aiqbench analysis
SurvPFN addresses a critical gap in tabular foundation models by natively handling censored data for time-to-event predictions, a limitation that previously rendered standard models unsuitable for real-world applications. By pretraining on synthetic tasks and utilizing distributional regression with a censored loss function, the model achieves competitive performance without requiring per-dataset fitting or specialized architectures. This approach effectively frames survival analysis as a continuous-time distributional regression problem, offering a robust solution for medical and reliability engineering domains.
SurvPFN: Towards Foundation Models for Survival Predictions
arXiv:2606.04564v1 Announce Type: new Abstract: Tabular foundation models (TFMs) have made rapid progress in standard classification and regression, but time-to-event survival prediction tasks have remained largely untouched. Unlike in standard regression tasks, survival prediction models must account for censored data. Standard TFMs cannot handle natively censored data, leading to biased and inaccurate predictions, making them unsuitable for real-world applications. To overcome this fundamental limitation, we propose \texttt{SurvPFN}, a prior-data fitted network (PFN), for survival prediction tasks. We pretrain \texttt{SurvPFN} on millions of synthetic survival prediction tasks to learn survival via distributional regression that accounts for censored data. \texttt{SurvPFN} works by (1) generating data with Weibull event times and a non-informative censoring mechanism; (2) integrating a censored event indicator; and (3) minimizing a censored negative log-likelihood. On SurvSet, a collection of real-world survival tasks, \texttt{SurvPFN} is highly competitive with classical and deep survival baselines without per-dataset fitting, a survival-specific architecture, or feature engineering. We show that survival can be treated as a continuous-time distributional regression problem with censored loss, unlocking the power of PFNs for time-to-event predictions.
Abstract:Tabular foundation models (TFMs) have made rapid progress in standard classification and regression, but time-to-event survival prediction tasks have remained largely untouched. Unlike in standard regression tasks, survival prediction models must account for censored data. Standard TFMs cannot handle natively censored data, leading to biased and inaccurate predictions, making them unsuitable for real-world applications. To overcome this fundamental limitation, we propose \texttt{SurvPFN}, a prior-data fitted network (PFN), for survival prediction tasks. We pretrain \texttt{SurvPFN} on millions of synthetic survival prediction tasks to learn survival via distributional regression that accounts for censored data. \texttt{SurvPFN} works by (1) generating data with Weibull event times and a non-informative censoring mechanism; (2) integrating a censored event indicator; and (3) minimizing a censored negative log-likelihood. On SurvSet, a collection of real-world survival tasks, \texttt{SurvPFN} is highly competitive with classical and deep survival baselines without per-dataset fitting, a survival-specific architecture, or feature engineering. We show that survival can be treated as a continuous-time distributional regression problem with censored loss, unlocking the power of PFNs for time-to-event predictions.
10 pages, 1 figure. Accepted to "Foundation Models for Structured Data" Workshop at the International Conference on Machine Learning (ICML) 2026
Machine Learning (cs.LG)
arXiv:2606.04564 [cs.LG]
(or arXiv:2606.04564v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.04564 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 210 words
Related coverage
A General-Purpose Molecular Foundation Model Transfers Across Diverse Olfactory Tasks
arXiv · Aug 27, 2026
Long-term Traffic Scene Prediction via Polynomial Representations in Autonomous Driving
arXiv · Aug 5, 2026
DoAtlas-2: A Foundation for Self-Evolving Causal Biomedical Discovery
arXiv · Sep 29, 2026
TabFORGE: A Tabular Foundation Model for Generative Modelling
arXiv · Oct 15, 2026