Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.10041 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918158683602944 |
|---|---|
| author | Han, Sahng-Min Kim, Minjae Cha, Jinho Choe, Se-woon Cha, Eunchan Daniel Choi, Jungwon Jung, Kyudong |
| author_facet | Han, Sahng-Min Kim, Minjae Cha, Jinho Choe, Se-woon Cha, Eunchan Daniel Choi, Jungwon Jung, Kyudong |
| contents | Deep learning in small and imbalanced biomedical datasets remains fundamentally constrained by unstable optimization and poor generalization. We present the first biomedical implementation of FOSSIL (Flexible Optimization via Sample-Sensitive Importance Learning), a regret-minimizing weighting framework that adaptively balances training emphasis according to sample difficulty. Using softmax-based uncertainty as a continuous measure of difficulty, we construct a four-stage curriculum (Easy-Very Hard) and integrate FOSSIL into both convolutional and transformer-based architectures for Mpox skin lesion diagnosis. Across all settings, FOSSIL substantially improves discrimination (AUC = 0.9573), calibration (ECE = 0.053), and robustness under real-world perturbations, outperforming conventional baselines without metadata, manual curation, or synthetic augmentation. The results position FOSSIL as a generalizable, data-efficient, and interpretable framework for difficulty-aware learning in medical imaging under data scarcity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_10041 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FOSSIL: Regret-Minimizing Curriculum Learning for Metadata-Free and Low-Data Mpox Diagnosis Han, Sahng-Min Kim, Minjae Cha, Jinho Choe, Se-woon Cha, Eunchan Daniel Choi, Jungwon Jung, Kyudong Machine Learning Artificial Intelligence Deep learning in small and imbalanced biomedical datasets remains fundamentally constrained by unstable optimization and poor generalization. We present the first biomedical implementation of FOSSIL (Flexible Optimization via Sample-Sensitive Importance Learning), a regret-minimizing weighting framework that adaptively balances training emphasis according to sample difficulty. Using softmax-based uncertainty as a continuous measure of difficulty, we construct a four-stage curriculum (Easy-Very Hard) and integrate FOSSIL into both convolutional and transformer-based architectures for Mpox skin lesion diagnosis. Across all settings, FOSSIL substantially improves discrimination (AUC = 0.9573), calibration (ECE = 0.053), and robustness under real-world perturbations, outperforming conventional baselines without metadata, manual curation, or synthetic augmentation. The results position FOSSIL as a generalizable, data-efficient, and interpretable framework for difficulty-aware learning in medical imaging under data scarcity. |
| title | FOSSIL: Regret-Minimizing Curriculum Learning for Metadata-Free and Low-Data Mpox Diagnosis |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2510.10041 |