|
advertisement |
|
|
|
|
|
|
Industrial Automatic Control Systems and Controllers Annotation << Back
Features of the Architecture of an Artificial Neural Network in a Cloud Service and the Principles of their Training |
O.A. Kovaleva, M.N. Timofeev
This article discusses the features of the architecture of neural networks and the principles of their training. The effectiveness of architectures and teaching methods for solving specifi c problems is analyzed. OpenMP directives were used to parallelize the learning algorithm of the multilayer perceptron at the sample level. Here are the details of the algorithm and methodology for estimating the parameters of parallel computing constraints in the cloud service. As a result, it was concluded that the development and trends in the field of parallel learning of neural networks are increasingly being used and developed in many areas. So, as they lead to a signifi cant reduction in the time of the training process, and, depending on the system on which the training is carried out, they can show a nearly linear increase in efficiency depending on the number of processors, cores, or system threads controlled by the cloud service.
Keywords: neural network models; parallel computing; neural network; cloud storage; sociological data; big data; algorithm; expert system; neural network learning process.
DOI: 10.25791/asu.2.2021.1256
Pp. 16-21. |
|
|
|
Last news:
Выставки по автоматизации и электронике «ПТА-Урал 2018» и «Электроника-Урал 2018» состоятся в Екатеринбурге Открыта электронная регистрация на выставку Дефектоскопия / NDT St. Petersburg Открыта регистрация на 9-ю Международную научно-практическую конференцию «Строительство и ремонт скважин — 2018» ExpoElectronica и ElectronTechExpo 2018: рост площади экспозиции на 19% и новые формы контент-программы Тематика и состав экспозиции РЭП на выставке "ChipEXPO - 2018" |