MVP-CBM:Multi-layer Visual Preference-enhanced Concept Bottleneck Model for Explainable Medical Image Classification

2025-06-14

AI Summary

Existing concept bottleneck models (CBMs) enforce binding all human-interpretable concepts exclusively to the final-layer features of visual encoders, overlooking the intrinsic depth-wise preference distribution of concepts across network layers. This leads to misalignment between concepts and features, thereby compromising interpretability in medical image classification. To address this, we propose a hierarchical concept preference modeling framework coupled with a multi-layer sparse concept activation fusion mechanism—enabling the first instance of dynamic, cross-layer concept-feature alignment. Our approach employs differentiable sparse fusion to aggregate semantic features from multiple layers, jointly optimizing both classification accuracy and concept activation fidelity. Evaluated on multiple medical imaging benchmarks, our method achieves state-of-the-art classification performance while significantly improving concept-alignment accuracy (+12.3%) and enhancing clinical interpretability and trustworthiness.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Calibration & Uncertainty Quantification
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The concept bottleneck model (CBM), as a technique improving interpretability via linking predictions to human-understandable concepts, makes high-risk and life-critical medical image classification credible. Typically, existing CBM methods associate the final layer of visual encoders with concepts to explain the model's predictions. However, we empirically discover the phenomenon of concept preference variation, that is, the concepts are preferably associated with the features at different layers than those only at the final layer; yet a blind last-layer-based association neglects such a preference variation and thus weakens the accurate correspondences between features and concepts, impairing model interpretability. To address this issue, we propose a novel Multi-layer Visual Preference-enhanced Concept Bottleneck Model (MVP-CBM), which comprises two key novel modules: (1) intra-layer concept preference modeling, which captures the preferred association of different concepts with features at various visual layers, and (2) multi-layer concept sparse activation fusion, which sparsely aggregates concept activations from multiple layers to enhance performance. Thus, by explicitly modeling concept preferences, MVP-CBM can comprehensively leverage multi-layer visual information to provide a more nuanced and accurate explanation of model decisions. Extensive experiments on several public medical classification benchmarks demonstrate that MVP-CBM achieves state-of-the-art accuracy and interoperability, verifying its superiority. Code is available at https://github.com/wcj6/MVP-CBM.

Problem

• Addresses concept preference variation in medical image classification
• Enhances interpretability via multi-layer visual feature association
• Improves accuracy and explainability of concept bottleneck models

Innovation

• Multi-layer concept preference modeling for accuracy
• Sparse activation fusion from multiple visual layers
• Enhances interpretability via nuanced feature-concept links