跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.25424 · v1

Coordinate Basis Linear Bandits with Sublinear Regret over Low-Rank Clusters

首次提交 2026-11-19 11:03
版本更新 2026-11-19 11:03
雷达观测 2026-11-20 10:50
arXiv ↗PDF ↗

摘要

We study the problem of computational efficiency for linear bandits in high-dimensional settings with a finite arm set. In linear bandits, the increase in the dimension dd of the feature vectors leads to growing computational costs of O(d2)O(d^2) at each round of update. Traditional low-rank approximation methods such as SOFUL reduce computation via fixed-size matrix sketching, yet run the risk of incurring vacuous linear regret when the spectral tail of the data is heavy and the approximation rank is inadequately selected. To guarantee regret convergence and effectively reduce computational costs, we introduce a clustering mechanism and propose the Coordinate Basis Linear Bandit (CBL) algorithm. Our method preserves the full covariance information in each cluster by maintaining an exact orthonormal basis of the cluster's column space together with a low-dimensional coordinate representation, which guarantees robust sublinear regret without spectral-tail vulnerabilities. It performs cluster switching by assigning a sentinel for each cluster, and reduces per-round update computation to O(rˉ2)O(\bar r^2) in the cluster's effective dimension rˉ=min⁡(r,l)\bar r=\min(r,l), where rr is the intrinsic rank of the data and l<dl<d is a tunable rank bound.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-19 11:03 · 最新观测版本

Coordinate Basis Linear Bandits with Sublinear Regret over Low-Rank Clusters

We study the problem of computational efficiency for linear bandits in high-dimensional settings with a finite arm set. In linear bandits, the increase in the dimension dd of the feature vectors leads to growing computational costs of O(d2)O(d^2) at each round of update. Traditional low-rank approximation methods such as SOFUL reduce computation via fixed-size matrix sketching, yet run the risk of incurring vacuous linear regret when the spectral tail of the data is heavy and the approximation rank is inadequately selected. To guarantee regret convergence and effectively reduce computational costs, we introduce a clustering mechanism and propose the Coordinate Basis Linear Bandit (CBL) algorithm. Our method preserves the full covariance information in each cluster by maintaining an exact orthonormal basis of the cluster's column space together with a low-dimensional coordinate representation, which guarantees robust sublinear regret without spectral-tail vulnerabilities. It performs cluster switching by assigning a sentinel for each cluster, and reduces per-round update computation to O(rˉ2)O(\bar r^2) in the cluster's effective dimension rˉ=min⁡(r,l)\bar r=\min(r,l), where rr is the intrinsic rank of the data and l<dl<d is a tunable rank bound.

查看此版本解读 →查看此版本原文 ↗