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 2304.09424cs.LGcs.AIstat.ML
收藏

损失最小化可为大型神经网络带来多重校准

Loss Minimization Yields Multicalibration for Large Neural Networks

ETH Zürich(苏黎世联邦理工学院);Apple(苹果公司);Stanford University(斯坦福大学);Microsoft Research(微软研究院)

Jarosław Błasiok, Parikshit Gopalan, Lunjia Hu, Adam Tauman Kalai, Preetum Nakkiran

AI总结:

本文研究受保护群体由较小神经网络表示、预测器为较大神经网络时的公平性问题,证明平方损失最小化除有限不幸规模外可实现多重校准。

摘要:

多重校准是预测器的一种公平性概念,要求预测器在大量受保护群体上都能提供经过校准的预测。众所周知,即使对于线性函数等简单预测器,多重校准也是一个不同于损失最小化的目标。在本工作中,我们考虑这样一种设定:受保护群体可由规模为$k$的神经网络表示,而预测器是规模为$n>k$的神经网络。我们证明,在所有规模为$n$的神经网络上最小化平方损失,会在除有限数量的不幸$n$值之外的所有情况下带来多重校准。我们还给出证据表明,在我们的证明技术下,关于不幸值数量的界是紧的。此前,损失最小化带来多重校准这类结果仅适用于接近真实情况的预测器,因此适用性相当有限。与这些结果不同,我们的结果依赖神经网络的表达能力,并利用了预测器的表示。

英文摘要:

Multicalibration is a notion of fairness for predictors that requires them to provide calibrated predictions across a large set of protected groups. Multicalibration is known to be a distinct goal than loss minimization, even for simple predictors such as linear functions. In this work, we consider the setting where the protected groups can be represented by neural networks of size $k$, and the predictors are neural networks of size $n > k$. We show that minimizing the squared loss over all neural nets of size $n$ implies multicalibration for all but a bounded number of unlucky values of $n$. We also give evidence that our bound on the number of unlucky values is tight, given our proof technique. Previously, results of the flavor that loss minimization yields multicalibration were known only for predictors that were near the ground truth, hence were rather limited in applicability. Unlike these, our results rely on the expressivity of neural nets and utilize the representation of the predictor.

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。

↑