Evidential Concept Embedding Models: Towards Reliable Concept Explanations for Skin Disease Diagnosis

TL;DR

This paper introduces an evidential concept embedding model (evi-CEM) that improves the reliability of concept explanations in medical image diagnosis by modeling concept uncertainty and rectifying misalignments, enhancing interpretability and performance.

Contribution

The paper proposes a novel evidential learning approach for concept embedding models, addressing reliability issues and concept misalignments in medical diagnosis applications.

Findings

evi-CEM outperforms existing models in concept prediction accuracy

Concept rectification reduces misalignments in label-efficient training

Enhanced interpretability through reliable concept explanations

Abstract

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.