Details

Autor(en) / Beteiligte

• Martín, Alejandro
• Lara-Cabrera, Raúl
• Fuentes-Hurtado, Félix
• Naranjo, Valery
• Camacho, David

Titel

EvoDeep: A new evolutionary approach for automatic Deep Neural Networks parametrisation

Ist Teil von

• Journal of parallel and distributed computing, 2018-07, Vol.117, p.180-191

Ort / Verlag

Elsevier Inc

Erscheinungsjahr

2018

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Deep Neural Networks (DNN) have become a powerful, and extremely popular mechanism, which has been widely used to solve problems of varied complexity, due to their ability to make models fitted to non-linear complex problems. Despite its well-known benefits, DNNs are complex learning models whose parametrisationand architecture are made usually by hand. This paper proposes a new Evolutionary Algorithm, named EvoDeep, devoted to evolve the parameters and the architecture of a DNN in order to maximise its classification accuracy, as well as maintaining a valid sequence of layers. This model is tested against a widely used dataset of handwritten digits images. The experiments performed using this dataset show that the Evolutionary Algorithm is able to select the parameters and the DNN architecture appropriately, achieving a 98.93% accuracy in the best run. •Definition of new evolutionary algorithm, named EvoDeep.•EvoDeep evolves parameters and architecture of Deep Neural Networks.•Classification accuracy is maximised ensuring a valid sequence of layers.•A detailed Evolutionary-based model design which uses a Finite State Machine.•Experimental evaluation of the EvoDeep algorithm in a well-known image dataset.