检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Tsirigotis, Aris Sarantoglou, George Deligiannidis, Stavros Sánchez, Erica Gutierrez, Ana Bogris, Adonis Capmany, Jose Mesaritakis, Charis Dept. Information and Communication Systems Engineering University of the Aegean Palama 2 str Samos Karlovasi83200 Greece Dept. Informatics and Computer Engineering University of West Attica Ag. Spiridonos Egaleo12243 Greece iPronics Programmable Photonics Valencia Spain Photonics Research Labs Universitat Politècnica de València Camino de Vera Valencia46022 Spain

In this work, we present and experimentally validate a passive photonic-integrated neuromorphic accelerator that uses a hardware-friendly optical spectrum slicing technique through a reconfigurable silicon photonic mesh. The proposed scheme acts as an analogue convolutional engine, enabling information preprocessing in the optical domain, dimensionality reduction and extraction of spatio-temporal features. Numerical results demonstrate that utilizing only 7 passive photonic nodes, critical modules of a digital convolutional neural network can be replaced. As a result, a 98.6% accuracy on the MNIST dataset was achieved, with a power consumption reduction of at least 26% compared to digital CNNs. Experimental results confirm these findings, achieving 97.7% accuracy with only 3 passive nodes. Copyright © 2024, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Automatic Decomposition of a Sequential Algorithm for MapReduce Frameworks

Automatic Decomposition of a Sequential Algorithm for MapRed...

引用

IEEE Region 8 International Conference on Computational Technologies in Electrical and Electronics engineering, SIBIRCON

作者： Victor S. Simonov Marat S. Khairetdinov Dept. Computer Engineering Novosibirsk State Technical University Novosibirsk Russian Federation Dept. Geophysical Informatics Institute of Computational Mathematics and Mathematical Geophysics SB RAS Novosibirsk Russian Federation

ISBN: (纸本)9781665464819

Effective programming of parallel architectures has always been a difficult t ask. T o d ate, programming languages and technologies have been developed that simplify the programmer’s work, but do not make parallelization automatic. MapReduce is a model of programming for the development of large-scale computations with intensive use of data. There are many frameworks where the implementation of this paradigm has been recently developed. There is a need to rewrite existing serial code to use the frameworks listed. The researcher must be familiar with the problems of parallelization, the API of the framework, and also have considerable experience. This prompted us to develop a new tool that automatically translates sequential programs into ready-made versions suitable for execution in the MapReduce paradigm. The code fragment from the serial version is converted in two stages. At the first stage, the synthesis of the program, the functional specification, was made. It was necessary to find information about the calculation structure for each block of code. The result was stored as a high-level intermediate language, reminiscent of the program format for MapReduce frameworks. Checking for semantic equivalence to the original has done using the proof of the theorem. At the second stage, executable code is created, which was the result of generation from a sequential program using the Hadoop or Spark instruction set. Creating a parallelizing compiler is one way to solve this problem. This will allow you to translate code written in a different paradigm (for example, imperative code) into a parallel version for the framework. Classical compilers, such as logical plan-to-physical compilers, use pattern matching rules. The compiler contains a set of rules that identify different patterns of code input (for example, list-sequential looping) and transform consistent code.

关键词： Codes Instruction sets Semantics Transforms Manuals Programming Data models

来源：评论

学校读者我要写书评

暂无评论

Towards Stagewise and Energy-Accuracy-balanced Client Selection and Resource Allocation over Dynamic Federated Learning Networks

Towards Stagewise and Energy-Accuracy-balanced Client Select...

引用

IEEE International Conference on Communications in China (ICCC)

作者： Minghong Wu Minghui Liwang Yuhan Su Yuliang Tang Zhenzhen Jiao Xianbin Wang School of Informatics Xiamen University Xiamen China College of Electronic and Information Engineering Tongji University Shanghai China School of Electronic Science and Engineering Xiamen University Xiamen China iF-Labs Beijing Teleinfo Technology Co. Ltd. CAICT China Dept. of Electrical and Computer Engineering Western University Ontario Canada

