检索结果-内蒙古大学图书馆

ACMFed: Fair Semi-Supervised Federated Learning With Additional Compromise Model

IEEE ACCESS 2025年 13卷 47734-47747页

作者： Kim, Dohyoung Lee, Kangyoon Lee, Youngho Woo, Hyekyung Gachon Univ Dept IT Convergence Engn Seongnam Si 13120 South Korea Gachon Univ Dept Comp Engn Seongnam Si 13120 South Korea Kongju Natl Univ Dept Hlth Adm Gongju Si 32588 South Korea

One of the major drawbacks of federated learning (FL) is data imbalance and uneven reliability of clients, which adversely impacts model performance and generalization ability. Among these problems, data imbalance can be addressed through semi-supervised federated learning (SSFL), in which fully labeled, and fully unlabeled clients are jointly trained to derive a global model. Existing approaches work well when local clients have independent identically distributed (IID) data but fail to generalize under a more practical SSFL setting, namely fairness, non-IID. Since the non-IID problem stems from unfairness, to address this issue, we propose ACMFed, a fair and improved concurrent federated learning method that balances client connections, computing resources, data quality and quantity based on fairness. ACMFed reaches the compromise by performing fair multiple sampling of clients in models with largely deviating labeled and unlabeled clients. Each client distills the model through a multi-sub-sampling process without aggregating the local model to the central server, and builds a global model based on it. The additional compromise model applies a fair adjustment method of the new distance-based dynamic model to reduce data bias and optimize performance. ACMFed outperforms other federated learning methods and achieves faster convergence on six benchmark datasets. To code is available at https://***/kimdohyoung96/ACMFed.

关键词： Data models Federated learning Computational modeling Servers Deep learning Data privacy Training Standards Solid modeling distributed databases distributed computing methodologies fair federated learning multi-sampling protected health information semi-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Commodity single board computer clusters and their applications

引用

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE 2018年 89卷 201-212页

作者： Johnston, Steven J. Basford, Philip J. Perkins, Colin S. Herry, Herry Tso, Fung Po Pezaros, Dimitrios Mullins, Robert D. Yoneki, Eiko Cox, Simon J. Singer, Jeremy Univ Southampton Fac Engn & Phys Sci Southampton SO167 QF Hants England Univ Glasgow Sch Comp Sci Glasgow G12 8QQ Lanark Scotland Loughborough Univ Dept Comp Sci Loughborough LE11 3TU Leics England Univ Cambridge Comp Lab Cambridge CB3 0FD England

Current commodity Single Board Computers (SBCs) are sufficiently powerful to run mainstream operating systems and workloads. Many of these boards may be linked together, to create small, low-cost clusters that replicate some features of large data center clusters. The Raspberry Pi Foundation produces a series of SBCs with a price/performance ratio that makes SBC clusters viable, perhaps even expendable. These clusters are an enabler for Edge/Fog Compute, where processing is pushed out towards data sources, reducing bandwidth requirements and decentralizing the architecture. In this paper we investigate use cases driving the growth of SBC clusters, we examine the trends in future hardware developments, and discuss the potential of SBC clusters as a disruptive technology. Compared to traditional clusters, SBC clusters have a reduced footprint, are low-cost, and have low power requirements. This enables different models of deployment particularly outside traditional data center environments. We discuss the applicability of existing software and management infrastructure to support exotic deployment scenarios and anticipate the next generation of SBC. We conclude that the SBC cluster is a new and distinct computational deployment paradigm, which is applicable to a wider range of scenarios than current clusters. It facilitates Internet of Things and Smart City systems and is potentially a game changer in pushing application logic out towards the network edge. (C) 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license.

关键词： Raspberry Pi Edge computing Networks cloud computing Centralization/decentralization distributed computing methodologies Multicore architectures Emerging architectures

来源：评论

学校读者我要写书评

暂无评论

Fast and scalable distributed deep convolutional autoencoder for fMRI big data analytics

引用

NEUROcomputing 2019年 325卷 20-30页

作者： Makkie, Milad Huang, Heng Zhao, Yu Vasilakos, Athanasios V. Liu, Tianming Univ Georgia Comp Sci Dept Athens GA 30602 USA Northwestern Polytech Univ Sch Automat Xian Shaanxi Peoples R China Lulea Univ Technol Comp Sci Dept Lulea Sweden

