Metraipolis › archive › dc_a1d33265ea61

Full-Covariance Smoothing of Bayesian Neural Networks for Online Adaptation — arXiv:2609.27244

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.27244
License (as archived)
CC0-1.0
Archived
2026-10-08T10:52:12.818Z
SHA-256 of the archived text
21155124b082bbdf0fe5c296a1c967bfcef40bcddde90cf5c6d76cccdb3aea3e
Stable URL of this page
https://metraipolis.com/documento/dc_a1d33265ea61

How to cite

Full-Covariance Smoothing of Bayesian Neural Networks for Online Adaptation — arXiv:2609.27244 — arXiv (en), https://arxiv.org/abs/2609.27244 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a1d33265ea61 (Metraipolis archive dc_a1d33265ea61, sha256 21155124b082bbdf0fe5c296a1c967bfcef40bcddde90cf5c6d76cccdb3aea3e)
BibTeX
@misc{dca1d33265ea61,
  title = {Full-Covariance Smoothing of Bayesian Neural Networks for Online Adaptation — arXiv:2609.27244},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.27244},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a1d33265ea61; sha256 21155124b082bbdf0fe5c296a1c967bfcef40bcddde90cf5c6d76cccdb3aea3e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.27244. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A neural network's layers can be treated as time steps of a state-space model, turning Bayesian training into a smoothing problem: a forward pass propagates Gaussian moments through the network, and a backward Rauch--Tung--Striebel pass updates the weight posteriors in closed form. Such methods learn from each observation in a single pass, in an uncertainty-aware manner, and without gradient-based iterations or replay, which makes them well suited for online adaptation and data-efficient learning. Existing smoothing-based methods, however, are restricted to diagonal covariances across activations, discarding correlations between neurons. We overcome this limitation via a cross-covariance identity that enables full-covariance propagation through a network's nonlinear activations. We derive a one-step-per-layer smoother that approximates as Gaussian only each layer's affine output, and that applies both to deterministic systems with noisy observations and to stochastic systems described by output statistics. We demonstrate this method in non-stationary classification, online dynamics learning, and policy adaptation of a vision-language-action model, and find that it is generally more accurate than other smoothing-based methods.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-24 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Systems and Control (eess.SY).
Pagina del lavoro: https://arxiv.org/abs/2609.27244 — 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 2026-10-08T10:06:16.715Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).