Metraipolis › archive › dc_ac669d0ea5b2
A Geometry-Based Capacity Theory for Finite-Feature Associative Memory — arXiv:2610.09056
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.09056
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T11:56:17.279Z
- SHA-256 of the archived text
b2605cf4024cfc5304ce1df897308137b71cd9feeb41758a6134411764da2d4e- Stable URL of this page
https://metraipolis.com/documento/dc_ac669d0ea5b2
How to cite
A Geometry-Based Capacity Theory for Finite-Feature Associative Memory — arXiv:2610.09056 — arXiv (en), https://arxiv.org/abs/2610.09056 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ac669d0ea5b2 (Metraipolis archive dc_ac669d0ea5b2, sha256 b2605cf4024cfc5304ce1df897308137b71cd9feeb41758a6134411764da2d4e)
BibTeX
@misc{dcac669d0ea5b2,
title = {A Geometry-Based Capacity Theory for Finite-Feature Associative Memory — arXiv:2610.09056},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.09056},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ac669d0ea5b2; sha256 b2605cf4024cfc5304ce1df897308137b71cd9feeb41758a6134411764da2d4e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09056. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We develop a geometry-based capacity theory for exact-key retrieval in compressed finite-feature Hebbian associative memory. For random or approximately isotropic values, retrieval interference separates into finite-feature noise, which decreases with feature dimension, and structural interference, which is determined by squared kernel overlap among stored keys and persists in the infinite-feature limit. This yields a fit-free prediction of retrieval quality, reveals a geometry-dependent capacity ceiling, and predicts the feature budget required for a target retrieval quality. When stored values are correlated, we show that retrieval depends jointly on the key kernel and value Gram matrix, and derive finite-feature approximations that account for this interaction. We validate the theory on synthetic, visual, and medical-image representations. Overall, the framework links representation geometry directly to memory capacity and distinguishes when performance can be improved by increasing the feature budget and when the representation itself must be changed. Across these settings, the predicted retrieval curves closely match empirical behavior and correctly identify changes in the preferred memory design. 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/2610.09056 — 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 2026-10-08T10:50:37.647Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).