Toward computing attributions for dimensionality reduction techniques
Description
SUMMARY: We describe the problem of computing local feature attributions for dimensionality reduction methods. We use one such method that is well established within the context of supervised classification-using the gradients of target outputs with
