NORMA, a conditional transformer framework, generates blood biomarker reference intervals by conditioning on patient history and population priors, outperforming pure personalization in predicting clinical outcomes such as mortality and kidney injury.
A multimodal and temporal foundation model for virtual patient representations at healthcare system scale.arXiv [cs.LG], April 2026
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
-
Learning Normal Representations for Blood Biomarkers
NORMA, a conditional transformer framework, generates blood biomarker reference intervals by conditioning on patient history and population priors, outperforming pure personalization in predicting clinical outcomes such as mortality and kidney injury.