Metraipolis › archive › dc_f059d2fb78f0

When Does Jacobian Fine-Tuning Stabilize Neural PDE Emulators? — arXiv:2610.42013

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.42013
License (as archived)
CC0-1.0
Archived
2026-11-03T08:12:37.329Z
SHA-256 of the archived text
2a5b3fad7d5ebb1ec4e3178f408108230a2957e8e47d8773bf4f97331a352b9c
Stable URL of this page
https://metraipolis.com/documento/dc_f059d2fb78f0

How to cite

When Does Jacobian Fine-Tuning Stabilize Neural PDE Emulators? — arXiv:2610.42013 — arXiv (en), https://arxiv.org/abs/2610.42013 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f059d2fb78f0 (Metraipolis archive dc_f059d2fb78f0, sha256 2a5b3fad7d5ebb1ec4e3178f408108230a2957e8e47d8773bf4f97331a352b9c)
BibTeX
@misc{dcf059d2fb78f0,
  title = {When Does Jacobian Fine-Tuning Stabilize Neural PDE Emulators? — arXiv:2610.42013},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.42013},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f059d2fb78f0; sha256 2a5b3fad7d5ebb1ec4e3178f408108230a2957e8e47d8773bf4f97331a352b9c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.42013. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Neural emulators of chaotic PDEs are trained on one-step prediction but deployed autoregressively, where their errors compound. Jacobian-based regularisers aim to curb this growth, but in a chaotic system part of the growth is physical and cannot be trained away. We separate the two sources of this error with a first-order decomposition. We compare the emulator's error with a counterfactual that propagates the same one-step mistakes through the solver's Jacobian, giving an amplification ratio: the factor by which the model's own dynamics amplify its errors. We study 21 configurations across four PDEs, five architectures and two model sizes (8-10k and 33k parameters), with errors measured at one Lyapunov time. Jacobian fine-tuning reduced the error by more than 10% for 1 of 34 seeds whose amplification ratio was at most 1.25, and for all 20 seeds whose ratio was at least four; for the latter, the error fell by a factor of 8 (small models) and 55 (33k models) in geometric mean over seeds. The gain approximately equals the one-step accuracy cost of fine-tuning divided by the amplification ratio. Matching the solver's Jacobian along the model's own error direction (EDJM) outperformed random directions at the same cost on 28 of 35 seeds; input-noise injection performed comparably to EDJM.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.42013 — 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 2026-11-03T08:02:38.948Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).