Metraipolis › archive › dc_7af4463891f3
Visible Tails: Characterizing Adaptation to Unknown Moments in Heavy-Tailed Bandits — arXiv:2611.12140
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.12140
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-10T18:43:56.257Z
- SHA-256 of the archived text
5f57dbf461a4cad12e4342d9792d87d27e7337292945bc43f507b37e9be163e1- Stable URL of this page
https://metraipolis.com/documento/dc_7af4463891f3
How to cite
Visible Tails: Characterizing Adaptation to Unknown Moments in Heavy-Tailed Bandits — arXiv:2611.12140 — arXiv (en), https://arxiv.org/abs/2611.12140 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7af4463891f3 (Metraipolis archive dc_7af4463891f3, sha256 5f57dbf461a4cad12e4342d9792d87d27e7337292945bc43f507b37e9be163e1)
BibTeX
@misc{dc7af4463891f3,
title = {Visible Tails: Characterizing Adaptation to Unknown Moments in Heavy-Tailed Bandits — arXiv:2611.12140},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.12140},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7af4463891f3; sha256 5f57dbf461a4cad12e4342d9792d87d27e7337292945bc43f507b37e9be163e1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.12140. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Heavy-tailed bandit learners without moment parameters face an obstruction: the rewards that carry an arm's mean can stay unseen. We address the remaining part of a COLT open problem by characterizing the optimal-arm laws that permit adaptation to unknown moments at constant-factor cost. The condition, variance-core visibility, is one-sided and translation-invariant: the mean hidden in any rare event is at most a constant multiple of the standard error of the rewards that remain. A quantile-trimmed UCB index that knows the visibility constants attains the central-moment benchmark on this class, and no single strategy adapts across all visibility levels. The same index achieves logarithmic expected regret on every instance with a visible optimal arm and finite-mean rewards. On an explicit family with no higher moment, gap $\Delta$ and constant $A$, its logarithmic coefficient and the information lower bound are of order $\Delta^{-1}e^{A/\Delta}$, exponential in the inverse gap within one visibility class.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.12140 — 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 2026-11-10T18:37:37.679Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).