Semi-supervised Estimation of Average Survival

This paper develops semi-supervised, transfer learning methods to improve efficiency of pragmatic trials when the study endpoint is time-to-event and is captured using imperfect surrogates in electronic health records.