Constructing Concept-based Models to Mitigate Spurious Correlations with Minimal Human Effort

TL;DR

This paper introduces a novel framework that leverages multiple foundation models to construct concept bottleneck models with minimal human effort, effectively reducing spurious correlations and enhancing interpretability.

Contribution

The paper presents a new method that exploits foundation models to build concept bottleneck models with little human annotation, addressing biases and spurious correlations.

Findings

Effective reduction of spurious correlations in models.

Maintains interpretability while improving robustness.

Seamless pipeline for dataset bias assessment and concept annotation.

Abstract

Enhancing model interpretability can address spurious correlations by revealing how models draw their predictions. Concept Bottleneck Models (CBMs) can provide a principled way of disclosing and guiding model behaviors through human-understandable concepts, albeit at a high cost of human efforts in data annotation. In this paper, we leverage a synergy of multiple foundation models to construct CBMs with nearly no human effort. We discover undesirable biases in CBMs built on pre-trained models and propose a novel framework designed to exploit pre-trained models while being immune to these biases, thereby reducing vulnerability to spurious correlations. Specifically, our method offers a seamless pipeline that adopts foundation models for assessing potential spurious correlations in datasets, annotating concepts for images, and refining the annotations for improved robustness. We evaluate the proposed method on multiple datasets, and the results demonstrate its effectiveness in reducing model reliance on spurious correlations while preserving its interpretability.