Invertible Concept-based Explanations for CNN Models with Non-negative Concept Activation Vectors

TL;DR

This paper introduces an invertible concept-based explanation framework for CNNs, utilizing non-negative matrix factorization to improve interpretability and fidelity of concept-level explanations.

Contribution

It proposes the ICE framework with NCAVs, enhancing explanation quality over previous methods like ACE through novel matrix factorization techniques.

Findings

NCAVs outperform other methods in interpretability

NCAVs achieve higher fidelity in explanations

Framework provides both local and global explanations

Abstract

Convolutional neural network (CNN) models for computer vision are powerful but lack explainability in their most basic form. This deficiency remains a key challenge when applying CNNs in important domains. Recent work on explanations through feature importance of approximate linear models has moved from input-level features (pixels or segments) to features from mid-layer feature maps in the form of concept activation vectors (CAVs). CAVs contain concept-level information and could be learned via clustering. In this work, we rethink the ACE algorithm of Ghorbani et~al., proposing an alternative invertible concept-based explanation (ICE) framework to overcome its shortcomings. Based on the requirements of fidelity (approximate models to target models) and interpretability (being meaningful to people), we design measurements and evaluate a range of matrix factorization methods with our framework. We find that non-negative concept activation vectors (NCAVs) from non-negative matrix factorization provide superior performance in interpretability and fidelity based on computational and human subject experiments. Our framework provides both local and global concept-level explanations for pre-trained CNN models.