arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.07055cs.LGstat.ML
收藏

具有长期记忆的在线多任务学习

Online Multitask Learning with Long-Term Memory

University College London(伦敦大学学院)

Mark Herbster, Stephen Pasteris, Lisa Tse

摘要:

我们介绍一种新颖的在线多任务设置。在这种设置中,每个任务都被划分为一系列学习者未知的片段。与每个段相关联的是来自某个假设类的假设。我们给出的算法,旨在利用的情况下,有许多这样的部分,但显着较少的相关假设。我们证明遗憾的界限,持有任何分割的任务和任何关联的假设的部分。在单任务设置中,这相当于在[Bousquet and Warfall; 2003]的意义上切换长期记忆。我们提供了一种算法,当假设类有限时,该算法在每次试验中的预测时间与假设数量呈线性关系。我们还考虑无限假设类从再生核希尔伯特空间,我们给一个算法,其每次试验的时间复杂度是立方的累积试验的数量。在单任务的特殊情况下,这是一个有效的遗憾有界的切换算法与长期记忆的非参数假设类的第一个例子。

英文摘要:

We introduce a novel online multitask setting. In this setting each task is partitioned into a sequence of segments that is unknown to the learner. Associated with each segment is a hypothesis from some hypothesis class. We give algorithms that are designed to exploit the scenario where there are many such segments but significantly fewer associated hypotheses. We prove regret bounds that hold for any segmentation of the tasks and any association of hypotheses to the segments. In the single-task setting this is equivalent to switching with long-term memory in the sense of [Bousquet and Warmuth; 2003]. We provide an algorithm that predicts on each trial in time linear in the number of hypotheses when the hypothesis class is finite. We also consider infinite hypothesis classes from reproducing kernel Hilbert spaces for which we give an algorithm whose per trial time complexity is cubic in the number of cumulative trials. In the single-task special case this is the first example of an efficient regret-bounded switching algorithm with long-term memory for a non-parametric hypothesis class.

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。

↑