Metraipolis › archive › dc_f9276fabcf2c

Coupled Scaling: A Representational Accessibility Framework for Neural Scaling Laws — arXiv:2609.03533

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.03533
License (as archived)
CC0-1.0
Archived
2026-10-17T02:07:59.486Z
SHA-256 of the archived text
f8677cfa80fac6ee669f8b0a05bb4e6965c3ea1c33114fca1ca51fddd31fffca
Stable URL of this page
https://metraipolis.com/documento/dc_f9276fabcf2c

How to cite

Coupled Scaling: A Representational Accessibility Framework for Neural Scaling Laws — arXiv:2609.03533 — arXiv (en), https://arxiv.org/abs/2609.03533 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f9276fabcf2c (Metraipolis archive dc_f9276fabcf2c, sha256 f8677cfa80fac6ee669f8b0a05bb4e6965c3ea1c33114fca1ca51fddd31fffca)
BibTeX
@misc{dcf9276fabcf2c,
  title = {Coupled Scaling: A Representational Accessibility Framework for Neural Scaling Laws — arXiv:2609.03533},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.03533},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f9276fabcf2c; sha256 f8677cfa80fac6ee669f8b0a05bb4e6965c3ea1c33114fca1ca51fddd31fffca},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.03533. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We ask when two learning systems trained on the same task under a common resource protocol should share a scaling rate and when their rates should differ. Coupled Scaling answers this through representational accessibility: the task-relevant geometry that a specified architecture-optimization system can reach and how that geometry is acquired as resources grow. In an orthogonal model, unsupported target energy sets the asymptotic floor, while unacquired supported energy sets the finite-budget residual. When acquisition can skip high-value directions, the largest fully acquired prefix no longer determines a unique exponent. For target powers $a_j \asymp j^{-b}$, $b>1$, and prefix log-growth rate $0<\rho\leq1$, the sharp attainable interval is $[\rho(b-1),\min\{b-1,\rho b\}]$, and every rate in this interval is realized by a fixed acquisition order on a common support. A rank-window condition identifies when the product law $\alpha=\rho(b-1)$ applies, while a fixed-kernel specialization expresses the same task-system dependence through the task-weighted spectral tail. We then audit released capability trajectories to diagnose measurement and comparison effects and propose a staged test in which independently measured geometry predicts held-out loss, scaling rates, and cross-task reversals.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.03533 — 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 2026-10-17T01:47:25.556Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).