Metraipolis › archive › dc_d77c97f18ffa
Bellman-Centric Learning: Near-Optimal Regret for Linear Bandits with Memory — arXiv:2610.05659
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.05659
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T05:31:12.439Z
- SHA-256 of the archived text
cdb8cc6f676303f2c191ee084a43704c943b4f0f30c37aa78737cc8412937941- Stable URL of this page
https://metraipolis.com/documento/dc_d77c97f18ffa
How to cite
Bellman-Centric Learning: Near-Optimal Regret for Linear Bandits with Memory — arXiv:2610.05659 — arXiv, https://arxiv.org/abs/2610.05659 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d77c97f18ffa (Metraipolis archive dc_d77c97f18ffa, sha256 cdb8cc6f676303f2c191ee084a43704c943b4f0f30c37aa78737cc8412937941)
BibTeX
@misc{dcd77c97f18ffa,
title = {Bellman-Centric Learning: Near-Optimal Regret for Linear Bandits with Memory — arXiv:2610.05659},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.05659},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d77c97f18ffa; sha256 cdb8cc6f676303f2c191ee084a43704c943b4f0f30c37aa78737cc8412937941},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.05659. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We study linear bandits with memory, where past actions induce endogenous nonstationarity through an arbitrary known, bounded matrix-valued memory map. To trade off exploration and exploitation while accounting for the memory dynamics, we develop RSM-LinUCB, a Bellman-centric algorithm that learns as in linear bandits and plans as in reinforcement learning. This design admits a novel regret decomposition which separates the memory-induced error from the cumulative reward estimation error along the learner's trajectory. We prove a high-probability regret bound of $\widetilde O\big(dRS(M+1)+\sigma d\sqrt T\big)$, where $T$ is the learning horizon, $d$ is the parameter dimension, $M$ is the memory length, $R$ and $S$ bound the memory-map operator norm and reward-parameter norm, respectively, and $\sigma$ is the sub-Gaussian noise scale. Our results reveal that the multiplicative memory-horizon coupling in prior bounds is not intrinsic: memory only contributes an additive cost, up to logarithmic factors. We also prove a matching minimax lower bound, establishing near-optimality. We further extend the algorithm to generalized linear rewards, preserving this separation with near-optimal memory and leading statistical dependence. Our algorithms outperform the baselines in numerical experiments on synthetic instances and semi-synthetic KV- and semantic-cache tasks. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.05659 — 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 2026-10-06T05:27:05.815Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).