Adaptive Variational Inference in Probabilistic Graphical Models: Beyond Bethe, Tree-Reweighted, and Convex Free Energies
Harald Leisenberger [email protected], Franz Pernkopf [email protected] · 2025
Read Adaptive Variational Inference in Probabilistic Graphical Models: Beyond Bethe, Tree-Reweighted, and Convex Free Energies by Harald Leisenberger [email protected], Franz Pernkopf [email protected] (2025) 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-2502-03341.txt.
Description
Harald Leisenberger and Franz Pernkopf work in 2025 on variational inference, a way to guess key values in probability models that are too hard to compute exactly. Well known shortcuts for this task can fail when a model is complex and its parts interact a lot.
The authors study two broad classes of shortcuts that contain the known ones as special cases. One class changes the model values, and the other changes how entropy is guessed. They weigh the gains and limits of each and sketch how an ideal free energy shortcut should be built. Their new shortcuts adapt to each model on their own and did well on a range of hard problems.
Catalog record
| CREATOR | Harald Leisenberger [email protected], Franz Pernkopf [email protected] |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-03341