Optimizing Convolutional Neural Networks Architecture Using a Modified Particle Swarm Optimization for Image Classification
| Autor: | D. Elhani, A.C. Megherbi, A. Zitouni, F. Dornaika, S. Sbaa, A. Taleb-Ahmed |
|---|---|
| Přispěvatelé: | University of Biskra Mohamed Khider, Laboratoire national des champs magnétiques intenses - Grenoble (LNCMI-G ), Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA), Ho Chi Minh City Open University, COMmunications NUMériques - IEMN (COMNUM - IEMN), INSA Institut National des Sciences Appliquées Hauts-de-France (INSA Hauts-De-France), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Institut d’Électronique, de Microélectronique et de Nanotechnologie - UMR 8520 (IEMN), Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS)-Université Polytechnique Hauts-de-France (UPHF)-JUNIA (JUNIA), Université catholique de Lille (UCL)-Université catholique de Lille (UCL)-Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS)-Université Polytechnique Hauts-de-France (UPHF)-JUNIA (JUNIA), Université catholique de Lille (UCL)-Université catholique de Lille (UCL), Institut d’Électronique, de Microélectronique et de Nanotechnologie - UMR 8520 (IEMN), No information found |
| Rok vydání: | 2022 |
| Předmět: |
[PHYS]Physics [physics]
History Polymers and Plastics Image classification Particle swarm optimization General Engineering Industrial and Manufacturing Engineering Computer Science Applications [SPI]Engineering Sciences [physics] Artificial Intelligence Neural network design Convolutional neural networks Business and International Management Hyperparameters |
| Zdroj: | Expert Systems with Applications Expert Systems with Applications, 2023, 229, part A, 120411. ⟨10.1016/j.eswa.2023.120411⟩ |
| ISSN: | 1556-5068 0957-4174 |
| Popis: | International audience; Although Convolutional Neural Networks (CNNs) have been shown to be highly effective in image classificationtasks, designing their architecture to achieve optimal results is often challenging. This process is time consuming, requires significant effort and expertise, and is complicated by the large number of hyperparameters. Toaddress this problem, in this work we propose an approach that reduces human intervention and automaticallygenerates the best CNN design. Our approach uses a variant of Particle Swarm Optimization (PSO), calledParticle Swarm Optimization without Velocity (PSWV), to speed up convergence and reduce the numberof iterations required to determine the optimal CNN hyperparameters. We developed a novel strategy todetermine the updated position of each particle using a linear combination of the best position of the particleand the best position of the swarm without relying on the velocity equation. Our algorithm harnesses thepower of the variable-length encoding strategy to represent particles within the population, thereby providingswift convergence towards the best architecture. We evaluate our proposed algorithm against several recentalgorithms in the literature by using nine benchmark datasets for classification tasks and comparing it to 27other algorithms, including state-of-the-art ones. Our experimental results show that our proposed method,pswvCNN, is able to quickly find effective CNN architectures that provide comparable performance to the bestcurrently available designs, indicating its significant potential. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect