Label-Free Concept Bottleneck Models

Link
Abstract

Concept bottleneck models (CBM) are a popular way of creating more interpretable neural networks by having hidden layer neurons correspond to human-understandable concepts. However, existing CBMs and their variants have two crucial limitations: first, they need to collect labeled data for each of the predefined concepts, which is time consuming and labor intensive; second, the accuracy of a CBM is often significantly lower than that of a standard neural network, especially on more complex datasets. This poor performance creates a barrier for adopting CBMs in practical real world applications. Motivated by these challenges, we propose Label-free CBM which is a novel framework to transform any neural network into an interpretable CBM without labeled concept data, while retaining a high accuracy. Our Label-free CBM has many advantages, it is: scalable - we present the first CBM scaled to ImageNet, efficient - creating a CBM takes only a few hours even for very large datasets, and automated - training it for a new dataset requires minimal human effort. Our code is available at https://github.com/Trustworthy-ML-Lab/Label-free-CBM. Finally, in Appendix B we conduct a large scale user evaluation of the interpretability of our method.

Synth

Problem:: 기존 Concept Bottleneck Model(CBM)은 Concept 레이블 데이터가 반드시 필요함 / 이로 인해 데이터 구축 비용과 시간이 많이 소요됨 / 또한, 일반 신경망에 비해 정확도가 저하되는 한계가 있음

Solution:: 데이터와 관련된 Concept을 GPT를 이용해 자동으로 생성 / CLIP을 이용해 Concept Text와 이미지간의 간의 활성화 패턴을 만들고 기존 Feature를 이용해 해당 활성화 패턴을 예측하도록 학습

Novelty:: GPT-3를 활용해 데이터셋에 맞는 Concept 집합을 자동으로 생성 / CBM을 ImageNet과 같은 대규모 데이터셋에 성공적으로 적용하고 높은 성능을 달성 / 모델의 해석 가능성을 이용해 수동으로 가중치를 편집하여 실제 모델 정확도를 향상시킨 사례 제시

Note:: 학습된 Feature를 해석 가능한 공간인 CLIP Feature Space와 연결하여 Interpretability를 향상시킴

Summary

Motivation

Method

file-20250618164749462.png|750

제안하는 Label-free CBM은 어떠한 신경망 Backbone이라도 총 4단계의 과정을 통해 해석 가능한 CBM으로 변환

Method 검증