Joshua D. Noe

This study develops a model-informed Deep Q-Network to automate infliximab dosing in pediatric Crohn's disease, achieving high target attainment with minimal overtreatment.

This paper develops a hybrid population PK-ML model to improve infliximab concentration predictions in pediatric and young adult Crohn's disease patients, showing better accuracy than Bayesian estimation alone.

The paper uses measured and predicted gene expression to identify ulcerative colitis patients at elevated risk of colectomy, showing both methods independently replicate in multiple datasets.