NeurIPS 2020

The Generalized Lasso with Nonlinear Observations and Generative Priors

Meta Review

This paper proposes a novel analysis for reconstruction of signals from nonlinear measurements using what's called K-Lasso, leveraging on generative priors. This is an important emerging area and this is a solid contribution. In our the updated reviews there are suggestions for improving the presentation, in particular section 4.