Open unnir opened 2 years ago
Hi,
Thanks for your question. ShapNet has been designed to be an intrinsic explanation model, which means that it explains itself and only itself by design.
However, you could learn a ShapNet that mimic the behavior of your model of choice, which could work in theory, but we have no idea how well.
Best.
Thank you for your answer! Also, I think the paper :+1:
Is it possible to explain a black box model with Shapley networks? Suppose, I have a trained CNN model, and I want to get feature attributions for a single image from IMAGENET data set.