Calibeating Prediction-Powered Inference
校准预测驱动推断
机构 * Department of Statistics, University of Washington(华盛顿大学统计学系) ; Department of Statistics and Division of Biostatistics, Center for Targeted Machine Learning and Causal Inference, University of California, Berkeley(加州大学伯克利分校统计学系与生物统计学系,目标机器学习与因果推断中心) ; Targeted ML Solutions, targetedml.com(Targeted ML Solutions公司)
AI总结 本文提出校准预测驱动推断方法,通过后处理校准预测分数提升半监督估计效率,证明线性与等距校准的优化性,并展示其在模拟和真实数据中优于PPI和AIPW。
Comments Paper website: https://larsvanderlaan.github.io/ppi-aipw/