![[2602.18502] Mitigating Shortcut Learning via Feature Disentanglement in Medical Imaging: A Benchmark Study](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.18502] Mitigating Shortcut Learning via Feature Disentanglement in Medical Imaging: A Benchmark Study
Summary
This study evaluates feature disentanglement methods to mitigate shortcut learning in medical imaging, enhancing model robustness and classification performance.
Why It Matters
As deep learning models in medical imaging often rely on spurious correlations, this research is crucial for ensuring models generalize effectively across diverse clinical settings. By addressing shortcut learning, the study contributes to safer and more reliable AI applications in healthcare.
Key Takeaways
- Feature disentanglement can reduce reliance on shortcut learning in medical imaging.
- Combining data-centric and model-centric approaches yields better performance.
- Latent space analyses reveal insights into representation quality beyond classification metrics.
- Robustness of models varies with the degree of confounding in training data.
- Computational efficiency remains comparable across different mitigation methods.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.18502 (cs) [Submitted on 17 Feb 2026] Title:Mitigating Shortcut Learning via Feature Disentanglement in Medical Imaging: A Benchmark Study Authors:Sarah Müller, Philipp Berens View a PDF of the paper titled Mitigating Shortcut Learning via Feature Disentanglement in Medical Imaging: A Benchmark Study, by Sarah M\"uller and 1 other authors View PDF HTML (experimental) Abstract:Although deep learning models in medical imaging often achieve excellent classification performance, they can rely on shortcut learning, exploiting spurious correlations or confounding factors that are not causally related to the target task. This poses risks in clinical settings, where models must generalize across institutions, populations, and acquisition conditions. Feature disentanglement is a promising approach to mitigate shortcut learning by separating task-relevant information from confounder-related features in latent representations. In this study, we systematically evaluated feature disentanglement methods for mitigating shortcuts in medical imaging, including adversarial learning and latent space splitting based on dependence minimization. We assessed classification performance and disentanglement quality using latent space analyses across one artificial and two medical datasets with natural and synthetic confounders. We also examined robustness under varying levels of confounding and compared computational efficiency a...
