?
Использование технологии CUDA в обучении сверточной нейросети для распознавания пыльцевых зерен
С. 70–81.
Замятина Елена Борисовна, Ханжина Н. Е.
In this work, we describe the problem of automated pollen recognition using images from lighting microscope. Automated pollen recognition related to such important tasks as honey quality control, air quality control for helping to asthma and allergy patients, paleopalynology, forensic palynology. We describe the problem solution based on machine learning and CUDA. Extracted features and preprocessing steps are described. Results are compared on dataset of 5 specie. The best model is convolutional neural network with 89% of accuracy. Its performance was particularly up twice using CUDA.
In book
Пермь: Пермский государственный национальный исследовательский университет, 2016.
Choi Y. R., Malkovsky S., Stegailov V., , in: 11th Russian Supercomputing Days, RuSCDays 2025, Moscow, Russia, September 29–30, 2025, Revised Selected Papers.: Springer, 2026. Ch. 3 P. 32–47.
Multi-GPU servers often exhibit uneven characteristics. For instance, the data transfer bandwidth between four NVIDIA V100 GPUs can vary due to the NVLink connecting these devices to a specific CPU in servers with IBM POWER 9 processors, which means that the communication bandwidth between other devices is comparably slower. To address this issue, the Multi-GPU ...
Added: January 3, 2026
Pikul A. S., Безопасность информационных технологий 2024 Т. 31 № 4 С. 116–127
This article explores the potential use of modern computer vision architectures for the task of deepfake detection. The following architectures are considered: EfficientNet, Vision Transformer (ViT), VisionLSTM (ViL), Vision KAN, and Mamba Vision. The novelty of the approach lies in the application and comparison of these architectures, as well as their combination into paired ensembles ...
Added: December 12, 2025
Choi Y. R., Мальковский С. И., Stegailov V., В кн.: "Информационные технологии и высокопроизводительные вычисления": Материалы VIII Международной научно-практической конференции, Хабаровск, 15-17 сентября 2025 г.: Хабаровск: Хабаровский Федеральный исследовательский центр, 2025. Гл. 81 С. 317–320.
Работа посвящена разработке и экспериментальному исследованию параллельных алгоритмов матричного умножения и матричной экспоненты с асинхронным обменом данными, использующих принцип наложения вычислений и коммуникаций для максимизации производительности, для систем с несколькими графическими ускорителями и неоднородной топологией. Также представлены теоретические модели оптимизации размера блоков для повышения эффективности расчетов. Алгоритм матричной экспоненты реализован с поддержкой комплексных матриц через ...
Added: October 15, 2025
Seleznev L. E., Chupakhin A. A., Kostenko V. A. et al., Optical Memory and Neural Networks (Information Optics) 2023 Vol. 32 No. 2 P. 73–85
We analyze a classification problem of mentally pronounced Russian phonemes based on data obtained by means of an electroencephalography device. We describe the data collection method as well as the methods of the obtained data processing. To solve the small sample size problem we present the augmentation techniques that use the time stretching and the ...
Added: October 2, 2025
Zabolotniy A., Chan R. W., Moiseeva V. et al., Frontiers in Neuroscience 2025 Vol. 19 Article 1623380
We demonstrated the feasibility of finger movement decoding with a tailored Convolutional Neural Network. The performance of our approach was comparable to complex deep learning architectures, while providing faster and interpretable outcome. This algorithmic strategy holds high potential for the investigation of the mechanisms underlying non-invasive neurophysiological recordings in cognitive neuroscience. ...
Added: October 2, 2025
Tsybina Y., Kastalskiy I., Krivonosov M. et al., Neural Computing and Applications 2023 Vol. 34 No. 11 P. 9147 –9160
Modeling the neuronal processes underlying short-term working memory remains the focus of many theoretical studies in neuroscience. In this paper, we propose a mathematical model of a spiking neural network (SNN) which simulates the way a fragment of information is maintained as a robust activity pattern for several seconds and the way it completely disappears ...
Added: April 9, 2025
Morozov D., Garipov T., Lyashevskaya O. et al., Journal of Language and Education 2024 Vol. 10 No. 4 P. 71–84
Introduction: Numerous algorithms have been proposed for the task of automatic morpheme segmentation of Russian words. Due to the differences in task formulation and datasets utilized, comparing the quality of these algorithms is challenging. It is unclear whether the errors in the models are due to the ineffectiveness of algorithms themselves or to errors and inconsistencies ...
Added: January 7, 2025
Teplyakov L., Kaymakov K., Shvets E. et al., SPIE, 2021.
Line detection is an important computer vision task traditionally solved by Hough Transform. With the advance of deep learning, however, trainable approaches to line detection became popular. In this paper we propose a lightweight CNN for line detection with an embedded parameter-free Hough layer, which allows the network neurons to have global strip-like receptive fields. ...
Added: November 5, 2024
Demidovskij A., Artyom Tugaryov, Aleksei Trutnev et al., Mathematics 2023 Vol. 14 No. 11 Article 3120
Due to industrial demands to handle increasing amounts of training data, lower the cost of computing one model at a time, and lessen the ecological effects of intensive computing resource consumption, the job of speeding the training of deep neural networks becomes exceedingly challenging. Adaptive Online Importance Sampling and IDS are two brand-new methods for ...
