arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.02240cs.LGcs.DM
收藏

可复制学习中的列表与证书复杂度

List and Certificate Complexities in Replicable Learning

Ben-Gurion University of Negev(内盖夫本-古里安大学);Iowa State University(爱荷华州立大学);University of Nebraska-Lincoln(内布拉斯加大学林肯分校)

Peter Dixon, A. Pavan, Jason Vander Woude, N. V. Vinodchandran

AI总结:

本文研究了可复制学习算法,针对强可复制性不可实现的问题,提出了列表可复制性和证书可复制性两种可行概念,并设计了在列表和证书复杂度上最优的学习算法,同时建立了匹配的不可能性结果。

摘要:

我们研究可复制学习算法。理想情况下,我们希望设计出在多次运行中输出相同规范模型的算法,即使不同运行观察到来自未知数据分布的不同样本集。通常,这种强可复制性概念是无法实现的。因此,我们考虑两种可行的可复制性概念,称为列表可复制性和证书可复制性。直观上,这些概念捕捉了(非)可复制性的程度。我们为某些学习问题设计了在列表和证书复杂度上最优的算法。我们建立了匹配的不可能性结果。

英文摘要:

We investigate replicable learning algorithms. Ideally, we would like to design algorithms that output the same canonical model over multiple runs, even when different runs observe a different set of samples from the unknown data distribution. In general, such a strong notion of replicability is not achievable. Thus we consider two feasible notions of replicability called list replicability and certificate replicability. Intuitively, these notions capture the degree of (non) replicability. We design algorithms for certain learning problems that are optimal in list and certificate complexity. We establish matching impossibility results.

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。

↑