Tabular foundation models are not a shortcut to calibrated clinical risk prediction: a pre-registered simulation against the true risk
stat.ME, cs.LG
Submitted: 2026-10-28
Updated: 2026-10-28
Comments: 37 pages, 7 figures, 1 table; Supplementary Information (55 pages) as an ancillary file. Pre-registered on OSF Registries (doi:10.17605/OSF.IO/8HMTZ). Code: doi:10.5281/zenodo.23737378; benchmark data: doi:10.5281/zenodo.23737416
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Related papers
- Learning Robust Treatment Rules for Censored Data
- Neural Estimation of Summary Statistics for Simulation-Based Inference with Non-Asymptotic Guarantees
- Spectral divide-and-conquer MCMC for long stationary time series
- Improving the Efficiency of Subgroup Analysis in Randomized Controlled Trials with TMLE
- Personalized Federated Vector Autoregression with Personalization Diversity
- Doubly Robust and Efficient Calibration of Prediction Sets for Right-Censored Time-to-Event Outcomes