Quaternion-based Deep Belief Networks fine-tuning

Autor(es)

Papa, Joao Paulo ^[1] ; Rosa, Gustavo H. ^[1] ; Pereira, Danillo R. ^[1] ; Yang, Xin-She ^[2]

Afiliação do(s) autor(es)

^[1] Sao Paulo State Univ, Dept Comp, Av Eng Luiz Edmundo Carrijo Coube 14-01, BR-17033360 Bauru, SP - Brazil

^[2] Middlesex Univ, Sch Sci & Technol, London NW4 4BT - England

• Tipo de documento: Artigo Científico
• Fonte: APPLIED SOFT COMPUTING; v. 60, p. 328-335, NOV 2017.
• Citações Web of Science: 6

Resumo

Deep learning techniques have been paramount in the last years, mainly due to their outstanding results in a number of applications. In this paper, we address the issue of fine-tuning parameters of Deep Belief Networks by means of meta-heuristics in which real-valued decision variables are described by quaternions. Such approaches essentially perform optimization in fitness landscapes that are mapped to a different representation based on hypercomplex numbers that may generate smoother surfaces. We therefore can map the optimization process onto a new space representation that is more suitable to learning parameters. Also, we proposed two approaches based on Harmony Search and quaternions that outperform the state-of-the-art results obtained so far in three public datasets for the reconstruction of binary images. (C) 2017 Elsevier B.V. All rights reserved. (AU)

• Processo FAPESP: 14/12236-1 - AnImaLS: Anotação de Imagem em Larga Escala: o que máquinas e especialistas podem aprender interagindo?
• Beneficiário: Alexandre Xavier Falcão
• Linha de fomento: Auxílio à Pesquisa - Temático
• Processo FAPESP: 14/16250-9 - Sobre a otimização de parâmetros em técnicas de aprendizado de máquina: avanços e paradigmas
• Beneficiário: João Paulo Papa
• Linha de fomento: Auxílio à Pesquisa - Regular
• Processo FAPESP: 15/25739-4 - Estudo de semântica em modelos de aprendizado em profundidade
• Beneficiário: Gustavo Henrique de Rosa
• Linha de fomento: Bolsas no Brasil - Mestrado