Metraipolis › archive › dc_49899072cb29
Measuring Memory and Generalization as Separable Geometric Channels: The Topo^2 Framework — arXiv:2608.30487
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.30487
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T07:52:44.070Z
- SHA-256 of the archived text
9486b1e9cfc7c6d96492f837cc5f9b36899710bd105e858ced2d459dae3ab82b- Stable URL of this page
https://metraipolis.com/documento/dc_49899072cb29
How to cite
Measuring Memory and Generalization as Separable Geometric Channels: The Topo^2 Framework — arXiv:2608.30487 — arXiv (en), https://arxiv.org/abs/2608.30487 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_49899072cb29 (Metraipolis archive dc_49899072cb29, sha256 9486b1e9cfc7c6d96492f837cc5f9b36899710bd105e858ced2d459dae3ab82b)
BibTeX
@misc{dc49899072cb29,
title = {Measuring Memory and Generalization as Separable Geometric Channels: The Topo^2 Framework — arXiv:2608.30487},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.30487},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_49899072cb29; sha256 9486b1e9cfc7c6d96492f837cc5f9b36899710bd105e858ced2d459dae3ab82b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30487. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Deep networks trained on noisy labels simultaneously generalize on clean data and memorize flipped labels. These are usually conflated as pressures on one capacity. We present Topo^2, a measurement framework that makes them causally separable, measurable, and law-governed. Persistent-homology H1 structure of the representation space separates into a within-class manifold channel (a function of the training stopping point) and a cross-class channel (a monotone readout of memorized flipped samples). An intervention, the FM0 prescription (zero loss on flipped samples from epoch 0), reaches each setting's generalization ceiling while memorizing essentially nothing. Within the framework we establish a law set with graded evidence: (L2) FM0 separation prescription (9/9); (L1) the within-channel as a training-position function (mid-rise 6/6; convergence-back CIFAR 3/3, SVHN 2/3); (L3) a ring-construction identity (definitional, not a law); and TLS (memory-generalization topological layering): memory is causally additive, anchored (silencing clean collapses the representation), invertible (stripping memory restores near-ceiling generalization), and quantitatively billable (the memorization cost law, effective slope coefficient C ~ 0.38 at the reference capacity: CIFAR-10 0.3801 / SVHN 0.3806 / CIFAR-100 0.384 / VGG 0.3715, capacity-dependent in general and traced to clean-sample feature displacement). We also publish the framework's boundaries: a falsification ledger of nine dead ends, and an instrument-vindication section that excludes six families of global statistics as explanations of the within-channel. The framework turns "memorization" from an ill-defined capacity into a measurable, separable, invertible topological layer. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.30487 — 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 2026-10-17T07:28:26.875Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).