Metraipolis › archive › dc_7184b5b933fa

Symmetry without a manifold: intrinsic dimension on orbits — arXiv:2609.17926

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.17926
License (as archived)
CC0-1.0
Archived
2026-10-15T15:04:30.760Z
SHA-256 of the archived text
633258d0c35833f8448e839f46765586af7ca4025e5c1d908857cc5099dfc1ba
Stable URL of this page
https://metraipolis.com/documento/dc_7184b5b933fa

How to cite

Symmetry without a manifold: intrinsic dimension on orbits — arXiv:2609.17926 — arXiv (en), https://arxiv.org/abs/2609.17926 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7184b5b933fa (Metraipolis archive dc_7184b5b933fa, sha256 633258d0c35833f8448e839f46765586af7ca4025e5c1d908857cc5099dfc1ba)
BibTeX
@misc{dc7184b5b933fa,
  title = {Symmetry without a manifold: intrinsic dimension on orbits — arXiv:2609.17926},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.17926},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7184b5b933fa; sha256 633258d0c35833f8448e839f46765586af7ca4025e5c1d908857cc5099dfc1ba},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.17926. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

The standard geometric derivation of neural scaling exponents takes the intrinsic dimension of a data manifold as its input. On modular addition in $\mathbb{Z}_p$ that derivation has no input. The exact algebraic solution is an orbit of $\mathbb{Z}_p$ acting by isometries. Transitivity alone makes the ratio statistic underlying the standard dimension estimator a point mass, so the estimator is undefined, and here the two nearest neighbour distances coincide exactly. Breaking the symmetry at scale $\epsilon$ returns a number, but one that tracks $1/\epsilon$ with no scale free plateau. We show that the failure is general, since on any finite orbit of a group acting by isometries the estimator reports the resolution at which the set is probed rather than a dimension. What replaces the power law is exponential in hidden width, $L(h)=L_\infty+A\exp(-c\,h^{\alpha})$, with $R^2$ between 0.982 and 0.995 against 0.857 to 0.906 for a power law admitting the same floor and fitted under the same protocol. Where the data supply is sufficient the rate belongs to the regulariser rather than to the group, since weight decay moves $c$ by a factor of 47 while group order moves it by 1.10, a residual below seed to seed resolution, for every fixed $\alpha$ between 0.75 and 2. The critical width falls with group order rather than rising, against capacity counting that assigns a fixed number of neurons to each irreducible representation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2609.17926 — 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 2026-10-15T14:08:19.085Z 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).