Metraipolis › archive › dc_cc400e91bcc5

The Rank the Task Demands: A Causal Rank Law for Matrix Memories Trained on Group Composition — arXiv:2609.12259

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.12259
License (as archived)
CC0-1.0
Archived
2026-10-16T03:15:37.737Z
SHA-256 of the archived text
77a85336ab2d688928ab1e0592b634d354f9c6046058aeeef58e658ae2407ed4
Stable URL of this page
https://metraipolis.com/documento/dc_cc400e91bcc5

How to cite

The Rank the Task Demands: A Causal Rank Law for Matrix Memories Trained on Group Composition — arXiv:2609.12259 — arXiv (en), https://arxiv.org/abs/2609.12259 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cc400e91bcc5 (Metraipolis archive dc_cc400e91bcc5, sha256 77a85336ab2d688928ab1e0592b634d354f9c6046058aeeef58e658ae2407ed4)
BibTeX
@misc{dccc400e91bcc5,
  title = {The Rank the Task Demands: A Causal Rank Law for Matrix Memories Trained on Group Composition — arXiv:2609.12259},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.12259},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cc400e91bcc5; sha256 77a85336ab2d688928ab1e0592b634d354f9c6046058aeeef58e658ae2407ed4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.12259. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Matrix-valued memories make rank the natural budget of a learned representation: the number of independent directions a state spans bounds what it can bind, compose, and track. We report causal evidence, on a group-composition testbed trained under a hard single-state bottleneck with a fixed decoder that cannot launder rank, that gradient descent recruits precisely the rank the task's algebra demands. A companion paper [Larson, 2026a] establishes the analogous recruitment and causal necessity pattern on a $K$-pair associative-binding testbed, where exact recovery provably requires state rank at least $K$; this paper inherits that instrument and extends the rank law from a scalar capacity bound to a representation-theoretic one. We train toward chosen minimal faithful reference representations embedded in larger matrices. On group-composition state tracking over five finite groups spanning the solvable/non-solvable divide, the recruited rank equals the group's minimal faithful real representation dimension $d_{\min}$ (Spearman $\rho = 0.9747$, the design's tie-capped maximum), the dimension-matched solvable/non-solvable pair $S_4$/$A_5$ is statistically equivalent under a pre-registered test, and a pre-registered force-rank test separates a guaranteed similarity ceiling from empirical recovery at the target dimension: one rank below $d_{\min}$, cosine similarity is capped by the target's tied unit spectrum at $\sqrt{(d_{\min}{-}1)/d_{\min}} \le 0.894$, below the $0.9$ threshold in every group by construction, with observed cells at 86-95% (mean 91%) of that ceiling; at $d_{\min}$, not guaranteed a priori, recovery clears the pre-registered anchor-relative bar at four seeds per group in all five groups. Within this testbed, measured effective rank tracks representation dimension; the matched-dimension $S_4$/$A_5$ comparison establishes equivalence within the pre-registered tolerance.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-14 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.12259 — 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 2026-10-16T03:05:44.028Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).