A metaheuristic-driven approach to fine-tune Deep Boltzmann Machines

Passos, Leandro Aparecido; Papa, Joao Paulo [UNESP]

A metaheuristic-driven approach to fine-tune Deep Boltzmann Machines

Data

2020-12-01

Autores

Passos, Leandro Aparecido

Papa, Joao Paulo

Editor

Elsevier B.V.

Tipo

Artigo

Resumo

Deep learning techniques, such as Deep Boltzmann Machines (DBMs), have received considerable attention over the past years due to the outstanding results concerning a variable range of domains. One of the main shortcomings of these techniques involves the choice of their hyperparameters, since they have a significant impact on the final results. This work addresses the issue of fine-tuning hyperparameters of Deep Boltzmann Machines using metaheuristic optimization techniques with different backgrounds, such as swarm intelligence, memoryand evolutionary-based approaches. Experiments conducted in three public datasets for binary image reconstruction showed that metaheuristic techniques can obtain reasonable results. (C) 2019 Elsevier B.V. All rights reserved.