Automated CNN Architectural Design: A Simple and Efficient Methodology for Computer Vision Tasks

Author:

Al Bataineh Ali1ORCID,Kaur Devinder2,Al-khassaweneh Mahmood34,Al-sharoa Esraa5ORCID

Affiliation:

1. Department of Electrical and Computer Engineering, Norwich University, Northfield, VT 05663, USA

2. Department of Electrical Engineering and Computer Science, University of Toledo, Toledo, OH 43606, USA

3. Engineering, Computing and Mathematical Sciences, Lewis University, Romeoville, IL 60446, USA

4. Computer Engineering Department, Yarmouk University, Irbid 21163, Jordan

5. Electrical Engineering Department, Jordan University of Science and Technology, Irbid 22110, Jordan

Abstract

Convolutional neural networks (CNN) have transformed the field of computer vision by enabling the automatic extraction of features, obviating the need for manual feature engineering. Despite their success, identifying an optimal architecture for a particular task can be a time-consuming and challenging process due to the vast space of possible network designs. To address this, we propose a novel neural architecture search (NAS) framework that utilizes the clonal selection algorithm (CSA) to automatically design high-quality CNN architectures for image classification problems. Our approach uses an integer vector representation to encode CNN architectures and hyperparameters, combined with a truncated Gaussian mutation scheme that enables efficient exploration of the search space. We evaluated the proposed method on six challenging EMNIST benchmark datasets for handwritten digit recognition, and our results demonstrate that it outperforms nearly all existing approaches. In addition, our approach produces state-of-the-art performance while having fewer trainable parameters than other methods, making it low-cost, simple, and reusable for application to multiple datasets.

Publisher

MDPI AG

Subject

General Mathematics,Engineering (miscellaneous),Computer Science (miscellaneous)

Reference69 articles.

1. Deep learning;LeCun;Nature,2015

2. A comparative analysis of nonlinear machine learning algorithms for breast cancer detection;Int. J. Mach. Learn. Comput.,2019

3. Goodfelow, I., Bengio, Y., and Courville, A. (2016). Deep Learning (Adaptive Computation and Machine Learning Series), MIT Press.

4. Al Bataineh, A., Mairaj, A., and Kaur, D. (2020). Autoencoder based semi-supervised anomaly detection in turbofan engines. Int. J. Adv. Comput. Sci. Appl. (IJACSA), 11.

5. High Performance Implementation of Neural Networks Learning Using Swarm Optimization Algorithms for EEG Classification Based on Brain Wave Data;Jarrah;Int. J. Appl. Metaheuristic Comput. (IJAMC),2022

Cited by 3 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3