NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5145
Title:A Benchmark for Interpretability Methods in Deep Neural Networks

Dear authors, congrats on the acceptance-- this paper was discussed extensively and controversially. In balance, we decided that the straightforward metric and empirical results will be interesting for Neurips, although at the same time it is clear that there is work to be done: The reviewers provided multiple comments and a lot of feedback, and it will be of critical importance that you revise the manuscript accordingly (and as you promised in your rebuttal)- in particular, this would require significant changes in the exposition, and also a more clear description of the novelty of the study in light of previous work (in addition to the suggestions by the reviewers, it strikes me that the classic literature on recursive feature elimination could really be mentioned at least).