NeurIPS 2020

GramGAN: Deep 3D Texture Synthesis From 2D Exemplars

Meta Review

Reviewers were unanimous in their appraisal of the submission's clarity and that the technical achievement was non-trivial and useful to the community, although views differed on the nature of what that achievement actually was (R2 contends that the "machine learning" contributions are thin, while this is disputed in the rebuttal; R2 did not update their review in response as of this writing). Regardless of the above, the paper's evaluations are strong enough for me to unreservedly recommend acceptance. I would nonetheless urge the authors to incorporate feedback of R3 and R4 in the camera-ready.