2e255d2d6bf9bb33030246d31f1a79ca-AuthorFeedback.pdf

Neural Information Processing Systems 

Reviewer 1 We thank the reviewer for their helpful comments. The reviewer's primary concern is this work's relationship with existing literature. Our paper includes a detailed comparison to Finlay et al. 2020 that addresses the reviewer's concerns about our Our regularizer, like theirs, is augmenting the loss with an extra penalty term. Empirically, we extensively compare to Finlay et al. 2020 in our experiments on As we explain in Section 6.2, we agree that K = 3 with a 3rd order solver shows a marginal improvement over other Reviewer 2 We will remove the qualifiers on l. 30 and add more detail on l. 47. We will include the detail on l. 64.