论文标题
黑盒分类器的信息理论视觉说明
Information-Theoretic Visual Explanation for Black-Box Classifiers
论文作者
论文摘要
在这项工作中，我们尝试从信息理论的角度解释任何黑盒分类器的预测。对于每个输入功能，我们使用两个信息理论指标将分类器输出使用或没有该功能进行比较。因此，我们获得了两个归因地图 - 信息增益（IG）图和一份点的互信息（PMI）图。 IG地图提供了与“每个像素有多信息？”的独立答案，而PMI地图提供了特定于班级的解释“每个像素支持特定类别？”与现有方法相比，我们的方法从定量度量标准方面提高了归因图的正确性。我们还使用建议的方法对ImageNet分类器进行了详细的分析，并且该代码可在线提供。
In this work, we attempt to explain the prediction of any black-box classifier from an information-theoretic perspective. For each input feature, we compare the classifier outputs with and without that feature using two information-theoretic metrics. Accordingly, we obtain two attribution maps--an information gain (IG) map and a point-wise mutual information (PMI) map. IG map provides a class-independent answer to "How informative is each pixel?", and PMI map offers a class-specific explanation of "How much does each pixel support a specific class?" Compared to existing methods, our method improves the correctness of the attribution maps in terms of a quantitative metric. We also provide a detailed analysis of an ImageNet classifier using the proposed method, and the code is available online.
