Adaptive Optimization of Training Datasets for Neural Network Image Classification
Advances in Cyber-Physical Systems · 29.05.2026 · 11 (1) · 97-102
DOI: 10.23939/acps2026.01.097 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advances in Cyber-Physical Systems, online erschienen am 25.05.2026. Verfasst von Oleksiy Lutsyk und 2 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben.
Abstract (Originalsprache)
This paper considers the problem of adaptive optimization of the training dataset for neural networks in image classification tasks. It has been shown that using the full training dataset is not the best solution. Different training samples have different value for the model. Some samples are representative, some are important for class boundaries, some preserve diversity, and some may be noisy or suspicious. A method for adaptive formation of the active training subset has been proposed. Unlike traditional methods, the proposed approach evaluates the role of each sample during training. For this purpose, the usefulness of a sample, its current role, and its noise risk have been taken into account. The method also uses a temporary buffer for suspicious samples with their later reevaluation. This makes it possible to move from static data selection to adaptive control of the training dataset. It has been shown that the proposed approach can reduce the amount of training data, decrease training time and computation cost, and keep classification accuracy at a high level. In some cases, it can also improve the generalization ability of the neural network, especially for complex multi class datasets.
Zitieren
APA
Lutsyk, O., Kosarevych, R., & Rusyn, B. (2026). Adaptive Optimization of Training Datasets for Neural Network Image Classification. Advances in Cyber-Physical Systems, 11(1), 97-102. https://doi.org/10.23939/acps2026.01.097
Vancouver
Lutsyk O, Kosarevych R, Rusyn B. Adaptive Optimization of Training Datasets for Neural Network Image Classification. Advances in Cyber-Physical Systems. 2026;11(1):97-102. doi:10.23939/acps2026.01.097
BibTeX
@article{Lutsyk2026,
author = {Lutsyk, Oleksiy and Kosarevych, Rostyslav and Rusyn, Bohdan},
title = {Adaptive Optimization of Training Datasets for Neural Network Image Classification},
journal = {Advances in Cyber-Physical Systems},
year = {2026},
volume = {11},
number = {1},
pages = {97--102},
doi = {10.23939/acps2026.01.097}
}
EUVIMED