arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2404.09413stat.MLcs.CRcs.LG
收藏

关于局部隐私线性上下文老虎机的最优遗憾

On the Optimal Regret of Locally Private Linear Contextual Bandit

Massachusetts Institute of Technology(麻省理工学院);University of Texas at Dallas(德克萨斯大学达拉斯分校)

Jiachun Li, David Simchi-Levi, Yining Wang

AI总结:

针对局部隐私约束下线性上下文老虎机能否达到经典$\tilde O(\sqrt{T})$遗憾上界的开放问题,提出结合平均绝对偏差误差分析与分层主成分回归的方案,首次实现该最优遗憾上界。

摘要:

具有线性奖励函数的上下文老虎机是老虎机与在线学习研究中被最广泛研究的模型之一。近年来,设计局部隐私(locally private)线性上下文老虎机算法的研究热度不断上升,这类算法可保护上下文与奖励中包含的敏感信息,防止其泄露给公众。经典线性上下文老虎机算法可通过多种替代方法达到$\tilde O(\sqrt{T})$的累积遗憾上界,但在存在局部隐私约束的情况下能否实现这类遗憾上界仍是未解难题,目前的最优结果为$\tilde O(T^{3/4})$。本文证明,局部隐私线性上下文老虎机确实可以实现$\tilde O(\\\\\/sqrt{T})$的遗憾上界。我们的方案依托若干全新的算法与分析思路,例如对平均绝对偏差误差的分析,以及为实现较小平均绝对偏差误差而采用的分层主成分回归方法。

英文摘要:

Contextual bandit with linear reward functions is among one of the most extensively studied models in bandit and online learning research. Recently, there has been increasing interest in designing \emph{locally private} linear contextual bandit algorithms, where sensitive information contained in contexts and rewards is protected against leakage to the general public. While the classical linear contextual bandit algorithm admits cumulative regret upper bounds of $\tilde O(\sqrt{T})$ via multiple alternative methods, it has remained open whether such regret bounds are attainable in the presence of local privacy constraints, with the state-of-the-art result being $\tilde O(T^{3/4})$. In this paper, we show that it is indeed possible to achieve an $\tilde O(\sqrt{T})$ regret upper bound for locally private linear contextual bandit. Our solution relies on several new algorithmic and analytical ideas, such as the analysis of mean absolute deviation errors and layered principal component regression in order to achieve small mean absolute deviation errors.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