Partially Shared Concept Bottleneck Models

TL;DR

This paper introduces PS-CBM, a novel framework for concept bottleneck models that improves interpretability and accuracy by integrating multimodal concepts, merging concepts strategically, and balancing accuracy with concept compactness.

Contribution

PS-CBM addresses key limitations of existing CBMs by proposing a multimodal concept generator, a partially shared concept strategy, and a new metric for balancing accuracy and interpretability.

Findings

Outperforms state-of-the-art CBMs on eleven datasets.

Improves classification accuracy by 1.0%-7.4%.

Enhances interpretability with fewer concepts and better metrics.

Abstract

Concept Bottleneck Models (CBMs) enhance interpretability by introducing a layer of human-understandable concepts between inputs and predictions. While recent methods automate concept generation using Large Language Models (LLMs) and Vision-Language Models (VLMs), they still face three fundamental challenges: poor visual grounding, concept redundancy, and the absence of principled metrics to balance predictive accuracy and concept compactness. We introduce PS-CBM, a Partially Shared CBM framework that addresses these limitations through three core components: (1) a multimodal concept generator that integrates LLM-derived semantics with exemplar-based visual cues; (2) a Partially Shared Concept Strategy that merges concepts based on activation patterns to balance specificity and compactness; and (3) Concept-Efficient Accuracy (CEA), a post-hoc metric that jointly captures both predictive accuracy and concept compactness. Extensive experiments on eleven diverse datasets show that PS-CBM consistently outperforms state-of-the-art CBMs, improving classification accuracy by 1.0%-7.4% and CEA by 2.0%-9.5%, while requiring significantly fewer concepts. These results underscore PS-CBM's effectiveness in achieving both high accuracy and strong interpretability.