Coarse-to-Fine Concept Bottleneck Models

TL;DR

This paper introduces a novel two-level concept discovery framework for interpretable deep learning models, leveraging vision-language models and Bayesian methods to improve concept granularity and hierarchy understanding.

Contribution

It proposes a coarse-to-fine concept selection approach using a hierarchical framework and Bayesian sparsity, enhancing interpretability and performance over existing CBMs.

Findings

Outperforms recent CBM approaches in experiments

Uncovers granular, patch-specific concepts within images

Provides a principled interpretability framework

Abstract

Deep learning algorithms have recently gained significant attention due to their impressive performance. However, their high complexity and un-interpretable mode of operation hinders their confident deployment in real-world safety-critical tasks. This work targets ante hoc interpretability, and specifically Concept Bottleneck Models (CBMs). Our goal is to design a framework that admits a highly interpretable decision making process with respect to human understandable concepts, on two levels of granularity. To this end, we propose a novel two-level concept discovery formulation leveraging: (i) recent advances in vision-language models, and (ii) an innovative formulation for coarse-to-fine concept selection via data-driven and sparsity-inducing Bayesian arguments. Within this framework, concept information does not solely rely on the similarity between the whole image and general unstructured concepts; instead, we introduce the notion of concept hierarchy to uncover and exploit more granular concept information residing in patch-specific regions of the image scene. As we experimentally show, the proposed construction not only outperforms recent CBM approaches, but also yields a principled framework towards interpetability.