A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention
The performance of Machine Learning (ML) models is highly sensitive to data quality, still the impact of label accuracy remains underexplored. In this study, a novel architecture, the Dual Selective Attention Network (DSAN), is proposed to improve robustness against mislabeled data in deep learning...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11062817/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849428612054253568 |
|---|---|
| author | Hasnain Hyder Gulsher Baloch Amreen Batool Yong-Woon Kim Yung-Cheol Byun |
| author_facet | Hasnain Hyder Gulsher Baloch Amreen Batool Yong-Woon Kim Yung-Cheol Byun |
| author_sort | Hasnain Hyder |
| collection | DOAJ |
| description | The performance of Machine Learning (ML) models is highly sensitive to data quality, still the impact of label accuracy remains underexplored. In this study, a novel architecture, the Dual Selective Attention Network (DSAN), is proposed to improve robustness against mislabeled data in deep learning tasks. DSAN incorporates Position Attention Module (PAM) and Channel Attention Module (CAM) to emphasize relevant spatial and channel level features, effectively suppressing the influence of incorrect labels. DSAN was evaluated alongside four baseline models Visual Geometry Group Network (VGG16), Convolutional Neural Network (CNN), Artificial Neural Network (ANN), and ResNet-50 on three datasets, with label noise introduced at 0%, 5%, 10%, 15%, and 20% to simulate real-world mislabeling scenarios. Model performance was assessed using accuracy, precision, recall (sensitivity), and F1-score. DSAN maintained over 96.5% accuracy under 20% label noise and outperformed all baseline models by 4–8% on average across performance metrics. While CNN showed moderate robustness and ResNet-50 exhibited better resilience due to its residual learning mechanism, both models experienced notable performance degradation as noise increased. VGG16 and ANN were particularly vulnerable, with sharp declines observed even under low noise levels. To further address mislabeling, the Latent Space Variation using Supervised Autoencoder (AQUAVS) technique was also applied to remove mislabeled data and assess model recovery. Although slight improvements were observed, AQUAVS still lagged behind DSAN in all scenarios. Additionally, an ablation study was conducted to evaluate the individual contributions of PAM and CAM, showing that their combination significantly enhances DSAN robustness to mislabeled data while maintaining high classification performance under label noise. These results highlight the importance of designing architectures that are inherently robust to label noise. By maintaining high reliability under noisy conditions, DSAN contributes to the development of more dependable and generalizable AI systems. |
| format | Article |
| id | doaj-art-b7b7f86a46b641898f2d42a960d1a4c5 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-b7b7f86a46b641898f2d42a960d1a4c52025-08-20T03:28:38ZengIEEEIEEE Access2169-35362025-01-011311560411562610.1109/ACCESS.2025.358528811062817A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective AttentionHasnain Hyder0https://orcid.org/0009-0004-1005-4229Gulsher Baloch1https://orcid.org/0000-0002-7346-6077Amreen Batool2https://orcid.org/0000-0002-6041-653XYong-Woon Kim3https://orcid.org/0000-0002-4759-0138Yung-Cheol Byun4https://orcid.org/0000-0003-1107-9941Department of Electronic Engineering, Institute of Information Science and Technology, Jeju National University, Jeju-si, South KoreaDepartment of Electrical Engineering, Sukkur IBA University, Sukkur, PakistanDepartment of Electronic Engineering, Institute of Information Science and Technology, Jeju National University, Jeju-si, South KoreaDepartment of Computer Engineering, Jeju National University, Jeju-si, South KoreaDepartment of Computer Engineering, Institute of Information Science and Technology, Jeju National University, Jeju-si, South KoreaThe performance of Machine Learning (ML) models is highly sensitive to data quality, still the impact of label accuracy remains underexplored. In this study, a novel architecture, the Dual Selective Attention Network (DSAN), is proposed to improve robustness against mislabeled data in deep learning tasks. DSAN incorporates Position Attention Module (PAM) and Channel Attention Module (CAM) to emphasize relevant spatial and channel level features, effectively suppressing the influence of incorrect labels. DSAN was evaluated alongside four baseline models Visual Geometry Group Network (VGG16), Convolutional Neural Network (CNN), Artificial Neural Network (ANN), and ResNet-50 on three datasets, with label noise introduced at 0%, 5%, 10%, 15%, and 20% to simulate real-world mislabeling scenarios. Model performance was assessed using accuracy, precision, recall (sensitivity), and F1-score. DSAN maintained over 96.5% accuracy under 20% label noise and outperformed all baseline models by 4–8% on average across performance metrics. While CNN showed moderate robustness and ResNet-50 exhibited better resilience due to its residual learning mechanism, both models experienced notable performance degradation as noise increased. VGG16 and ANN were particularly vulnerable, with sharp declines observed even under low noise levels. To further address mislabeling, the Latent Space Variation using Supervised Autoencoder (AQUAVS) technique was also applied to remove mislabeled data and assess model recovery. Although slight improvements were observed, AQUAVS still lagged behind DSAN in all scenarios. Additionally, an ablation study was conducted to evaluate the individual contributions of PAM and CAM, showing that their combination significantly enhances DSAN robustness to mislabeled data while maintaining high classification performance under label noise. These results highlight the importance of designing architectures that are inherently robust to label noise. By maintaining high reliability under noisy conditions, DSAN contributes to the development of more dependable and generalizable AI systems.https://ieeexplore.ieee.org/document/11062817/Mislabelled datadata-centric approachvisual geometry group (VGG) 16dual selective attention network (DSAN)data qualityResNet-50 |
| spellingShingle | Hasnain Hyder Gulsher Baloch Amreen Batool Yong-Woon Kim Yung-Cheol Byun A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention IEEE Access Mislabelled data data-centric approach visual geometry group (VGG) 16 dual selective attention network (DSAN) data quality ResNet-50 |
| title | A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention |
| title_full | A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention |
| title_fullStr | A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention |
| title_full_unstemmed | A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention |
| title_short | A Robust Deep Learning Framework for Mitigating Label Noise With Dual Selective Attention |
| title_sort | robust deep learning framework for mitigating label noise with dual selective attention |
| topic | Mislabelled data data-centric approach visual geometry group (VGG) 16 dual selective attention network (DSAN) data quality ResNet-50 |
| url | https://ieeexplore.ieee.org/document/11062817/ |
| work_keys_str_mv | AT hasnainhyder arobustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT gulsherbaloch arobustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT amreenbatool arobustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT yongwoonkim arobustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT yungcheolbyun arobustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT hasnainhyder robustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT gulsherbaloch robustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT amreenbatool robustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT yongwoonkim robustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention AT yungcheolbyun robustdeeplearningframeworkformitigatinglabelnoisewithdualselectiveattention |