Hyperparameter optimization of pre-trained convolutional neural networks using adolescent identity search algorithm


Akkus E., Bal U., Kocoglu F. Ö., Beyhan S.

NEURAL COMPUTING & APPLICATIONS, vol.36, no.4, pp.1523-1537, 2024 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 36 Issue: 4
  • Publication Date: 2024
  • Doi Number: 10.1007/s00521-023-09121-8
  • Journal Name: NEURAL COMPUTING & APPLICATIONS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Applied Science & Technology Source, Biotechnology Research Abstracts, Compendex, Computer & Applied Sciences, Index Islamicus, INSPEC, zbMATH
  • Page Numbers: pp.1523-1537
  • Istanbul University Affiliated: No

Abstract

Convolutional neural networks (CNNs) are widely used deep learning (DL) models for image classification. The selected hyperparameters for training convolutional neural network (CNN) models have a significant effect on the performance. Therefore, hyperparameter optimization (HPO) is an important process to design optimal CNN models. In this study, Adolescent Identity Search Algorithm (AISA) and Bayesian Optimization (BO) methods were applied for HPO of pre-trained CNN models to improve their classification performance. Diabetic retinopathy (DR) classification was chosen as the application problem of the study and Kaggle Diabetic Retinopathy Detection dataset was used. We used pre-trained CNN models named AlexNet, MobileNetV2, ResNet18, and GoogLeNet. To the best of our knowledge, this study represents the first use of AISA-based HPO for DR classification. The results show that hybrid models incorporating AISA-based HPO achieve better accuracy with fewer iterations than BO-based HPO hybridized models.