Temporally-Consistent Survival Analysis

Lucas Maystre, Daniel Russo

Keywords: survival analysis, time-to-event, dynamic survival, temporal-difference learning, policy evaluation

TL;DR: We take advantage of temporal consistency to learn more data-efficient & more accurate survival models from sequential data.

Abstract: We study survival analysis in the dynamic setting: We seek to model the time to an event of interest given sequences of states. Taking inspiration from temporal-difference learning, a central idea in reinforcement learning, we develop algorithms that estimate a discrete-time survival model by exploiting a temporal-consistency condition. Intuitively, this condition captures the fact that the survival distribution at consecutive states should be similar, accounting for the delay between states. Our method can be combined with any parametric survival model and naturally accommodates right-censored observations. We demonstrate empirically that it achieves better sample-efficiency and predictive performance compared to approaches that directly regress the observed survival outcome.