1c336b8080f82bcc2cd2499b4c57261d-AuthorFeedback.pdf
–Neural Information Processing Systems
Re: Reviewer#2 Thank you for the positive comments. About line 160: We are sorry for the misunderstanding here. Indeed, showing the equivalence of NNs trained by38 gradient flow and NTK kernel predictor is non-trivial. This result is rigorously proved in [22].
Neural Information Processing Systems
Feb-7-2026, 16:46:28 GMT