Grad-CAM++ is Equivalent to Grad-CAM With Positive Gradients
This is an incremental analysis for researchers in interpretable AI, clarifying the novelty of existing methods.
The paper tackles the problem of evaluating the Grad-CAM++ method for visual explanations in deep networks, showing that it is equivalent to a simple variation of Grad-CAM using positive gradients.
The Grad-CAM algorithm provides a way to identify what parts of an image contribute most to the output of a classifier deep network. The algorithm is simple and widely used for localization of objects in an image, although some researchers have point out its limitations, and proposed various alternatives. One of them is Grad-CAM++, that according to its authors can provide better visual explanations for network predictions, and does a better job at locating objects even for occurrences of multiple object instances in a single image. Here we show that Grad-CAM++ is practically equivalent to a very simple variation of Grad-CAM in which gradients are replaced with positive gradients.