Please use this identifier to cite or link to this item: http://repositorio.inesctec.pt/handle/123456789/6074
Full metadata record
DC FieldValueLanguage
dc.contributor.authorTeresa Finisterra Araújoen
dc.contributor.authorGuilherme Moreira Arestaen
dc.contributor.authorBernardo Almada-Loboen
dc.contributor.authorAna Maria Mendonçaen
dc.contributor.authorAurélio Campilhoen
dc.date.accessioned2018-01-14T17:05:02Z-
dc.date.available2018-01-14T17:05:02Z-
dc.date.issued2017en
dc.identifier.urihttp://repositorio.inesctec.pt/handle/123456789/6074-
dc.identifier.urihttp://dx.doi.org/10.1007/978-3-319-59876-5_41en
dc.description.abstractAn unsupervised method for convolutional neural network (CNN) architecture design is proposed. The method relies on a variable neighborhood search-based approach for finding CNN architectures and hyperparameter values that improve classification performance. For this purpose, t-Distributed Stochastic Neighbor Embedding (t-SNE) is applied to effectively represent the solution space in 2D. Then, k-Means clustering divides this representation space having in account the relative distance between neighbors. The algorithm is tested in the CIFAR-10 image dataset. The obtained solution improves the CNN validation loss by over 15% and the respective accuracy by 5%. Moreover, the network shows higher predictive power and robustness, validating our method for the optimization of CNN design. © Springer International Publishing AG 2017.en
dc.languageengen
dc.relation6381en
dc.relation6071en
dc.relation5428en
dc.relation6321en
dc.relation6320en
dc.rightsinfo:eu-repo/semantics/embargoedAccessen
dc.titleImproving convolutional neural network design via variable neighborhood searchen
dc.typeconferenceObjecten
dc.typePublicationen
Appears in Collections:C-BER - Articles in International Conferences
CEGI - Articles in International Conferences

Files in This Item:
File Description SizeFormat 
P-00M-WJ8.pdf
  Restricted Access
887.12 kBAdobe PDFView/Open Request a copy


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.