Added: September 12, 2023
Vilkov E. A., Byshevski-Konopko O. A., Kalyabin D. et al., Journal of Physics: Condensed Matter 2023 Vol. 35 No. 43 Article 435001
The spectral properties of gap electroacoustic waves in a PT-symmetric structure of piezoelectrics of symmetry class 6 mm separated by a gap are theoretically investigated. The spectra were calculated for lead germanate (non-zero transverse piezoactivity) and barium titanate (symmetry class 4 mm—zero transverse piezoactivity). It has been established that at a certain level of losses ...
Added: August 18, 2023
Choi Y. R., Stegailov V., , in: 22nd International Conference, MMST 2022, Nizhny Novgorod, Russia, November 14–17, 2022, Revised Selected Papers.: Springer, 2022. Ch. 23 P. 281–292.
Modern types of multi-GPU servers combine up to 8 A100 GPUs connected by NVLink 3.0 links through NVSwitch. This connectivity provides unprecedented capabilities for multi-GPU algorithms. In this work, we analyze the performance of matrix-matrix multiplication algorithm developed by us previously. Tuning principles and limits for maximum performance are discussed. Algorithm performance for much more ...
Added: December 26, 2022
Головин Р. Д., Zunin V., В кн.: Межвузовская научно-техническая конференция студентов, аспирантов и молодых специалистов имени Е.В. Арменского. Материалы конференции.: М.: МИЭМ НИУ ВШЭ, 2021. С. 120–122.
В работе проводится обзор различных способов реализации нейронной сети для распознавания изображения на примере набора данных MNIST. Дается анализ используемых библиотек и методов их применения в области распознавания объектов. ...
Added: September 25, 2022
Choi Y. R., Nikolskiy V., Stegailov V., , in: Parallel Computational Technologies: 16th International Conference, PCT 2022, Dubna, Russia, March 29–31, 2022, Revised Selected Papers.: Springer, 2022. Ch. 12 P. 158–171.
Added: August 11, 2022
Savchenko A., Belova N. S., Expert Systems with Applications 2022 Vol. 207 Article 117885
In this paper, the computational complexity of the probabilistic neural network for the classification of high-dimensional data is improved. At first, the class probability densities are estimated by using only a few principal components of an observed point. The Gaussian–Parzen kernel is replaced by the orthogonal series estimates of class-conditional densities for each principal component using the Fourier series to speed ...
Added: June 29, 2022
Ziganurova L., Shchur L., Lobachevskii Journal of Mathematics 2022 Vol. 43 No. 2 P. 513–518
The Lattice Boltzmann method (LBM) is the alternative approach for hydrodynamic
equation solving. Two factors make it a favorite approach nowadays. Firstly, the attractive feature
of LBM is that it is intrinsic for parallel simulations due to the linear structure of the algorithm.
Secondly, what makes LBM special for the research, it is well applicable to the simulations ...
Added: May 25, 2022
Demochkina P., Savchenko A., , in: Pattern Recognition. ICPR International Workshops and Challenges. Virtual Event, January 10–15, 2021, Proceedings, Part V.: Springer, 2021. P. 266–274.
In this paper, we address the emotion classification problem in videos using a two-stage approach. At the first stage, deep features are extracted from facial regions detected in each video frame using a MobileNet-based image model. This network has been preliminarily trained to identify the age, gender, and identity of a person, and further fine-tuned ...
Added: April 10, 2022
Turtsynskiy M., Вестник Московского университета. Серия 1: Математика. Механика 2021 № 1 С. 24–29
Исследован специальный подкласс решений трехмерной системы уравнений идеального политропного газа, отвечающей модели атмосферы. Свойства таких решений полностью характеризуются нелинейной системой обыкновенных дифференциальных уравнений высокого порядка. Установлено, что, в отличие от соответствующей двумерной модели, все особые точки такой системы неустойчивы. Найдены некоторые первые интегралы. Показано, что в случае осевой симметрии система сводится к одному уравнению. При ...
Added: October 31, 2021
Turtsynskiy M., Управление большими системами: сборник трудов 2020 № 84 С. 51–65
Рассмотрена двумерная по пространству система уравнений идеального политропного газа на вращающейся плоскости, возникающая в задачах динамики атмосферы. В общей постановке система очень сложна, однако она допускает решения с линейным по пространственным переменным профилем скорости (отвечающим движениям с однородной деформацией), нахождение которых сводится к решению квадратично-нелинейной системы обыкновенных дифференциальных уравнений. Эта система обладает двумя семействами особых ...
Added: October 31, 2021
Turtsynskiy M., Вестник Московского университета. Серия 1: Математика. Механика 2020 № 2 С. 37–43
Для системы уравнений идеального политропного газа на равномерно вращающейся плоскости, записанной в лагранжевых координатах, найдены первые интегралы, соответствующие движению с однородной деформацией. Показано, что в случае показателя адиабаты, равного двум, исходная система из четырех нелинейных обыкновенных дифференциальных уравнений второго порядка может быть сведена к одному уравнению первого порядка и решение может быть найдено как функция ...
Added: October 31, 2021
Churaev E., Savchenko A., , in: 2021 International Russian Automation Conference (RusAutoCon).: IEEE, 2021. P. 633–638.
In this paper, we examine the issue of video-based facial emotion recognition algorithms which show excellent performance on some benchmarks, but have much worse accuracy in practical applications. For example, the typical error rate of contemporary deep neural networks on the RAVDESS dataset is less than 5%. We argue that such results are obtained only ...
Added: October 7, 2021