Metraipolis › archive › dc_d2758d02207a

Reserve-Aware Contrast Certificates for Conservative Bandits with Uncertain Baselines — arXiv:2609.39106

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.39106
License (as archived)
CC0-1.0
Archived
2026-10-07T12:42:28.992Z
SHA-256 of the archived text
b1981e031b4477c3c04d764270e4a5024bef6705b71d6bd0559ff7c74b3660cf
Stable URL of this page
https://metraipolis.com/documento/dc_d2758d02207a

How to cite

Reserve-Aware Contrast Certificates for Conservative Bandits with Uncertain Baselines — arXiv:2609.39106 — arXiv (en), https://arxiv.org/abs/2609.39106 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d2758d02207a (Metraipolis archive dc_d2758d02207a, sha256 b1981e031b4477c3c04d764270e4a5024bef6705b71d6bd0559ff7c74b3660cf)
BibTeX
@misc{dcd2758d02207a,
  title = {Reserve-Aware Contrast Certificates for Conservative Bandits with Uncertain Baselines — arXiv:2609.39106},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.39106},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d2758d02207a; sha256 b1981e031b4477c3c04d764270e4a5024bef6705b71d6bd0559ff7c74b3660cf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39106. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Conservative bandits must improve an incumbent policy without exhausting a prescribed performance budget. When the incumbent is uncertain, separately bounding candidate and baseline rewards can charge twice for shared estimation error. We develop Reserve-C4B around the baseline-relative contrast itself. A shared confidence set yields an exact expression for this avoidable penalty and a tighter admissibility test at every fixed history. A reserve ledger separates statistical evidence from permitted performance deficit; a prefix-refresh extension recertifies accumulated decisions under the current confidence set without discarding previously certified credit. For linear rewards, self-normalized confidence sets provide simultaneous validity over time and adaptively generated candidates, and the resulting policy satisfies a conditional-mean performance constraint with high probability. Reproducible experiments isolate certificate coupling, prefix refresh, and historical information, showing large reductions in baseline fallback while exposing the limitations of frozen certificates.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.39106 — 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 2026-10-07T12:14:13.731Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).