SurvICL: a Foundation Model for in-context Survival Analysis
Abstract
Tabular foundation models (TFMs) now outperform gradient-boosted trees in classification and regression. Yet in survival analysis, which predicts when an event occurs from censored data, often on small clinical cohorts, we find that no existing TFM beats the half-century-old Cox model. We address this issue by introducing SurvICL, a TFM that predicts survival curves in-context. SurvICL is pretrained on purely synthetic datasets generated from a novel flexible proportional-odds prior that outperforms other prior types in our ablation study. Unlike most other survival models, SurvICL is trained with a plain quantile regression loss, which is possible because its synthetic pre-training data generator can provide uncensored event times. Additionally, we introduce the largest survival benchmark to date spanning 103 datasets. On the 63 datasets held out from all design decisions, SurvICL outperforms 20 tree-based, deep, and foundation-model baselines without any tuning: it wins 82.9\% of pairwise comparisons in integrated Brier score, against 71.3\% for the runner-up, the Cox model, and is the first model to rank above Cox on small datasets.
Related
- TabCUQ: Context-Uncertainty Quantification for Tabular Foundation Models
- SurvFM enables tabular foundation models for right-censored survival prediction
- Evaluating Time-Series Foundation Models and Multimodal Dietary Context for CGM Forecasting
- What masking geometry works best for EEG foundation models?
- Limits and promises of earth observation foundation models in predicting multi-trophic soil biodiversity
BibTeX
@misc{alberge2026survicl,
title = {{SurvICL: a Foundation Model for in-context Survival Analysis}},
author = {Julie Alberge and Linus Bleistein and Judith Abécassis and Marine Le Morvan and Gaël Varoquaux and David Holzmüller},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.22242},
eprint = {2610.22242},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.22242},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - SurvICL: a Foundation Model for in-context Survival Analysis
AU - Julie Alberge
AU - Linus Bleistein
AU - Judith Abécassis
AU - Marine Le Morvan
AU - Gaël Varoquaux
AU - David Holzmüller
PY - 2026
DA - 2026/10/15
PB - arXiv
DO - 10.48550/arxiv.2610.22242
UR - https://arxiv.org/abs/2610.22242
AB - Tabular foundation models (TFMs) now outperform gradient-boosted trees in classification and regression. Yet in survival analysis, which predicts when an event occurs from censored data, often on small clinical cohorts, we find that no existing TFM beats the half-century-old Cox model. We address this issue by introducing SurvICL, a TFM that predicts survival curves in-context. SurvICL is pretrained on purely synthetic datasets generated from a novel flexible proportional-odds prior that outperforms other prior types in our ablation study. Unlike most other survival models, SurvICL is trained with a plain quantile regression loss, which is possible because its synthetic pre-training data generator can provide uncensored event times. Additionally, we introduce the largest survival benchmark to date spanning 103 datasets. On the 63 datasets held out from all design decisions, SurvICL outperforms 20 tree-based, deep, and foundation-model baselines without any tuning: it wins 82.9\% of pairwise comparisons in integrated Brier score, against 71.3\% for the runner-up, the Cox model, and is the first model to rank above Cox on small datasets.
ER -