MADL: A Multilevel Architecture of Deep Learning

Authors

Samir Brahim Belhaouari^{1, *,} , Hafsa Raissouli²

¹College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar

²College of Computer Science and Information Technology, University Putra Malaysia, Serdang, Malaysia

^*Corresponding author. Email: sbelhaouari@hbku.edu.qa

DOI

10.2991/ijcis.d.201216.003How to use a DOI?

Keywords

Convolutional neural network; Multilevel architecture of deep learning; Advanced activation function; CIFAR-10; MADL

Abstract

Deep neural networks (DNN) are a powerful tool that is used in many real-life applications. Solving complicated real-life problems requires deeper and larger networks, and hence, a larger number of parameters to optimize. This paper proposes a multilevel architecture of deep learning (MADL) that breaks down the optimization to different levels and steps where networks are trained and optimized separately. Two approaches of passing the features from level to level are discussed. The first approach uses the output layer of level as input to level and the second approach discusses introducing an additional fully connected layer to pass the features from it directly to the next level. The experimentations showed that the second approach, that is the use of the features in the additional fully connected layer, gives a higher improvement. The paper also discusses an advanced customizable activation function that is comparable in its performance to rectified linear unit (ReLU). MADL is experimented using CIFAR-10 and exhibited an improvement of 0.84% compared to a single network resulting in an accuracy of 98.04%.

Copyright

© 2021 The Authors. Published by Atlantis Press B.V.

Open Access

This is an open access article distributed under the CC BY-NC 4.0 license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)
View full text (HTML)

TY  - JOUR
AU  - Samir Brahim Belhaouari
AU  - Hafsa Raissouli
PY  - 2021
DA  - 2021/02/08
TI  - MADL: A Multilevel Architecture of Deep Learning
JO  - International Journal of Computational Intelligence Systems
SP  - 693
EP  - 700
VL  - 14
IS  - 1
SN  - 1875-6883
UR  - https://doi.org/10.2991/ijcis.d.201216.003
DO  - 10.2991/ijcis.d.201216.003
ID  - Belhaouari2021
ER  -