Metraipolis › archive › dc_44aba6561605

Preconditioned Physics-Informed Neural Operator Training — arXiv:2609.36216

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.36216
License (as archived)
CC0-1.0
Archived
2026-10-07T17:28:49.033Z
SHA-256 of the archived text
2655d3155a3cae92ad90802dd159868cd2db498e26a424e79e4935704524fee5
Stable URL of this page
https://metraipolis.com/documento/dc_44aba6561605

How to cite

Preconditioned Physics-Informed Neural Operator Training — arXiv:2609.36216 — arXiv (en), https://arxiv.org/abs/2609.36216 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_44aba6561605 (Metraipolis archive dc_44aba6561605, sha256 2655d3155a3cae92ad90802dd159868cd2db498e26a424e79e4935704524fee5)
BibTeX
@misc{dc44aba6561605,
  title = {Preconditioned Physics-Informed Neural Operator Training — arXiv:2609.36216},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.36216},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_44aba6561605; sha256 2655d3155a3cae92ad90802dd159868cd2db498e26a424e79e4935704524fee5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36216. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Neural operators are typically trained in a supervised fashion, which requires a dataset to be generated with a classical solver. Training them physics-informed, i.e., purely from the governing equations, removes this large offline cost and allows fresh samples to be drawn at every optimization step, but has so far been limited to simplified problems and trails supervised training in accuracy. The obstacle is the ill-conditioning of physics-informed losses, which differential operators induce and which worsens as the discretization is refined. We therefore propose a preconditioned residual loss function and show mesh-independent conditioning for elliptic problems and greatly improved conditioning for saddle point problems. Realized through geometric and algebraic multigrid, the construction applies to linear and nonlinear equations, steady or time-dependent, on structured and unstructured meshes, is agnostic to the neural operator architecture, and adds no cost at inference. On the Poisson, Allen-Cahn and stationary Stokes equations, the resulting label-free training matches supervised training and is four to twenty-five times more accurate than previous physics-informed operator learning methods.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Numerical Analysis (math.NA).
Pagina del lavoro: https://arxiv.org/abs/2609.36216 — 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 2026-10-07T16:59:48.120Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).