ISBN: (数字)9798350378412

ISBN: (纸本)9798350378429

Federated Learning (FL) represents a popular distributed learning architecture that facilitates data privacy by enabling clients (e.g., mobile devices) to train a FL model collaboratively without data sharing. Existing efforts mainly focus on accuracy, delay and energy consumption over a rather stable network with certain clients, with less emphasis on the impact of frequent decision-making processes during model training. Inspired by this, this paper considers a dynamic FL network with uncertain participation of clients, while we jointly optimize client selection and communication resource allocation to achieve a balance between energy cost and FL model accuracy, as proved to be NP-hard. To facilitate timely model training, we propose a prediction-based two-stage asynchronous programming mechanism, which decouples the problem in two subproblems, corresponding to two stages. In particular, the former stage determines some long-term clients which are more stable to join in prior to practical model training process, by estimating the online probability of clients. Then, the latter stage can be implemented by involving some temporary clients as backups when long-term ones are not able to show up. Such a well-designed mechanism offers a unique veiw on FL, while enabling a responsive and cost-effective decision-making process. Comprehensive simulations regarding both IID and non-IID data distributions on MNIST and CIFAR-10 datasets can prove our commendable performance on time efficiency, energy cost and accuracy.

关键词： Training Accuracy Costs Federated learning Decision making Programming Dynamic scheduling

来源：评论

学校读者我要写书评

暂无评论

Õ(n+poly(k))-time Algorithm for Bounded Tree Edit Distance

Õ(n+poly(k))-time Algorithm for Bounded Tree Edit Distance

引用

Annual IEEE Symposium on Foundations of computer Science

作者： Debarati Das Jacob Gilbert MohammadTaghi Hajiaghayi Tomasz Kociumaka Barna Saha Hamed Saleh Dept. of Computer Science & Engineering Pennsylvania State University State College PA United States Dept. of Computer Science University of Maryland College Park MD United States Max Planck Institute for Informatics Saarland Informatics Campus Saarbrücken Germany Dept. of Computer Science & Engineering Halıcıoğlu Data Science Institute University of California San Diego San Diego CA United States

ISBN: (数字)9781665455190

ISBN: (纸本)9781665455206

Computing the edit distance of two strings is one of the most basic problems in computer science and combinatorial optimization. Tree edit distance is a natural generalization of edit distance in which the task is to compute a measure of dissimilarity between two (unweighted) rooted trees with node labels. Perhaps the most notable recent application of tree edit distance is in NoSQL big databases, such as MongoDB, where each row of the database is a JSON document represented as a labeled rooted tree and finding dissimilarity between two rows is a basic operation. Until recently, the fastest algorithm for tree edit distance ran in cubic time (Demaine, Mozes, Rossman, Weimann; TALG’10); however, Mao (FOCS’21) broke the cubic barrier for the tree edit distance problem using fast matrix *** a parameter k as an upper bound on the distance, an $\mathcal{O}(n+k^{2})$-time algorithm for edit distance has been known since the 1980s due to works of Myers (Algorithmica’86) and Landau and Vishkin (JCSS’88). The existence of an $\tilde{\mathcal{O}}(n+poly(k))$-time algorithm for tree edit distance has been posed as open question, e.g., by Akmal and Jin (ICALP’21), who give a stateof-the-art $O(nk^{2})$-time algorithm. In this paper, we answer this question positively.

关键词： computer science Upper bound Costs Databases Task analysis Optimization

来源：评论

学校读者我要写书评

暂无评论

pylspack: Parallel algorithms and data structures for sketching, column subset selection, regression and leverage scores

arXiv

引用

arXiv 2022年

作者： Sobczyk, Aleksandros Gallopoulos, Efstratios IBM Research Europe ETH Zürich Zürich Switzerland HPCLAB Computer Engineering and Informatics Dept. University of Patras Patras Greece

MSC Codes 65F50, 65F08, 65F20, 68W10, 68W20, 68P05We present parallel algorithms and data structures for three fundamental operations in Numerical Linear Algebra: (i) Gaussian and CountSketch random projections and their combination, (ii) computation of the Gram matrix and (iii) computation of the squared row norms of the product of two matrices, with a special focus on "tall-and-skinny" matrices, which arise in many applications. We provide a detailed analysis of the ubiquitous CountSketch transform and its combination with Gaussian random projections, accounting for memory requirements, computational complexity and workload balancing. We also demonstrate how these results can be applied to column subset selection, least squares regression and leverage scores computation. These tools have been implemented in pylspack, a publicly available Python package, whose core is written in C++ and parallelized with OpenMP, and which is compatible with standard matrix data structures of SciPy and NumPy. Extensive numerical experiments indicate that the proposed algorithms scale well and significantly outperform existing libraries for tall-and-skinny matrices. © 2022, CC BY-NC-ND.

