用少量标注的局部部件，生成大规模模型决策的全局可解释说明。
Generating Part-Based Global Explanations Via Correspondence
- 通过少量图像中的部件标签，迁移至更大数据集
- 聚合局部部件解释，生成全局符号化解释
- 适合需要可解释性的工业级模型部署场景
深度学习模型通常缺乏透明性。现有解释方法多聚焦于单个图像的局部视觉解释；而概念型解释虽能提供全局洞察，却需大量人工标注，成本高昂。本文提出一种新方法：利用少量图像中用户定义的部件标签，高效迁移到更大规模数据集上，通过聚合基于部件的局部解释，生成全局符号化解释，从而实现对大规模模型决策的人类可理解解释。
原文摘要 · Abstract (English)
Deep learning models are notoriously opaque. Existing explanation methods often focus on localized visual explanations for individual images. Concept-based explanations, while offering global insights, require extensive annotations, incurring significant labeling cost. We propose an approach that leverages user-defined part labels from a limited set of images and efficiently transfers them to a larger dataset. This enables the generation of global symbolic explanations by aggregating part-based local explanations, ultimately providing human-understandable explanations for model decisions on a large scale.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。