NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:7779
Title:Calculating Optimistic Likelihoods Using (Geodesically) Convex Optimization

The paper seems well-written and motivated. Overall evaluations are positive, but the authors are strongly encouraged to take the concerns raised into account. Also, adding a compact ball as a constraint leads to a projection based method, where the authors should present results about the extra computational costs incurred due to that.