Improving Convolutional Neural Network Design via Variable Neighborhood Search

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2017

Autores

Araujo, T (Autor) Outra; Aresta, G (Autor) Outra; Bernardo Almada Lobo (Autor) FEUP; Ana Maria Mendonça (Autor) FEUP; Aurélio Campilho (Autor) FEUP

Ata de Conferência Internacional

Título: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Páginas: 371-379

14th International Conference on Image Analysis and Recognition, ICIAR 2017

5 July 2017 through 7 July 2017

Indexação

ISI Web of Knowledge - 4 Citações

Scopus - 3 Citações

Outras Informações

ID Authenticus: P-00M-WJ8

DOI: 10.1007/978-3-319-59876-5_41

Abstract (EN): An 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.

Idioma: Inglês

Tipo (Avaliação Docente): Científica