ProtoPShare: Prototypical Parts Sharing for Similarity Discovery in Interpretable Image Classification
Knowledge Discovery and Data Mining(2021)
摘要
ABSTRACTIn this work, we introduce an extension to ProtoPNet called ProtoPShare which shares prototypical parts between classes. To obtain prototype sharing we prune prototypical parts using a novel data-dependent similarity. Our approach substantially reduces the number of prototypes needed to preserve baseline accuracy and finds prototypical similarities between classes. We show the effectiveness of ProtoPShare on the CUB-200-2011 and the Stanford Cars datasets and confirm the semantic consistency of its prototypical parts in user-study.
更多查看译文
关键词
prototypical parts, interpretability, explainability, neural networks
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要