Tabular Foundation Models Can Do Survival Analysis
This work addresses the problem of enabling tabular foundation models to perform survival analysis for applications like healthcare and finance, representing an incremental advancement by extending existing models to a new task.
The paper tackles the challenge of adapting tabular foundation models to survival analysis by developing a classification-based framework that handles right-censoring through discretized event times, and it demonstrates that this approach outperforms classical and deep learning baselines across 53 real-world datasets on multiple survival metrics.
While tabular foundation models have achieved remarkable success in classification and regression, adapting them to model time-to-event outcomes for survival analysis is non-trivial due to right-censoring, where data observations may end before the event occurs. We develop a classification-based framework that reformulates both static and dynamic survival analysis as a series of binary classification problems by discretizing event times. Censored observations are naturally handled as examples with missing labels at certain time points. This classification formulation enables existing tabular foundation models to perform survival analysis through in-context learning without explicit training. We prove that under standard censoring assumptions, minimizing our binary classification loss recovers the true survival probabilities as the training set size increases. We demonstrate through evaluation across $53$ real-world datasets that off-the-shelf tabular foundation models with this classification formulation outperform classical and deep learning baselines on average over multiple survival metrics.