AS-XAI: Self-Supervised Automatic Semantic Interpretation for CNN

Changqi Sun

doi:10.1002/aisy.202400359

2024-09-30

Changqi Sun

H. Xu

Yuntian Chen

张东晓 (Dongxiao Zhang) *

DOI：10.1002/aisy.202400359

原文链接

原文求助

摘要

En 中文

Explainable artificial intelligence (XAI) aims to develop transparent explanatory approaches for black-box deep learning models. However, it remains difficult for existing methods to achieve the trade-off of the three key criteria in interpretability, namely, reliability, understandability, and usability, which hinder their practical applications. In this article, we propose a self-supervised automatic semantic interpretable explainable artificial intelligence (AS-XAI) framework, which utilizes transparent orthogonal embedding semantic extraction spaces and row-centered principal component analysis (PCA) for global semantic interpretation of model decisions in the absence of human interference, without additional computational costs. In addition, the invariance of filter feature high-rank decomposition is used to evaluate model sensitivity to different semantic concepts. Extensive experiments demonstrate that robust and orthogonal semantic spaces can be automatically extracted by AS-XAI, providing more effective global interpretability for convolutional neural networks (CNNs) and generating human-comprehensible explanations. The proposed approach offers broad fine-grained extensible practical applications, including shared semantic interpretation under out-of-distribution (OOD) categories, auxiliary explanations for species that are challenging to distinguish, and classification explanations from various perspectives. In a systematic evaluation by users with varying levels of AI knowledge, AS-XAI demonstrated superior glass box characteristics. Explainable artificial intelligence (XAI) provides transparent deep learning explanations. This article introduces self-supervised automatic semantic interpretable XAI (AS-XAI), a framework using orthogonal embedding spaces and principal component analysis (PCA) for global semantic interpretation, and offers effective interpretability for convolutional neural networks (CNNs), including out-of-distribution (OOD) category interpretation and species classification, with minimal computational cost.image (c) 2024 WILEY-VCH GmbH

Keyword:

convolutional neural network

interpretable machine learning

out-of-distribution detection

self-supervised learning

semantic space