Stepwise Variational Inference with Vine Copulas

Elisabeth Griesbauer, Leiv Rønneberg, Arnoldo Frigessi, Claudia Czado, Ingrid Hobæk Haff · 2026

Read Stepwise Variational Inference with Vine Copulas by Elisabeth Griesbauer, Leiv Rønneberg, Arnoldo Frigessi, Claudia Czado, Ingrid Hobæk Haff (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-22959.txt.

Description

Griesbauer and co-authors propose stepwise variational inference with vine copulas. Variational inference approximates a posterior, often with weak dependence structure. Vine copulas build dependence from nested trees of simple copulas. The authors estimate the approximation tree by tree, step by step. The usual divergence fails to recover the right parameters here. They define the training bound with Rényi divergence instead. A stopping rule decides when to add more trees. The method sits between mean-field and full dependence, using few parameters.

Catalog record

CREATORElisabeth Griesbauer, Leiv Rønneberg, Arnoldo Frigessi, Claudia Czado, Ingrid Hobæk Haff
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE134083
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.22959

Browse the full library · About the Tower of Babel