Hybrid token-based binning for numeric values in EHR transformers is more robust than explicit interaction modeling, with optimal bin count following an empirically derived power-law in dataset size.
Advances in Neural Information Processing Systems (NeurIPS) , volume=
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
-
How Should Transformers Encode Numeric Values in Electronic Health Records?
Hybrid token-based binning for numeric values in EHR transformers is more robust than explicit interaction modeling, with optimal bin count following an empirically derived power-law in dataset size.