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 2203.05067cs.LGstat.ML
收藏

对抗响应下的通用回归

Universal Regression with Adversarial Responses

MIT(麻省理工学院)

Moïse Blanchard, Patrick Jaillet

AI总结:

该研究针对一般可分度量空间下大类非独立同分布实例序列的对抗响应回归问题,提出最少假设下的算法与可学习性刻画,给出乐观通用学习规则及无界损失下的可积条件,还得到无矩约束的对抗均值估计学习规则。

摘要:

我们针对一般可分度量空间下的大类非独立同分布实例序列,提出了对抗响应回归问题的算法,且仅采用可证明的最少假设。我们还给出了该回归场景下可学习性的刻画。我们研究了通用一致性,即要求学习器在不对值响应施加限制的条件下仍具备强一致性。分析表明,相较于平稳过程,该目标可在范围大得多的实例序列类别上实现,同时揭示了值空间的根本二分性:有限时域均值估计是否可实现。我们进一步提出了乐观通用学习规则,即若该规则无法实现通用一致性,则其他任何算法也同样无法实现。针对无界损失,我们提出了一个温和的可积性条件,在该条件下,大类非独立同分布实例序列上的对抗回归问题存在可行算法。此外,我们的分析还给出了一种适用于一般度量空间的均值估计学习规则,它在对抗响应下具备一致性,且对序列无任何矩条件约束,这一结果本身具有独立研究价值。

英文摘要:

We provide algorithms for regression with adversarial responses under large classes of non-i.i.d. instance sequences, on general separable metric spaces, with provably minimal assumptions. We also give characterizations of learnability in this regression context. We consider universal consistency which asks for strong consistency of a learner without restrictions on the value responses. Our analysis shows that such an objective is achievable for a significantly larger class of instance sequences than stationary processes, and unveils a fundamental dichotomy between value spaces: whether finite-horizon mean estimation is achievable or not. We further provide optimistically universal learning rules, i.e., such that if they fail to achieve universal consistency, any other algorithms will fail as well. For unbounded losses, we propose a mild integrability condition under which there exist algorithms for adversarial regression under large classes of non-i.i.d. instance sequences. In addition, our analysis also provides a learning rule for mean estimation in general metric spaces that is consistent under adversarial responses without any moment conditions on the sequence, a result of independent interest.

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。

↑