Towards Achieving Concept Completeness for Textual Concept Bottleneck Models

TL;DR

This paper introduces CT-CBM, an unsupervised method for creating complete and interpretable concept bases in textual concept bottleneck models, improving interpretability and concept detection without human-labeled data.

Contribution

It presents a novel unsupervised approach to generate complete concept bases in TCBMs, reducing reliance on human annotations and enhancing interpretability.

Findings

Outperforms competitors in concept basis completeness

Achieves high concept detection accuracy

Eliminates need for human-labeled concepts

Abstract

Textual Concept Bottleneck Models (TCBMs) are interpretable-by-design models for text classification that predict a set of salient concepts before making the final prediction. This paper proposes Complete Textual Concept Bottleneck Model (CT-CBM), a novel TCBM generator building concept labels in a fully unsupervised manner using a small language model, eliminating both the need for predefined human labeled concepts and LLM annotations. CT-CBM iteratively targets and adds important and identifiable concepts in the bottleneck layer to create a complete concept basis. CT-CBM achieves striking results against competitors in terms of concept basis completeness and concept detection accuracy, offering a promising solution to reliably enhance interpretability of NLP classifiers.