ASMIL: Attention-Stabilized Multiple Instance Learning for Whole Slide Imaging
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915840189792256 |
|---|---|
| author | Ye, Linfeng Hamidi, Shayan Mohajer Chi, Zhixiang Li, Guang Pilanci, Mert Ogawa, Takahiro Haseyama, Miki Plataniotis, Konstantinos N. |
| author_facet | Ye, Linfeng Hamidi, Shayan Mohajer Chi, Zhixiang Li, Guang Pilanci, Mert Ogawa, Takahiro Haseyama, Miki Plataniotis, Konstantinos N. |
| contents | Attention-based multiple instance learning (MIL) has emerged as a powerful framework for whole slide image (WSI) diagnosis, leveraging attention to aggregate instance-level features into bag-level predictions. Despite this success, we find that such methods exhibit a new failure mode: unstable attention dynamics. Across four representative attention-based MIL methods and two public WSI datasets, we observe that attention distributions oscillate across epochs rather than converging to a consistent pattern, degrading performance. This instability adds to two previously reported challenges: overfitting and over-concentrated attention distribution. To simultaneously overcome these three limitations, we introduce attention-stabilized multiple instance learning (ASMIL), a novel unified framework. ASMIL uses an anchor model to stabilize attention, replaces softmax with a normalized sigmoid function in the anchor to prevent over-concentration, and applies token random dropping to mitigate overfitting. Extensive experiments demonstrate that ASMIL achieves up to a 6.49\% F1 score improvement over state-of-the-art methods. Moreover, integrating the anchor model and normalized sigmoid into existing attention-based MIL methods consistently boosts their performance, with F1 score gains up to 10.73\%. All code and data are publicly available at https://github.com/Linfeng-Ye/ASMIL. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_06658 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ASMIL: Attention-Stabilized Multiple Instance Learning for Whole Slide Imaging Ye, Linfeng Hamidi, Shayan Mohajer Chi, Zhixiang Li, Guang Pilanci, Mert Ogawa, Takahiro Haseyama, Miki Plataniotis, Konstantinos N. Computer Vision and Pattern Recognition Attention-based multiple instance learning (MIL) has emerged as a powerful framework for whole slide image (WSI) diagnosis, leveraging attention to aggregate instance-level features into bag-level predictions. Despite this success, we find that such methods exhibit a new failure mode: unstable attention dynamics. Across four representative attention-based MIL methods and two public WSI datasets, we observe that attention distributions oscillate across epochs rather than converging to a consistent pattern, degrading performance. This instability adds to two previously reported challenges: overfitting and over-concentrated attention distribution. To simultaneously overcome these three limitations, we introduce attention-stabilized multiple instance learning (ASMIL), a novel unified framework. ASMIL uses an anchor model to stabilize attention, replaces softmax with a normalized sigmoid function in the anchor to prevent over-concentration, and applies token random dropping to mitigate overfitting. Extensive experiments demonstrate that ASMIL achieves up to a 6.49\% F1 score improvement over state-of-the-art methods. Moreover, integrating the anchor model and normalized sigmoid into existing attention-based MIL methods consistently boosts their performance, with F1 score gains up to 10.73\%. All code and data are publicly available at https://github.com/Linfeng-Ye/ASMIL. |
| title | ASMIL: Attention-Stabilized Multiple Instance Learning for Whole Slide Imaging |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.06658 |