NeurIPS 2020

Ultrahyperbolic Representation Learning

Meta Review

The paper was carefully reviewed by four domain experts, who found in particular the theoretical contributions sound and useful. Both the reviews and the post-rebuttal discussion indicated, however, that the authors would very much have liked to see some indication, intuition or understanding of for which type of this particular representation would be of use. I encourage the authors to think about this for the camera ready version; this could significantly improve the impact of the paper.