多组半监督学习在某些组上误差可能随组数线性上升，远差于单组情况。
The price of multi-group transductive learning
- 分析多组半监督学习的误差下界，发现组数越多，最差组误差越大
- 误差惩罚可高达样本量平方根级别，随组数线性增长
- 适用于关注公平性与鲁棒性的机器学习研究者
我们证明，在半监督设定中，任何多组学习器在某些组上的误差率都可能相对于单组设定产生乘法级惩罚，且该惩罚随组数线性增长，最高可达样本量平方根量级。这与在类似（组可实现）统计设定中表现最优的多组学习器形成鲜明对比：后者惩罚仅对数级增长，且与组数无关。
原文摘要 · Abstract (English)
We show every multi-group learner in the transductive setting may incur a multiplicative penalty in its error rate on some group relative to the error rate achievable in the single-group setting, and the penalty can increasing linearly with the number of groups, up to roughly the square-root of the sample size. This stands in stark contrast to optimal multi-group learners in an analogous (group-realizable) statistical setting, where the penalty is always at most logarithmic in the sample size and independent of the number of groups.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。