Griffiths, Nathan Univ Warwick Dept Comp Sci Coventry CV4 7AL W Midlands England

ISBN: (纸本)9780769550091

Recent work has highlighted how consideration of learning styles in the process of group formation for collaborative learning can have a positive impact. This paper investigates the performance of similar learning style groups and diverse learning style groups in group work, and explores how a grouping algorithm based on students' learning styles affects students' learning achievements and processes. An empirical study with current undergraduate students in the UK has been conducted. The participating students were invited to accomplish two group discussion tasks. Each student performed the two activities, in a separate group each time, once in a group consisting of students with similar learning styles, once in a group with diverse learning styles. No volunteer shared both groups with any individual student. This paper focuses on analysing the learning achievements and collaboration processes for the two types of groups, particularly with respect to the quality of group interactions. A significant difference was found between the percentages of time spent on meaningful interactions by the two types of groups, revealing that diverse learning style groups tend to spend significantly more time on meaningful interactions than similar learning style groups. In addition, the diverse learning style groups had demonstrated significantly less negative social-emotional reactions in showing disagreements.

关键词： learning styles group formation collaborative learning grouping algorithm group work

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：