Metraipolis › archive › dc_fff3ce0f66e6
Amortized Bayesian Inference on Multilevel Models of Arbitrary Structure — arXiv:2609.40024
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.40024
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T13:47:51.284Z
- SHA-256 of the archived text
8f1ac9d96522a2261a3602b2da8ce8b0f95d1507880d65b5800f257046df7847- Stable URL of this page
https://metraipolis.com/documento/dc_fff3ce0f66e6
How to cite
Amortized Bayesian Inference on Multilevel Models of Arbitrary Structure — arXiv:2609.40024 — arXiv (en), https://arxiv.org/abs/2609.40024 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fff3ce0f66e6 (Metraipolis archive dc_fff3ce0f66e6, sha256 8f1ac9d96522a2261a3602b2da8ce8b0f95d1507880d65b5800f257046df7847)
BibTeX
@misc{dcfff3ce0f66e6,
title = {Amortized Bayesian Inference on Multilevel Models of Arbitrary Structure — arXiv:2609.40024},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.40024},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fff3ce0f66e6; sha256 8f1ac9d96522a2261a3602b2da8ce8b0f95d1507880d65b5800f257046df7847},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.40024. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We develop a general method for amortized Bayesian inference on multilevel models of arbitrary structure. Given a generative model specified as a directed acyclic graph, our method automatically derives valid factorizations of the joint posterior and matching neural network architectures. The key steps, graph expansion and graph inversion, yield an inverse graph that determines how inference networks are stacked and conditioned, producing factorizations that amortize over the number of groups and the number of observations within each group. Unlike approaches that simplify the dependency structure to speed up learning or inference, our method preserves all conditional independence and exchangeability assumptions of the generative model. Across three case studies, it closely matches gold-standard samplers on models with more than 6,500 parameters while reducing inference to a near-instant forward pass once trained. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG). Categorie arXiv: Machine Learning (stat.ML) ; Machine Learning (cs.LG); Computation (stat.CO). Pagina del lavoro: https://arxiv.org/abs/2609.40024 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T13:18:10.688Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).