用可解释的视觉概念比较深度模型差异，发现它们各自独有的特征。
Representational Similarity via Interpretable Visual Concepts
- 通过可解释的视觉概念对比两模型表征
- 发现某些差异源于单个模型独有的概念
- 适用于不同架构与训练方式的模型对比
深度神经网络如何做出决策？模型间的表征相似性测量长期悬而未决。现有方法仅提供单一数值，无法揭示差异来源。本文提出可解释表征相似性方法（RSVC），用于发现两个模型共享与独有的视觉概念。结果显示，部分模型差异可归因于某一模型独有的概念在另一模型中未能充分表示。我们在多种视觉模型架构和训练协议下进行广泛评估，验证了该方法的有效性。
原文摘要 · Abstract (English)
How do two deep neural networks differ in how they arrive at a decision? Measuring the similarity of deep networks has been a long-standing open question. Most existing methods provide a single number to measure the similarity of two networks at a given layer, but give no insight into what makes them similar or dissimilar. We introduce an interpretable representational similarity method (RSVC) to compare two networks. We use RSVC to discover shared and unique visual concepts between two models. We show that some aspects of model differences can be attributed to unique concepts discovered by one model that are not well represented in the other. Finally, we conduct extensive evaluation across different vision model architectures and training protocols to demonstrate its effectiveness.
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。