Metraipolis › archive › dc_035a54928f80
An Accuracy--Information Tradeoff for Loss-Difference Conditional Mutual Information — arXiv:2610.09206
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.09206
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T11:56:30.540Z
- SHA-256 of the archived text
52e97de0a398d0fc61e2cc0608c41afd0dd5803231912da8c2f7599817cedbb9- Stable URL of this page
https://metraipolis.com/documento/dc_035a54928f80
How to cite
An Accuracy--Information Tradeoff for Loss-Difference Conditional Mutual Information — arXiv:2610.09206 — arXiv (en), https://arxiv.org/abs/2610.09206 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_035a54928f80 (Metraipolis archive dc_035a54928f80, sha256 52e97de0a398d0fc61e2cc0608c41afd0dd5803231912da8c2f7599817cedbb9)
BibTeX
@misc{dc035a54928f80,
title = {An Accuracy--Information Tradeoff for Loss-Difference Conditional Mutual Information — arXiv:2610.09206},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.09206},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_035a54928f80; sha256 52e97de0a398d0fc61e2cc0608c41afd0dd5803231912da8c2f7599817cedbb9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09206. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Loss-difference conditional mutual information (ld-CMI) uses the smallest of the standard observations in the supersample hierarchy of generalization bounds: it measures what a learner's loss differences reveal about which candidate of each pair it was trained on. Accuracy is known to force information into the model; data processing does not carry such lower bounds to losses. We show, by bounding three moments of the loss differences, that accuracy also forces ld-CMI. For linear predictors with a smooth convex loss of nonzero slope at zero, such as the logistic loss, plus a regularizer whose curvature and growth are both of power $r\ge2$, on product distributions over a scaled sign cube in dimension at least linear in $n$, every proper learner with expected excess risk at most $\varepsilon$ on these distributions at the optimal sample size $n\asymp\varepsilon^{-2+2/r}$ has worst-case ld-CMI of order $n$ bits, and $\Theta(n/(1+(\tau/\varepsilon)^2))$ bits under Gaussian noise of standard deviation $\tau$ on the loss differences. The same holds without a regularizer, at $n\asymp\varepsilon^{-2}$. Consequently, range-scaled ld-CMI bounds cannot vanish on these distributions, although every proper learner's generalization gap is $O(n^{-1/2})$. We also show that model-level information does not determine noisy loss-difference information, and that the growth, slope and dimension conditions are needed, the last up to a logarithm.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Information Theory (cs.IT); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.09206 — 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 2026-10-08T10:50:37.669Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).