arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.06326cs.LG
收藏

随机线性Bands中的多环境元学习

Multi-Environment Meta-Learning in Stochastic Linear Bandits

University of California, Santa Barbara(加州大学圣巴巴拉分校);Google Research(谷歌研究院);Stanford University(斯坦福大学);University of British Columbia(不列颠哥伦比亚大学)

Ahmadreza Moradipari, Mohammad Ghavamzadeh, Taha Rajabzadeh, Christos Thrampoulidis, Mahnoosh Alizadeh

摘要:

在这项工作中,我们研究元学习(或学习学习)的方法在多任务线性随机强盗问题,可以起源于多个环境。受[1]关于线性强盗问题序列中的元学习的工作的启发,该线性强盗问题的参数是从单个分布中采样的(即,单个环境),这里我们考虑当任务参数从混合分布中提取时元学习的可行性。对于这个问题,我们提出了一个正则化版本的OFUL算法,当训练的任务与标记的环境,实现低遗憾的新任务,而不需要的环境知识,新任务的起源。具体来说,我们对新算法的遗憾范围捕获了环境误分类的影响,并强调了单独学习每个任务或在不识别不同混合成分的情况下进行元学习的好处。

英文摘要:

In this work we investigate meta-learning (or learning-to-learn) approaches in multi-task linear stochastic bandit problems that can originate from multiple environments. Inspired by the work of [1] on meta-learning in a sequence of linear bandit problems whose parameters are sampled from a single distribution (i.e., a single environment), here we consider the feasibility of meta-learning when task parameters are drawn from a mixture distribution instead. For this problem, we propose a regularized version of the OFUL algorithm that, when trained on tasks with labeled environments, achieves low regret on a new task without requiring knowledge of the environment from which the new task originates. Specifically, our regret bound for the new algorithm captures the effect of environment misclassification and highlights the benefits over learning each task separately or meta-learning without recognition of the distinct mixture components.

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。

↑