A Variational Formula for Rényi Divergences

07/07/2020
by   Jeremiah Birrell, et al.
0

We derive a new variational formula for the Rényi family of divergences, R_α(QP), generalizing the classical Donsker-Varadhan variational formula for the Kullback-Leibler divergence. The objective functional in this new variational representation is expressed in terms of expectations under Q and P, and hence can be estimated using samples from the two distributions. We illustrate the utility of such a variational formula by constructing neural-network estimators for the Rényi divergences.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
04/21/2021

A note on some information-theoretic divergences between Zeta distributions

In this short communication, we first report a closed-form formula for c...
research
10/10/2022

Function-space regularized Rényi divergences

We propose a new family of regularized Rényi divergences parametrized no...
research
10/05/2022

Rediscovery of Numerical Lüscher's Formula from the Neural Network

We present that by predicting the spectrum in discrete space from the ph...
research
06/18/2012

Tighter Variational Representations of f-Divergences via Restriction to Probability Measures

We show that the variational representations for f-divergences currently...
research
04/30/2020

The Canny-Emiris conjecture for the sparse resultant

We present a product formula for the initial parts of the sparse resulta...
research
08/04/2019

Improved GM(1,1) model based on Simpson formula and its applications

The classical GM(1,1) model is an efficient tool to make accurate foreca...
research
01/05/2018

Variable-Length Intrinsic Randomness Allowing Positive Value of the Average Variational Distance

This paper considers the problem of variable-length intrinsic randomness...

Please sign up or login with your details

Forgot password? Click here to reset