Quaternion-based deep belief networks fine-tuning

Made available in DSpace on 2018-11-26T17:42:02Z (GMT). No. of bitstreams: 0 Previous issue date: 2017-11-01 Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) 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. Sao Paulo State Univ, Dept Comp, Av Eng Luiz Edmundo Carrijo Coube 14-01, BR-17033360 Bauru, SP, Brazil Middlesex Univ, Sch Sci & Technol, London NW4 4BT, England Sao Paulo State Univ, Dept Comp, Av Eng Luiz Edmundo Carrijo Coube 14-01, BR-17033360 Bauru, SP, Brazil FAPESP: 2014/12236-1 FAPESP: 2014/16250-9 FAPESP: 2015/25739-4 CNPq: 470571/2013-6 CNPq: 306166/2014-3