SDPM: Survival Diffusion Probabilistic Model for Continuous-Time Survival Analysis

The paper proposes SDPM, a generative model using denoising diffusion to estimate continuous-time survival (time-to-event) distributions without parametric hazard assumptions or time discretization. It is evaluated on ten real datasets against tree-based, boosting, and neural baselines, showing competitive C-index, AUC, and Brier score performance.