NeurIPS 2020

Nonconvex Sparse Graph Learning under Laplacian Constrained Graphical Model

Meta Review

All reviewers but one agree that the contribution of the paper is novel, interesting and that the experiments are compelling. Bit all reviewers also agree that the paper lacks either an intuitive or even formal explanation of the surprizing properties of the L1 norm regularization in this setting. The proofs provided show that the stated facts are correct, but this is not explained in the text. The authors are strongly encouraged to include an explanation in the final version of the manuscript.