Synthetic Training Data
Training on richer and more diverse measurement distributions improved TEMPO's generalization. Synthetic experiments covered nine generative configurations varying event timing, stage type, stage prior, measurement model, and biomarker dis…
1 sources - 4 claims
Training on richer and more diverse measurement distributions improved TEMPO's generalization. Synthetic experiments covered nine generative configurations varying event timing, stage type, stage prior, measurement model, and biomarker distribution. TEMPO depends on the quality of assumed synthetic disease-progression frameworks. Synthetic cohorts provide known biomarker sequences and patient stages for supervised training.