Metraipolis › archive › dc_041427d05f4e

Acceleration and its Limits in Gaussian Variational Inference — arXiv:2610.21020

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21020
License (as archived)
CC0-1.0
Archived
2026-10-16T10:25:08.025Z
SHA-256 of the archived text
0118871dffa621f5318dfa0025a0adddf0ee9c8ad0c98cf86683c67610f26385
Stable URL of this page
https://metraipolis.com/documento/dc_041427d05f4e

How to cite

Acceleration and its Limits in Gaussian Variational Inference — arXiv:2610.21020 — arXiv (en), https://arxiv.org/abs/2610.21020 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_041427d05f4e (Metraipolis archive dc_041427d05f4e, sha256 0118871dffa621f5318dfa0025a0adddf0ee9c8ad0c98cf86683c67610f26385)
BibTeX
@misc{dc041427d05f4e,
  title = {Acceleration and its Limits in Gaussian Variational Inference — arXiv:2610.21020},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21020},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_041427d05f4e; sha256 0118871dffa621f5318dfa0025a0adddf0ee9c8ad0c98cf86683c67610f26385},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21020. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We study whether the classical momentum-based acceleration that improves the condition-number dependence of gradient descent from $\kappa$ to $\sqrt{\kappa}$ for strongly convex, smooth objectives extends to Gaussian variational inference (GVI) for strongly log-concave, log-smooth target distributions. With exact Gaussian expectations, a convex composite formulation of GVI yields the classical accelerated $\sqrt{\kappa}$ convergence rate. However, when only pointwise derivatives are available, we construct a family of strongly log-concave targets for which any randomized algorithm requires $\kappa^{1-o(1)}$ queries to achieve a fixed objective accuracy, even when the initial objective gap is bounded independently of $\kappa$. Thus curvature bounds alone cannot guarantee uniform accelerated total-query complexity. We then ask which additional structure permits acceleration and give positive examples: under a Hessian-Lipschitz assumption, variance-reduced stochastic methods yield accelerated guarantees when the Hessian variation is sufficiently small. Finally, experiments on real datasets show substantial empirical speedups from acceleration across a broad range of methods.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (stat.ML) ; Machine Learning (cs.LG); Optimization and Control (math.OC).
Pagina del lavoro: https://arxiv.org/abs/2610.21020 — 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 2026-10-16T08:48:02.878Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).