Metraipolis › archive › dc_2840344d727d

Does LinUCB Need Valid Confidence Sets? — arXiv:2611.04641

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04641
License (as archived)
CC0-1.0
Archived
2026-11-04T21:40:07.395Z
SHA-256 of the archived text
45def5af207ede76a25e25151632d051a8801687c534f045859009797a8be9bd
Stable URL of this page
https://metraipolis.com/documento/dc_2840344d727d

How to cite

Does LinUCB Need Valid Confidence Sets? — arXiv:2611.04641 — arXiv (en), https://arxiv.org/abs/2611.04641 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2840344d727d (Metraipolis archive dc_2840344d727d, sha256 45def5af207ede76a25e25151632d051a8801687c534f045859009797a8be9bd)
BibTeX
@misc{dc2840344d727d,
  title = {Does LinUCB Need Valid Confidence Sets? — arXiv:2611.04641},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04641},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2840344d727d; sha256 45def5af207ede76a25e25151632d051a8801687c534f045859009797a8be9bd},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04641. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We show that replacing LinUCB's usual confidence width $\beta$ by any fixed $\alpha>0$ preserves sublinear regret when the action set is $M$-smooth and $m$-strongly convex. Our high-probability regret bound for this generalisation, $\alpha$-LinUCB, establishes that valid confidence sets are not necessary for the success of LinUCB. We identify the mechanism behind this result: action-set curvature drives $\alpha$-LinUCB to explore every direction, even when optimism is not guaranteed. To analyse this mechanism, we introduce a logarithmically capped spectral potential that, unlike the standard log-determinant potential, caps contributions from well-explored directions. This yields lower bounds on the smallest eigenvalue of the design matrix. At the standard confidence width $\alpha=\beta\asymp\sqrt{d\log T}$, our $T$-round regret bound in $d$ dimensions is $O\bigl(\frac{M}{m}d\sqrt{T\log T}\bigr)$, up to iterated-logarithmic factors, improving the horizon dependence of the classical LinUCB guarantee by a factor of $\sqrt{\log T}$.

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/2611.04641 — 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 2026-11-04T21:35:49.679Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).