Evidential Concept Embedding Models: Towards Reliable Concept Explanations for Skin Disease Diagnosis
This addresses the need for reliable interpretable methods in high-stakes medical diagnosis, offering improvements for both supervised and label-efficient settings.
The paper tackled the problem of unreliable concept predictions in Concept Bottleneck Models for medical image analysis by proposing an evidential Concept Embedding Model that models concept uncertainty and rectifies misalignments, achieving superior performance in concept prediction and mitigating misalignments in label-efficient training.
Due to the high stakes in medical decision-making, there is a compelling demand for interpretable deep learning methods in medical image analysis. Concept Bottleneck Models (CBM) have emerged as an active interpretable framework incorporating human-interpretable concepts into decision-making. However, their concept predictions may lack reliability when applied to clinical diagnosis, impeding concept explanations' quality. To address this, we propose an evidential Concept Embedding Model (evi-CEM), which employs evidential learning to model the concept uncertainty. Additionally, we offer to leverage the concept uncertainty to rectify concept misalignments that arise when training CBMs using vision-language models without complete concept supervision. With the proposed methods, we can enhance concept explanations' reliability for both supervised and label-efficient settings. Furthermore, we introduce concept uncertainty for effective test-time intervention. Our evaluation demonstrates that evi-CEM achieves superior performance in terms of concept prediction, and the proposed concept rectification effectively mitigates concept misalignments for label-efficient training. Our code is available at https://github.com/obiyoag/evi-CEM.