Semi-supervised Concept Bottleneck Models

TL;DR

This paper introduces SSCBM, a semi-supervised framework for Concept Bottleneck Models that reduces reliance on extensive labeled data by leveraging unlabeled data and alignment strategies, maintaining high accuracy.

Contribution

The paper proposes SSCBM, a novel semi-supervised approach that improves concept bottleneck models by effectively utilizing unlabeled data and addressing annotation alignment issues.

Findings

Achieves comparable accuracy with only 10% labeled data

Effectively generates pseudo labels and aligns concepts

Maintains high task and concept accuracy with limited annotations

Abstract

Concept Bottleneck Models (CBMs) have garnered increasing attention due to their ability to provide concept-based explanations for black-box deep learning models while achieving high final prediction accuracy using human-like concepts. However, the training of current CBMs is heavily dependent on the precision and richness of the annotated concepts in the dataset. These concept labels are typically provided by experts, which can be costly and require significant resources and effort. Additionally, concept saliency maps frequently misalign with input saliency maps, causing concept predictions to correspond to irrelevant input features - an issue related to annotation alignment. To address these limitations, we propose a new framework called SSCBM (Semi-supervised Concept Bottleneck Model). Our SSCBM is suitable for practical situations where annotated data is scarce. By leveraging joint training on both labeled and unlabeled data and aligning the unlabeled data at the concept level, we effectively solve these issues. We proposed a strategy to generate pseudo labels and an alignment loss. Experiments demonstrate that our SSCBM is both effective and efficient. With only 10% labeled data, our model's concept and task accuracy on average across four datasets is only 2.44% and 3.93% lower, respectively, compared to the best baseline in the fully supervised learning setting.