Building Ensemble of Deep Networks: Convolutional Networks and Transformers
This paper presents a study on an automated system for image classification, which is based on the fusion of various deep learning methods. The study explores how to create an ensemble of different Convolutional Neural Network (CNN) models and transformer topologies that are fine-tuned on several da...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10309107/ |
| _version_ | 1827763119261220864 |
|---|---|
| author | Loris Nanni Andrea Loreggia Leonardo Barcellona Stefano Ghidoni |
| author_facet | Loris Nanni Andrea Loreggia Leonardo Barcellona Stefano Ghidoni |
| author_sort | Loris Nanni |
| collection | DOAJ |
| description | This paper presents a study on an automated system for image classification, which is based on the fusion of various deep learning methods. The study explores how to create an ensemble of different Convolutional Neural Network (CNN) models and transformer topologies that are fine-tuned on several datasets to leverage their diversity. The research question addressed in this work is whether different optimization algorithms can help in developing robust and efficient machine learning systems to be used in different domains for classification purposes. To do that, we introduce novel Adam variants. We employed these new approaches, coupled with several CNN topologies, for building an ensemble of classifiers that outperforms both other Adam-based methods and stochastic gradient descent. Additionally, the study combines the ensemble of CNNs with an ensemble of transformers based on different topologies, such as Deit, Vit, Swin, and Coat. To the best of our knowledge, this is the first work in which an in-depth study of a set of transformers and convolutional neural networks in a large set of small/medium-sized images is carried out. The experiments performed on several datasets demonstrate that the combination of such different models results in a substantial performance improvement in all tested problems. All resources are available at <uri>https://github.com/LorisNanni</uri>. |
| first_indexed | 2024-03-11T10:47:50Z |
| format | Article |
| id | doaj.art-377458e5525d4efa98a8b00001efe8c3 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-11T10:47:50Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-377458e5525d4efa98a8b00001efe8c32023-11-14T00:00:53ZengIEEEIEEE Access2169-35362023-01-011112496212497410.1109/ACCESS.2023.333044210309107Building Ensemble of Deep Networks: Convolutional Networks and TransformersLoris Nanni0https://orcid.org/0000-0002-3502-7209Andrea Loreggia1https://orcid.org/0000-0002-9846-0157Leonardo Barcellona2https://orcid.org/0000-0003-4281-0610Stefano Ghidoni3https://orcid.org/0000-0003-3406-8719Department of Information Engineering (DEI), University of Padova, Padua, ItalyDepartment of Information Engineering (DII), University of Brescia, Brescia, ItalyDepartment of Information Engineering (DEI), University of Padova, Padua, ItalyDepartment of Information Engineering (DEI), University of Padova, Padua, ItalyThis paper presents a study on an automated system for image classification, which is based on the fusion of various deep learning methods. The study explores how to create an ensemble of different Convolutional Neural Network (CNN) models and transformer topologies that are fine-tuned on several datasets to leverage their diversity. The research question addressed in this work is whether different optimization algorithms can help in developing robust and efficient machine learning systems to be used in different domains for classification purposes. To do that, we introduce novel Adam variants. We employed these new approaches, coupled with several CNN topologies, for building an ensemble of classifiers that outperforms both other Adam-based methods and stochastic gradient descent. Additionally, the study combines the ensemble of CNNs with an ensemble of transformers based on different topologies, such as Deit, Vit, Swin, and Coat. To the best of our knowledge, this is the first work in which an in-depth study of a set of transformers and convolutional neural networks in a large set of small/medium-sized images is carried out. The experiments performed on several datasets demonstrate that the combination of such different models results in a substantial performance improvement in all tested problems. All resources are available at <uri>https://github.com/LorisNanni</uri>.https://ieeexplore.ieee.org/document/10309107/Convolutional neural networkstransformersoptimizationensemble |
| spellingShingle | Loris Nanni Andrea Loreggia Leonardo Barcellona Stefano Ghidoni Building Ensemble of Deep Networks: Convolutional Networks and Transformers IEEE Access Convolutional neural networks transformers optimization ensemble |
| title | Building Ensemble of Deep Networks: Convolutional Networks and Transformers |
| title_full | Building Ensemble of Deep Networks: Convolutional Networks and Transformers |
| title_fullStr | Building Ensemble of Deep Networks: Convolutional Networks and Transformers |
| title_full_unstemmed | Building Ensemble of Deep Networks: Convolutional Networks and Transformers |
| title_short | Building Ensemble of Deep Networks: Convolutional Networks and Transformers |
| title_sort | building ensemble of deep networks convolutional networks and transformers |
| topic | Convolutional neural networks transformers optimization ensemble |
| url | https://ieeexplore.ieee.org/document/10309107/ |
| work_keys_str_mv | AT lorisnanni buildingensembleofdeepnetworksconvolutionalnetworksandtransformers AT andrealoreggia buildingensembleofdeepnetworksconvolutionalnetworksandtransformers AT leonardobarcellona buildingensembleofdeepnetworksconvolutionalnetworksandtransformers AT stefanoghidoni buildingensembleofdeepnetworksconvolutionalnetworksandtransformers |