NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5897
Title:The Broad Optimality of Profile Maximum Likelihood

The paper shows that profile maximum likelihood, an idea from the distribution estimation literature from a couple of years ago, enjoys optimality properties for a large class of property estimation tasks. The class of tasks includes a number of popular problems studied in the distribution learning literature. All reviewers liked the paper and advocate acceptance. Please do go over the reviews and incorporate any feedback for the camera ready.