Review for NeurIPS paper: Learning Kernel Tests Without Data Splitting

The reviewers found the theoretical work sound. In the final version, the potential extensions of the method to other frameworks could be addressed.