关键词： Matrix algebra

来源：评论

学校读者我要写书评

暂无评论

A Load Balancing Algorithm for 5G Vehicular Cloud Computing Systems 24

A Load Balancing Algorithm for 5G Vehicular Cloud Computing ...

引用

24th Pan-Hellenic Conference on informatics, PCI 2020

作者： Zoumi, Eirini Skondras, Emmanouil Veliu, David Michalas, Angelos Vergados, Dimitrios D. Dept. of Informatics University of Piraeus Piraeus Greece Dept. of Electrical and Computer Engineering University of Western Macedonia Kozani Greece

ISBN: (纸本)9781450388979

5G Vehicular Cloud Computing (5G-VCC) infrastructures are evolving rapidly. In a 5G-VCC system, Cloud resources should be efficiently distributed to provide satisfactory Quality of Service (QoS) for modern services with increased requirements. Furthermore, the workload should be fairly distributed to the available Virtual Machines (VMs). This paper proposes an algorithm for performing load balancing in Cloud infrastructures that exist in 5G-VCC systems. The algorithm is called Modified Ant Colony Optimization (MACO), as its functionality is influenced by the natural behaviour of ants. Specifically, the MACO algorithm assigns a pheromone (weight) value to each VM. Subsequently, for each service request the VM with the highest pheromone is selected in a way similar to the one that ants apply to select optimal routes. The selection of each VM results in the decrement its pheromone value, considering the workload of the assigned service. Evaluation results show that the proposed algorithm outperforms existing load balancing algorithms in terms of the processing time required for serving the user requests. © 2020 ACM.

关键词： 5G mobile communication systems

来源：评论

学校读者我要写书评

暂无评论

Observing Generator Performance by Finite Element Analysis Considering Coil Numbers

Observing Generator Performance by Finite Element Analysis C...

引用

Technology and Policy in Energy and Electric Power (ICT-PEP), International Conference on

作者： Muhammad Afnan Habibi Mochamad Farhan Ali Irfani Langlang Gumilar Arya Kusumawardana Aripriharta Angger Baskoro Study Program of Power Plant Engineering Technology Universitas Negeri Malang Malang Indonesia Department of Electrical Engineering and Informatics Universitas Negeri Malang Malang Indonesia Dept. Electrical Engineering and Computer Sciences National Yang Ming Chiao Tung University Hsinchu Taiwan

This paper conducts a comprehensive analysis of electrical generator performance using Finite Element Analysis (FEA), with a specific emphasis on the role of coil numbers in influencing generator efficiency and functionality. In the context of modern energy systems, where efficient power generation is paramount, this research aims to elucidate the relationship between the number of coils within a generator and its overall performance, including power output and electromagnetic behavior. Through systematic FEA simulations that vary coil numbers while keeping other parameters constant, this study provides valuable insights into the trade-offs associated with increased coil numbers and enhanced efficiency. These findings have significant implications for optimizing generator designs across various applications, from renewable energy systems to industrial power generation, ultimately advancing our understanding of generator dynamics and contributing to more sustainable and efficient power generation technologies.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Performance Comparison of 15-Level Multilevel Inverter Topologies

Performance Comparison of 15-Level Multilevel Inverter Topol...

引用

International Conference on Electronics, computers and Artificial Intelligence (ECAI)

作者： G. Anusha Krishan Arora Himanshu Sharma Surya Prakash Thota Chaman Verma Maria Simona Raboaca Constantin Filote Dept. of EEE Lovely Professional University Punjab India Dept. of EEE GNITS Hyderabad India Dept. of Media and Educational Informatics Faculty of Informatics Eötvös Loránd University Budapest Hungary ICSI Energy Department National Research and Development Institute for Cryogenics and Isotopic Technologies Ramnicu Valcea Romania Faculty of Electrical Engineering and Computer Science Stefan cel Mare University of Suceava Suceava Romania

