Bibliografia publikacji pracowników
Państwowej Szkoły Wyższej w Białej Podlaskiej
Baza tworzona przez Bibliotekę Akademii Bialskiej im. Jana Pawła II.
Zapytanie:
UNSUPERVISED LEARNING Liczba odnalezionych rekordów: 2
Przejście do opcji zmiany formatu | Wyświetl/ukryj etykiety | Wyświetlenie wyników w wersji do druku | Pobranie pliku do edytora | Nowe wyszukiwanie Streszczenie: Unsupervised learning based on restricted Boltzmann machine or autoencoders has become an important research domain in the area of neural networks. In this paper mathematical expressions to adaptive learning step calculation for RBM with ReLU transfer function are proposed. As a result, we can automatically estimate the step size that minimizes the loss function of the neural network and correspondingly update the learning step in every iteration. We give a theoretical justification for the proposed adaptive learning rate approach, which is based on the steepest descent method. The proposed technique for adaptive learning rate estimation is compared with the existing constant step and Adam methods in terms of generalization ability and loss function. We demonstrate that the proposed approach provides better performance.
2/2
Nr opisu: 0000046270 Autorzy: Vladimir Golovko, Egor Mikhno, Aliaksandr Kroshchanka, Marta Chodyka, Piotr Lichograj. Tytuł pracy: Adaptive Learning Rate for Unsupervised Learning of Deep Neural Networks Tytuł całości: W: 2023 International Joint Conference on Neural Networks (IJCNN) Proceedings, 18-23 June 2023, Gold Coast, Australia Miejsce wydania: [Piscataway] Wydawca: Institute of Electrical and Electronics Engineers Inc. Rok wydania: 2023 Strony zajęte przez pracę: P. 1-6 ISBN: 978-1-6654-8867-9 Charakterystyka formalna: referat w materiałach pokonferencyjnych zagranicznych Charakterystyka merytoryczna: konferencja naukowa międzynarodowa Charakterystyka wg MNiSW: publikacja w recenzowanych mat. konf. międzynar. z wykazu Język publikacji: ENG Punktacja ministerstwa: 70.000 Słowa kluczowe ang.: adaptive training step ; deep learning ; unsupervised learning https://ieeexplore.ieee.org/document/10191642 DOI: 10.1109/IJCNN54540.2023.10191642 Streszczenie: In this paper an approach for adaptive learning step calculation using ReLU transfer function in neural network is proposed. This adaptive learning rate aims to automatically choose the step size that minimizes the objective function of neural network. We give a theoretical justification for the proposed adaptive learning rate approach, which is based on the steepest descent method. The main contribution of this paper is a novel technique for adaptive learning rate calculation, if we use ReLU transfer function. The experiments in data compression datasets show that proposed approach provides better generalization capability (test set accuracy) and permits to choose the learning rate automatically.