arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.09379stat.ME
收藏

通过多数投票合并不确定性集合

Merging uncertainty sets via majority vote

Matteo Gasparin, Aaditya Ramdas

AI总结:

针对如何黑盒合并 $K$ 个任意依赖的不确定性集合,提出一种多数投票程序,保留误差保证并支持加权平均与随机化优化,且能有效对 split conformal prediction 等方法去随机化。

摘要:

给定 $K$ 个任意依赖的不确定性集合——例如,使用 $K$ 个不同估计量获得的未知参数的置信区间,或基于 $K$ 个不同算法在共享数据上通过 conformal prediction 获得的预测集——我们解决如何以黑盒方式高效合并它们以产生单一不确定性集合的问题。我们提出了一种简单且广泛适用的多数投票程序,产生的合并集具有与输入集几乎相同的误差保证。随后我们从几个方面扩展了这一核心思想:我们证明加权平均是融入先验信息的强大方法,且简单的随机化技巧能在不改变覆盖保证的情况下产生严格更小的合并集。如果集合是可交换的,则可获得进一步改进。我们还证明,split conformal prediction、median of means、HulC 和交叉拟合的“double machine learning”等现代方法均可利用这些思想进行有效的去随机化。

英文摘要:

Given $K$ uncertainty sets that are arbitrarily dependent -- for example, confidence intervals for an unknown parameter obtained with $K$ different estimators, or prediction sets obtained via conformal prediction based on $K$ different algorithms on shared data -- we address the question of how to efficiently combine them in a black-box manner to produce a single uncertainty set. We present a simple and broadly applicable majority vote procedure that produces a merged set with nearly the same error guarantee as the input sets. We then extend this core idea in a few ways: we show that weighted averaging can be a powerful way to incorporate prior information, and a simple randomization trick produces strictly smaller merged sets without altering the coverage guarantee. Further improvements can be obtained if the sets are exchangeable. We also show that many modern methods, like split conformal prediction, median of means, HulC and cross-fitted ``double machine learning'', can be effectively derandomized using these ideas.

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。

↑