CROCODILE: Causality aids RObustness via COntrastive DIsentangled LEarning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909282690138112 |
|---|---|
| author | Carloni, Gianluca Tsaftaris, Sotirios A Colantonio, Sara |
| author_facet | Carloni, Gianluca Tsaftaris, Sotirios A Colantonio, Sara |
| contents | Due to domain shift, deep learning image classifiers perform poorly when applied to a domain different from the training one. For instance, a classifier trained on chest X-ray (CXR) images from one hospital may not generalize to images from another hospital due to variations in scanner settings or patient characteristics. In this paper, we introduce our CROCODILE framework, showing how tools from causality can foster a model's robustness to domain shift via feature disentanglement, contrastive learning losses, and the injection of prior knowledge. This way, the model relies less on spurious correlations, learns the mechanism bringing from images to prediction better, and outperforms baselines on out-of-distribution (OOD) data. We apply our method to multi-label lung disease classification from CXRs, utilizing over 750000 images from four datasets. Our bias-mitigation method improves domain generalization and fairness, broadening the applicability and reliability of deep learning models for a safer medical image analysis. Find our code at: https://github.com/gianlucarloni/crocodile. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_04949 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | CROCODILE: Causality aids RObustness via COntrastive DIsentangled LEarning Carloni, Gianluca Tsaftaris, Sotirios A Colantonio, Sara Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition Machine Learning I.2; I.4; I.5; J.3; J.6 Due to domain shift, deep learning image classifiers perform poorly when applied to a domain different from the training one. For instance, a classifier trained on chest X-ray (CXR) images from one hospital may not generalize to images from another hospital due to variations in scanner settings or patient characteristics. In this paper, we introduce our CROCODILE framework, showing how tools from causality can foster a model's robustness to domain shift via feature disentanglement, contrastive learning losses, and the injection of prior knowledge. This way, the model relies less on spurious correlations, learns the mechanism bringing from images to prediction better, and outperforms baselines on out-of-distribution (OOD) data. We apply our method to multi-label lung disease classification from CXRs, utilizing over 750000 images from four datasets. Our bias-mitigation method improves domain generalization and fairness, broadening the applicability and reliability of deep learning models for a safer medical image analysis. Find our code at: https://github.com/gianlucarloni/crocodile. |
| title | CROCODILE: Causality aids RObustness via COntrastive DIsentangled LEarning |
| topic | Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition Machine Learning I.2; I.4; I.5; J.3; J.6 |
| url | https://arxiv.org/abs/2408.04949 |