发表机构
Stanford University; Apple; University of Oxford; Institute for Advanced Study; Harvard University（斯坦福大学; 苹果公司; 牛津大学; 普林斯顿高等研究院; 哈佛大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究探讨全知预测与多组公平性之间的关系，证明普通全知预测不要求多组公平性，而损失结果不可区分性等价于校准多准确性。
AI 中文摘要
全知预测（Omniprediction）是一种学习保证，要求单个预测器相对于基准类别中的最优假设，在从损失函数族中选取的任何损失下都具有竞争力。损失结果不可区分性（简称loss OI）是一个更强的概念，蕴含全知预测。它要求预测的标签分布与真实分布在依赖于损失函数和基准类别的测试下不可区分。多准确性和多校准是多组公平性概念，推广了经典的校准和期望准确性概念。大多数已知的全知预测学习算法（无论是标准概念还是如loss OI的强化版本）都依赖于这些多组公平性概念的某些版本，或依赖于一个称为校准多准确性的中间概念。我们询问这是否必要：全知预测是否要求某种形式的多组公平性？我们证明，对于（普通）全知预测，答案是否定的；而对于loss OI，答案是肯定的。首先，一系列工作表明多校准或校准多准确性蕴含全知预测。我们排除了甚至弱逆命题，通过证明对于适当损失的全知预测甚至不蕴含期望准确性，而期望准确性是比校准、多准确性或多校准弱得多的概念。其次，先前的工作展示了如何从校准和多准确性的组合实现loss OI。我们证明了一个逆命题：loss OI等价于一种校准多准确性形式。
英文摘要
Omniprediction is a learning guarantee which requires a single predictor to be competitive relative to the best hypothesis from a benchmark class for any loss chosen from a family of loss functions. Loss Outcome Indistinguishability (loss OI for short) is a stronger notion that implies omniprediction. It requires the predicted distribution on labels to be indistinguishable from the true distribution to tests that depend on the loss functions and the benchmark class. Multiaccuracy and multicalibration are multigroup fairness notions that generalize classical notions of calibration and accuracy in expectation. Most known learning algorithms for omniprediction (both for the standard notion and for strengthenings like loss OI) rely on some version of these multigroup fairness notions, or on an intermediate notion called calibrated multiaccuracy. We ask if this is necessary: Does omniprediction require some form of multigroup fairness? We show that the answer is no for (plain) omniprediction, and yes for loss OI. First, a sequence of works shows that multicalibration or calibrated multiaccuracy imply omniprediction. We rule out even a weak converse, by showing that omniprediction for proper losses does not imply even accuracy in expectation, a much weaker notion than any of calibration, multiaccuracy, or multicalibration. Second, prior work showed how to achieve loss OI from a combination of calibration and multiaccuracy. We show a converse: loss OI is equivalent to a form of calibrated multiaccuracy.
CommentsAccepted for presentation at NeurIPS 2026
