Paper proposes efficient approximation for KL divergence between discrete normal distributions

FrnkNlsn · x · 2026-08-15

This paper from 2021 discusses the calculation of Kullback-Leibler divergence between discrete normal distributions, which are defined as maximum entropy distributions on integer lattices with prescribed means and covariance matrices. The study presents formulas for common statistical divergences and proposes an efficient approximation technique for calculating the KL divergence via Rényi $\alpha$-divergences or projective $\gamma$-divergences.

Original post →

More from Research

Research channel →