trends.barnas.net

method · first seen 22 Oct 2026

Bayesian learning rule

Also: BLR

An approach deriving efficient natural-gradient-style additive updates for exponential-family variational posteriors using Bayesian updating and KL divergence.

Mentions per week

0119 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

Bregman divergenceinformation-geometric optimization