95e62984b87e90645a5cf77037395959-AuthorFeedback.pdf

We thank all four reviewers for their great reviews. We provide our feedback for each reviewer as follows. Due to the difficulty of evaluation, we can only use small scale datasets (e.g. CIFAR) to generate figures like Figure 1 (a-c) in our paper to show the correctness. MNIST and CIFAR are standard datasets in influence function or data cleansing literature.