In recent years, analyzing task-based fMRI (tfMRI) data has become an essential tool for understanding brain function and networks. However, due to the sheer size of tfMRI data, its intrinsic complex structure, and lack of ground truth of underlying neural activities, modeling tfMRI data is hard and challenging. Previously proposed data modeling methods including Independent Component Analysis (ICA) and Sparse Dictionary Learning only provided shallow models based on blind source separation under the strong assumption that original fMRI signals could be linearly decomposed into time series components with corresponding spatial maps. Given the Convolutional Neural Network (CNN) successes in learning hierarchical abstractions from low-level data such as tfMRI time series, in this work we propose a novel scalable distributed deep CNN autoencoder model and apply it for fMRI big data analysis. This model aims to both learn the complex hierarchical structures of the tfMRI big data and to leverage the processing power of multiple GPUs in a distributed fashion. To deploy such a model, we have created an enhanced processing pipeline on the top of Apache Spark and Tensorflow, leveraging from a large cluster of GPU nodes over cloud. Experimental results from applying the model on the Human Connectome Project (HCP) data show that the proposed model is efficient and scalable toward tfMRI big data modeling and analytics, thus enabling data-driven extraction of hierarchical neuroscientific information from massive fMRI big data. (C) 2018 Elsevier B.V. All rights reserved.

关键词： Data mining Neural networks distributed computing methodologies Machine learning

来源：评论

学校读者我要写书评

暂无评论

Caching in Location Based Services: Approaches, Challenges and Emerging Trends

引用

WIRELESS PERSONAL COMMUNICATIONS 2024年第3期135卷 1581-1615页

作者： Gupta, Ajay K. Shanker, Udai M M M Univ Technol Dept CSE Gorakhpur 273010 India

Location based service (LBS) is related basically to a local, continuous, and spatially confined computation in the context-aware mobile environment. Hitherto, serving a query within the specified timeline becomes possible with the help of caching the data at the client and/or server sites. It enhances the performance of LBSs by the reduction in the network traffic, access latency, server load etc.;however, at the same time, it faces difficulty in maintaining the database consistency as it works in the environment, where frequent disconnections occur. As a result, many fascinating LBS caching issues created in the mobile environment have forced the mobile database specialists and researchers to do extensive research efforts to improve the response time and efficient storage resources by developing the effective location-based technologies. Thus, this study specifically discusses the aforementioned issues with comparative analysis over the common attributes such as highlighting limitations/strength, recent advancement etc. and also possible research directions for the further investigation of the unanswered questions.

关键词： Applications & services Location-based service Mobile computing Mobile environment Spatial database Spatio-temporal caching distributed computing methodologies

来源：评论

学校读者我要写书评

暂无评论

Adaptive Compression-Aware Split Learning and Inference for Enhanced Network Efficiency

引用

ACM TRANSACTIONS ON INTERNET TECHNOLOGY 2024年第4期24卷 1-26页

作者： Mudvari, Akrit Vainio, Antero Ofeidis, Iason Tarkoma, Sasu Tassiulas, Leandros Yale Univ New Haven CT 06520 USA Univ Helsinki Helsinki Finland

The growing number of AI-driven applications in mobile devices has led to solutions that integrate deep learning models with the available edge-cloud resources. Due to multiple benefits such as reduction in on-device energy consumption, improved latency, improved network usage, and certain privacy improvements, split learning, where deep learning models are split away from the mobile device and computed in a distributed manner, has become an extensively explored topic. Incorporating compression-aware methods (where learning adapts to compression level of the communicated data) has made split learning even more advantageous. This method could even offer a viable alternative to traditional methods, such as federated learning techniques. In this work, we develop an adaptive compression-aware split learning method ("deprune") to improve and train deep learning models so that they are much more network-efficient, which would make them ideal to deploy in weaker devices with the help of edge-cloud resources. This method is also extended ("prune") to very quickly train deep learning models through a transfer learning approach, which tradesoff little accuracy for much more network-efficient inference abilities. We show that the "deprune" method can reduce network usage by 4x when compared with a split-learning approach (that does not use our method) without loss of accuracy, while also improving accuracy over compression-aware split-learning by up to 4 percent. Lastly, we show that the "prune" method can reduce the training time for certain models by up to 6x without affecting the accuracy when compared against a compression-aware split-learning approach.

