Fast and Accurate Multi-Neural Network Ensemble Model
Gespeichert in:
| Titel: | Fast and Accurate Multi-Neural Network Ensemble Model |
|---|---|
| Autoren: | Nakci, Veli, Altun, M. |
| Quelle: | 2025 21st International Conference on Synthesis, Modeling, Analysis and Simulation Methods, and Applications to Circuits Design (SMACD). :1-4 |
| Verlagsinformationen: | IEEE, 2025. |
| Publikationsjahr: | 2025 |
| Schlagwörter: | Deep Neural Networks, Image Classification, Neural Networks, Transfer Learning Methods, Ensemble Models, High Accuracy, Transfer Learning, Learning Systems, Image Enhancement, Multi-Neural Networks, Ensemble Techniques, Neural Networks Ensemble, Training Time, Ensemble Technique |
| Beschreibung: | In image classification, having a high accuracy is a significant metric for a model. Therefore, some certain methods such as ensemble technique etc. are commonly used for this objective. However, while trying to achieve high accuracy, other important metrics such as training time must also be considered. Transfer learning method is widely applied in image classification to reduce training time and enhance model efficiency. Even though transfer learning with models such as AlexNet, VGG16, and DenseNet121 is applied on some image datasets, it requires a great amount of training time to achieve high accuracy. In this study, we propose a model that utilizes weighted voting ensemble technique with an auxiliary network. We evaluate our model and pre-trained models - Alexnet, VGG1, and DenseNet121 - on CIFAR-10 dataset. The results show that the proposed model outperforms pre-trained models in terms of achieving high accuracies and requiring less training time. To achieve 80% accuracy, our model requires 15,38%, 10%, and 87.78% of the training time used by Alexnet, VGG16 and DenseNet121, respectively. While the proposed model achieves 85% and 90% accuracy, AlexNet and VGG16 cannot. In addition, it achieves 90% accuracy in 38.23 min, whereas DenseNet121 - more efficient than the other two pre-trained models - only reaches 87% accuracy in over three hours. © 2025 Elsevier B.V., All rights reserved. Ankasys; Atek Midas; CDT; Cirrus Logic; et al.; EuroPractice |
| Publikationsart: | Article Conference object |
| DOI: | 10.1109/smacd65553.2025.11091980 |
| Zugangs-URL: | https://hdl.handle.net/20.500.11779/3086 https://doi.org/10.1109/SMACD65553.2025.11091980 |
| Rights: | STM Policy #29 |
| Dokumentencode: | edsair.doi.dedup.....1af709dfc48aa500fb5fa1766e40bc83 |
| Datenbank: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstract: | In image classification, having a high accuracy is a significant metric for a model. Therefore, some certain methods such as ensemble technique etc. are commonly used for this objective. However, while trying to achieve high accuracy, other important metrics such as training time must also be considered. Transfer learning method is widely applied in image classification to reduce training time and enhance model efficiency. Even though transfer learning with models such as AlexNet, VGG16, and DenseNet121 is applied on some image datasets, it requires a great amount of training time to achieve high accuracy. In this study, we propose a model that utilizes weighted voting ensemble technique with an auxiliary network. We evaluate our model and pre-trained models - Alexnet, VGG1, and DenseNet121 - on CIFAR-10 dataset. The results show that the proposed model outperforms pre-trained models in terms of achieving high accuracies and requiring less training time. To achieve 80% accuracy, our model requires 15,38%, 10%, and 87.78% of the training time used by Alexnet, VGG16 and DenseNet121, respectively. While the proposed model achieves 85% and 90% accuracy, AlexNet and VGG16 cannot. In addition, it achieves 90% accuracy in 38.23 min, whereas DenseNet121 - more efficient than the other two pre-trained models - only reaches 87% accuracy in over three hours. © 2025 Elsevier B.V., All rights reserved.<br />Ankasys; Atek Midas; CDT; Cirrus Logic; et al.; EuroPractice |
|---|---|
| DOI: | 10.1109/smacd65553.2025.11091980 |