arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2111.04746cs.LGstat.ML
收藏

可实现学习就是你所需要的一切

Realizable Learning is All You Need

Yale University(耶鲁大学)

Max Hopkins, Daniel M. Kane, Shachar Lovett, Gaurav Mahajan

AI总结:

本文提出首个模型无关的三行黑盒归约框架,统一证明可实现与不可知可学习性等价,并将其推广为学习算法性质从有限假设类到可学习假设类的泛化现象。

摘要:

可实现可学习性与不可知可学习性的等价性是学习理论中的一个基本现象。其变体涵盖从PAC学习和回归等经典设定,到对抗鲁棒学习等近期趋势;令人惊讶的是,我们仍缺乏统一理论:传统的等价性证明往往彼此分散,并依赖于强的、特定于模型的假设,例如一致收敛和样本压缩。在本文中,我们给出了第一个独立于模型的框架,用于解释可实现可学习性与不可知可学习性的等价性:这是一个三行的黑盒归约,简化、统一并扩展了我们在多种设定下的理解。这包括尚无可学习性刻画的模型,例如在任意分布假设和更一般损失函数下的学习,以及许多其他流行设定,如鲁棒学习、部分学习、公平学习和统计查询模型。更一般地,我们主张可实现学习与不可知学习的等价性实际上是一个更广泛现象的特例,我们称之为性质泛化:学习算法的任何理想性质(例如噪声容忍性、隐私性、稳定性),若能在有限假设类上得到满足,就会(可能以某种变体)扩展到任何可学习的假设类。

英文摘要:

The equivalence of realizable and agnostic learnability is a fundamental phenomenon in learning theory. With variants ranging from classical settings like PAC learning and regression to recent trends such as adversarially robust learning, it's surprising that we still lack a unified theory; traditional proofs of the equivalence tend to be disparate, and rely on strong model-specific assumptions like uniform convergence and sample compression. In this work, we give the first model-independent framework explaining the equivalence of realizable and agnostic learnability: a three-line blackbox reduction that simplifies, unifies, and extends our understanding across a wide variety of settings. This includes models with no known characterization of learnability such as learning with arbitrary distributional assumptions and more general loss functions, as well as a host of other popular settings such as robust learning, partial learning, fair learning, and the statistical query model. More generally, we argue that the equivalence of realizable and agnostic learning is actually a special case of a broader phenomenon we call property generalization: any desirable property of a learning algorithm (e.g. noise tolerance, privacy, stability) that can be satisfied over finite hypothesis classes extends (possibly in some variation) to any learnable 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。

↑