Profile Maximum Likelihood is Optimal for Estimating KL Divergence.

Jayadev Acharya

Browse the full ISIT paper archive.