关键词： computing methodologies- Machine learning distributed computing methodologies Networks- Network management

来源：评论

学校读者我要写书评

暂无评论

distributed Percolation Analysis for Turbulent Flows 9

Distributed Percolation Analysis for Turbulent Flows

引用

9th IEEE Symposium on Large Data Analysis and Visualization (LDAV)

作者： Friederici, Anke Kopp, Wiebke Atzori, Marco Vinuesa, Ricardo Schlatter, Philipp Weinkauf, Tino KTH Royal Inst Technol Div Computat Sci & Technol Stockholm Sweden KTH Mech Linne FLOW Ctr Stockholm Sweden

ISBN: (纸本)9781728126050

Percolation analysis is a valuable tool to study the statistical properties of turbulent flows. It is based on computing the percolation function for a derived scalar field, thereby quantifying the relative volume of the largest connected component in a superlevel set for a decreasing threshold. We propose a novel memory-distributed parallel algorithm to finely sample the percolation function. It is based on a parallel version of the union-find algorithm interleaved with a global synchronization step for each threshold sample. The efficiency of this algorithm stems from the fact that operations in-between threshold samples can be freely reordered, are mostly local and thus require no inter-process communication. Our algorithm is significantly faster than previous algorithms for this purpose, and is neither constrained by memory size nor number of compute nodes compared to the conceptually related algorithm for extracting augmented merge trees. This makes percolation analysis much more accessible in a large range of scenarios. We explore the scaling of our algorithm for different data sizes, number of samples and number of MPI processes. We demonstrate the utility of percolation analysis using large turbulent flow data sets.

关键词： computing methodologies distributed computing methodologies distributed algorithms Mathematics of computing Discrete mathematics Graph theory Paths and connectivity problems

来源：评论

学校读者我要写书评

暂无评论

Causality Tracking Trade-offs for distributed Storage 19

Causality Tracking Trade-offs for Distributed Storage

引用

IEEE 19th International Symposium on Network computing and Applications (NCA)

作者： Guerreiro, Hugo Rodrigues, Luis Preguica, Nuno Quental, Nivia Univ Lisbon Inst Super Tecn INESC ID Lisbon Portugal Univ Nova Lisboa Fac Ciencias & Tecnol NOVA LINCS Lisbon Portugal INESC ID Lisbon Portugal

ISBN: (纸本)9781728183268

After the seminal paper by L. Lamport, which introduced (scalar) logical clocks, several other data structures for keeping track of causality in distributed systems have been proposed, including vector and matrix clocks. These are able to capture causal dependencies with more detail but, unfortunately, also consume a substantially larger amount of network bandwidth and storage space than Lamport clocks. This raises the question of whether the benefits of these more complex structures are worth their cost. We address this question in the context of partially replicated systems. We show that for some workloads the use of more expensive clocks does bring significant benefits and that for other workloads no visible benefits can be observed. The paper provides a characterization of the scenarios where each type of clock is more beneficial, helping designers to develop more efficient distributed storage systems.

关键词： distributed computing methodologies distributed Storage Network algorithm causal Consistency

来源：评论

学校读者我要写书评

暂无评论

SNTA'22: The 5thWorkshop on Systems and Network Telemetry and Analytics 22

SNTA'22: The 5thWorkshop on Systems and Network Telemetry an...

引用

31st International Symposium on High-Performance Parallel and distributed computing (HPDC)

作者： Kim, Jinoh Cafaro, Massimo Chou, Jerry Sim, Alex Texas A&M Univ Commerce TX 75428 USA Univ Salento Lecce LE Italy Natl Tsing Hua Univ Hsinchu Taiwan Lawrence Berkeley Natl Lab Berkeley CA USA

ISBN: (纸本)9781450391993

HPC and distributed systems are the driving force for the advancement of many emerging technologies, such as exascale systems, quantum machines, terabit networking, 5G/6G wireless, and cloud/edge computing. The tasks of systems and network telemetry are a key element for effective operations and management of the advancement of many emerging systems and technologies, and require more scalable telemetry and analysis techniques for comprehensive monitoring and analysis. Various input sources such as end systems, switches, firewalls, intrusion sensors and the emerging network elements speaking with different syntax and semantics make organizing and incorporating the generated data challenging for the quantitative and qualitative analysis. This workshop looks for new approaches and methods at the intersection of HPC systems and data sciences to address these difficult challenges of emerging technologies from the diverse angles of systems/network performance, availability, reliability, and security.

关键词： distributed computing methodologies parallel computing methodologies intrusion/anomaly detection and malware mitigation network security network performance evaluation

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：