NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:4084
Title:Neural Spline Flows

This paper proposes a spline-based parametric approximation for monotonic functions as a module to be used in flow-based models. All reviewers were impressed with the work, which has exciting practical relevance across a variety of applications. Reviewers also appreciated the author response.