Multilevel inverters (MLI) have gained a significant emphasis in the area of renewable energy applications in recent years. The popularity of MLIs is due to their ability to efficiently convert the sustainable sources of energy, such as wind and solar power, into usable electrical energy. One key advantage of using reduced multilevel inverter configurations is that it can help to reduce the system overall cost by decreasing number of components required. The reduction of components in the inverter circuit results in simplification of design and lower expenditure. This paper proposes a fifteen-level modified multilevel inverter with cascaded h-bridge with a minimized switch count. The simulation results of the proposed inverter demonstrate that it is able to effectively attenuate harmonics and with reduced number of components required while still providing good performance. The suggested inverter design is also evaluated on its harmonic reduction characteristics, providing a cleaner output waveform and less distortion.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Healthcare Industry 5.0: Pareto Optimal IoT-Based Health Monitoring With Federated Learning

引用

IEEE Internet of Things Journal 2025年

作者： Diamantoulaki, Ioanna Tegos, Sotiris A. Bouzinis, Pavlos S. Sarigiannidis, Panagiotis Chatzakis, Christos Petousis, Stamatios Maglaveras, Nicos Karagiannidis, George K. Aristotle University of Thessaloniki Department of Electrical and Computer Engineering Thessaloniki54124 Greece Kila Kozani MetaMind Innovation P.C. Kozani50100 Greece University of Western Macedonia Department of Informatics and Telecommunications Engineering Kozani50100 Greece Aristotle University of Thessaloniki AUTH Campus Medical School 2nd Dept. of Obstetrics and Gynaecology Thessaloniki54636 Greece

With the recent expansion of patient data availability and storage capabilities, healthcare entities tend to store an increasing amount of medical data locally. In addition, ongoing advances in the era of healthcare industry 5.0 and Artificial Intelligence of Things can lead to more efficient use of medical data, resulting in better health monitoring at lower costs. However, due to strict privacy restrictions related to the sensitive nature of medical data, it is often only used locally and ultimately underutilized. To this end, federated learning (FL) offers a promising solution for the efficient use of medical data, facilitating the development of reliable and robust healthcare tools. This can be achieved thanks to its decentralized nature, which allows participating entities to collaborate and thus develop and train a centralized shared model without requiring data sharing. Taking this into account, we present a Pareto-front optimization framework for FL-based health monitoring that is able to mitigate false negative predictions for the required level of false positives. By applying the proposed framework to four different medical applications, it is shown that the risk of misdiagnosis is significantly reduced, providing an additional tool for medical professionals. © 2014 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

High-speed photonic neuromorphic computing using recurrent optical spectrum slicing neural networks

引用

Communications engineering 2022年第1期1卷 1-10页

作者： Sozos, Kostas Bogris, Adonis Bienstman, Peter Sarantoglou, George Deligiannidis, Stavros Mesaritakis, Charis Dept. of Informatics and Computer Engineering University of West Attica Egaleo Greece Dept. of Information Technology Ghent University-imec Gent Belgium Dept. Information and Communication Systems Engineering Engineering School University of the Aegean Samos Greece

Neuromorphic computing using photonic hardware is a promising route towards ultrafast processing while maintaining low power consumption. Here we present and numerically evaluate a hardware concept for realizing photonic recurrent neural networks and reservoir computing architectures. Our method, called Recurrent Optical Spectrum Slicing Neural Networks (ROSS-NNs), uses simple optical filters placed in a loop, where each filter processes a specific spectral slice of the incoming optical signal. The synaptic weights in our scheme are equivalent to the filters’ central frequencies and bandwidths. Numerical application to high baud rate optical signal equalization (>100 Gbaud) reveals that ROSS-NN extends optical signal transmission reach to > 60 km, more than four times that of two state-of-the-art digital equalizers. Furthermore, ROSS-NN relaxes complexity, requiring less than 100 multiplications/bit in the digital domain, offering tenfold reduction in power consumption with respect to these digital counterparts. ROSS-NNs hold promise for efficient photonic hardware accelerators tailored for processing high-bandwidth (>100 GHz) optical signals in optical communication and high-speed imaging applications. © The Author(s) 2022.

关键词： computer science Electrical and electronic engineering Photonic devices

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：