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.
View on arXiv