NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:3011
Title:Invertible Convolutional Flow


		
This is a very good submission in the area of generative probabilistic models. The authors introduce normalizing flows using circular and symmetric convolutions leading to efficient Jacobian determinant computation. Several experiments have been conducted to demonstrate the model showing good results. Please revise the paper in light of the pre and post rebuttal comments of the reviewers. Congratulations on this excellent piece of work!