Leakage-aware distillation transfers at least 90% of tabular foundation model AUC to lightweight students across 19 health datasets, with 26x CPU speedup and preserved calibration/fairness.
Obtaining well calibrated probabilities using Bayesian binning into quantiles
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Distilling Tabular Foundation Models for Structured Health Data
Leakage-aware distillation transfers at least 90% of tabular foundation model AUC to lightweight students across 19 health datasets, with 26x CPU speedup and preserved calibration/fairness.