Metraipolis › archive › dc_db2e5c0f6ed4

Stability-Constrained Approximation in Spline KANs: Exact Layer Balancing and Budget-Compatible Saturation — arXiv:2609.17619

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.17619
License (as archived)
CC0-1.0
Archived
2026-10-15T15:05:05.824Z
SHA-256 of the archived text
6fec989efeb866bb18822e190a31010660400aeca237c23ad91be2d19e88d225
Stable URL of this page
https://metraipolis.com/documento/dc_db2e5c0f6ed4

How to cite

Stability-Constrained Approximation in Spline KANs: Exact Layer Balancing and Budget-Compatible Saturation — arXiv:2609.17619 — arXiv (en), https://arxiv.org/abs/2609.17619 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db2e5c0f6ed4 (Metraipolis archive dc_db2e5c0f6ed4, sha256 6fec989efeb866bb18822e190a31010660400aeca237c23ad91be2d19e88d225)
BibTeX
@misc{dcdb2e5c0f6ed4,
  title = {Stability-Constrained Approximation in Spline KANs: Exact Layer Balancing and Budget-Compatible Saturation — arXiv:2609.17619},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.17619},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db2e5c0f6ed4; sha256 6fec989efeb866bb18822e190a31010660400aeca237c23ad91be2d19e88d225},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.17619. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Deep spline superposition networks face a tension between approximation order and stability across depth. We study approximation under a hard layerwise Lipschitz budget, and organise it around two quantities: the factorisation stability complexity of a given deep factorisation, and the budget-compatible approximation complexity of a discretisation operator. First, we solve exactly the finite-depth diagonal balancing problem for a fixed chain of nonnegative envelope matrices: the optimal uniform layer budget equals $\|M_{L-1}\cdots M_0\|_{\infty\to\infty}^{1/L}$, attained by an explicit one-pass minimiser, for rectangular layers, with a complete treatment of degeneracies and non-attainment. The optimum can be arbitrarily larger than the Lipschitz constant of the network itself, because passing to envelopes destroys sign cancellation. Second, we give a constructive spline discretisation theorem preserving the budget up to a controlled slack, with an explicit grid threshold. Conversely, for linear spline-valued operators that preserve the budget exactly, we prove budget-compatible minimax lower bounds on classes constrained simultaneously in the first and third derivative norms -- a constraint pair that is forced by the problem and that rules out the usual scaling escapes. Finally, we show that the corresponding layer errors need not cancel under composition: for every operator of the class there is a stable depth-$L$ tower realising a constant fraction of the accumulated error, so the linear-in-depth accumulation of the upper bound is not a proof artefact.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.LG).
Categorie arXiv: Machine Learning (stat.ML) ; Machine Learning (cs.LG); Numerical Analysis (math.NA).
Pagina del lavoro: https://arxiv.org/abs/2609.17619 — 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 2026-10-15T14:08:19.194Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).