Metraipolis › archive › dc_0aaf09bb4f28

The Tilted Sampling Problem: Optimism, Posterior Sampling, and Logarithmic Regret — arXiv:2610.21908

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21908
License (as archived)
CC0-1.0
Archived
2026-10-16T10:24:42.499Z
SHA-256 of the archived text
634cf8ff0113507c6a992f29a04f2ef721b75695dbf66d1e3b854fcad1495a17
Stable URL of this page
https://metraipolis.com/documento/dc_0aaf09bb4f28

How to cite

The Tilted Sampling Problem: Optimism, Posterior Sampling, and Logarithmic Regret — arXiv:2610.21908 — arXiv (en), https://arxiv.org/abs/2610.21908 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0aaf09bb4f28 (Metraipolis archive dc_0aaf09bb4f28, sha256 634cf8ff0113507c6a992f29a04f2ef721b75695dbf66d1e3b854fcad1495a17)
BibTeX
@misc{dc0aaf09bb4f28,
  title = {The Tilted Sampling Problem: Optimism, Posterior Sampling, and Logarithmic Regret — arXiv:2610.21908},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21908},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0aaf09bb4f28; sha256 634cf8ff0113507c6a992f29a04f2ef721b75695dbf66d1e3b854fcad1495a17},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21908. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Many modern learning systems are used not to select a single best action, but to sample diverse high-quality outputs. Examples include entropy-regularized reinforcement learning, RLHF, and reward-guided generative modeling. Motivated by these settings, we study tilted sampling: an online learning problem in which the goal is to sample from the Gibbs distribution induced by an unknown reward function. We define sampling regret as the temperature-normalized cumulative KL divergence between the learner's distribution and the Gibbs target, corresponding to the suboptimality gap in entropy-regularized reward maximization. Our framework generalizes recent sampling-based and KL-regularized formulations in bandits and reinforcement learning. Our main contribution is a general regret framework that captures the different uncertainty structure of tilted sampling and classical regret minimization. While standard regret scales as $\widetilde{O}(\sqrt{T})$ due to first-order uncertainty, sampling regret is governed by squared confidence widths, yielding logarithmic-in-$T$ regret in standard settings at fixed temperature. We develop sampling analogues of UCB and Thompson sampling (UCS and TS$^2$) and provide unified regret decompositions that explain and extend recent logarithmic guarantees, including the first Bayesian sampling-regret decomposition based on new bounds for posterior-random Gibbs distributions. Our results show that learning to sample from a soft-optimal distribution is statistically distinct from learning to identify a hard optimum.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Optimization and Control (math.OC); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.21908 — 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 2026-10-16T08:48:02.768Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).