Statistical Test for Attention Map in Vision Transformer
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911760719544320 |
|---|---|
| author | Shiraishi, Tomohiro Miwa, Daiki Katsuoka, Teruyuki Duy, Vo Nguyen Le Taji, Kouichi Takeuchi, Ichiro |
| author_facet | Shiraishi, Tomohiro Miwa, Daiki Katsuoka, Teruyuki Duy, Vo Nguyen Le Taji, Kouichi Takeuchi, Ichiro |
| contents | The Vision Transformer (ViT) demonstrates exceptional performance in various computer vision tasks. Attention is crucial for ViT to capture complex wide-ranging relationships among image patches, allowing the model to weigh the importance of image patches and aiding our understanding of the decision-making process. However, when utilizing the attention of ViT as evidence in high-stakes decision-making tasks such as medical diagnostics, a challenge arises due to the potential of attention mechanisms erroneously focusing on irrelevant regions. In this study, we propose a statistical test for ViT's attentions, enabling us to use the attentions as reliable quantitative evidence indicators for ViT's decision-making with a rigorously controlled error rate. Using the framework called selective inference, we quantify the statistical significance of attentions in the form of p-values, which enables the theoretically grounded quantification of the false positive detection probability of attentions. We demonstrate the validity and the effectiveness of the proposed method through numerical experiments and applications to brain image diagnoses. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_08169 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Statistical Test for Attention Map in Vision Transformer Shiraishi, Tomohiro Miwa, Daiki Katsuoka, Teruyuki Duy, Vo Nguyen Le Taji, Kouichi Takeuchi, Ichiro Machine Learning The Vision Transformer (ViT) demonstrates exceptional performance in various computer vision tasks. Attention is crucial for ViT to capture complex wide-ranging relationships among image patches, allowing the model to weigh the importance of image patches and aiding our understanding of the decision-making process. However, when utilizing the attention of ViT as evidence in high-stakes decision-making tasks such as medical diagnostics, a challenge arises due to the potential of attention mechanisms erroneously focusing on irrelevant regions. In this study, we propose a statistical test for ViT's attentions, enabling us to use the attentions as reliable quantitative evidence indicators for ViT's decision-making with a rigorously controlled error rate. Using the framework called selective inference, we quantify the statistical significance of attentions in the form of p-values, which enables the theoretically grounded quantification of the false positive detection probability of attentions. We demonstrate the validity and the effectiveness of the proposed method through numerical experiments and applications to brain image diagnoses. |
| title | Statistical Test for Attention Map in Vision Transformer |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2401.08169 |