DivNet : Efficient Convolutional Neural Network via Multilevel Hierarchical Architecture Design
Kaddar, Bachir (University Ibn Khaldoun. Department of Natural and Life Sciences)
Fizazi, Hadria (Université des Sciences et de la Technologie dOran Mohamed Boudiaf. Department of Computer Sciences)
Hernández-Cabronero, Miguel (Universitat Autònoma de Barcelona. Departament d'Enginyeria de la Informació i de les Comunicacions)
Sánchez Silva, Víctor Francisco (University of Warwick. Department of Computer Science)
Serra-Sagristà, Joan (Universitat Autònoma de Barcelona. Departament d'Enginyeria de la Informació i de les Comunicacions)

Fecha:	2021
Resumen:	Designing small and efficient mobile neural networks is difficult because the challenge is to determine the architecture that achieves the best performance under a given limited computational scenario. Previous lightweight neural networks rely on a cell module that is repeated in all stacked layers across the network. These approaches do not permit layer diversity, which is critical for achieving strong performance. This paper presents an experimental study to develop an efficient mobile network using a hierarchical architecture. Our proposed mobile network, called Diversity Network (DivNet), has been shown to perform better than the basic architecture generally employed by the best high-efficiency models-with simply stacked layers-, regarding complexity cost and performance. A set of architectural design decisions are described that reduce the proposed model size while yielding a significant performance improvement. Our experiments on image classification show that compared to, respectively, MobileNetV2, SqueezeNet, and ShuffleNetV2, our proposal DivNet can improve accuracy by 2. 09%, 0. 76%, and 0. 66% on the CIFAR100 dataset, and by 0. 05%, 4. 96%, and 1. 13% on the CIFAR10 dataset. On more complex datasets, e. g. , ImageNet, our proposal DivNet achieves 70. 65% Top-1 accuracy and 90. 23% Top-5 accuracy, still better than other small models like MobilNet, SqueezeNet, ShuffleNet.
Ayudas:	Generalitat de Catalunya 2017/SGR-463 Generalitat de Catalunya 2018/BP-00008 European Commission 801370 Agencia Estatal de Investigación RTI2018-095287-B-I00
Derechos:	Aquest document està subjecte a una llicència d'ús Creative Commons. Es permet la reproducció total o parcial, la distribució, la comunicació pública de l'obra i la creació d'obres derivades, fins i tot amb finalitats comercials, sempre i quan es reconegui l'autoria de l'obra original.
Lengua:	Anglès
Documento:	Article ; recerca ; Versió publicada
Materia:	Deep neural network ; Mobile network ; Network compression
Publicado en:	IEEE Access, Vol. 9 (July 2021) , p. 105892-105901, ISSN 2169-3536

DOI: 10.1109/ACCESS.2021.3099952

10 p, 1.1 MB

El registro aparece en las colecciones:
Documentos de investigación > Documentos de los grupos de investigación de la UAB > Centros y grupos de investigación (producción científica) > Ingeniería > Group on Interactive Coding of Images (GICI)
Artículos > Artículos de investigación
Artículos > Artículos publicados