The R2D2 project proposes to study how formal or semi-formal representations of clinical protocols for diagnosis can be automatically learned from patient data. To this end, we (1) study two families of methods to infer such decision processes from clinical data warehouses, and (2) will define a framework for evaluating and comparing learned and pre-existing protocols.

Contacts:
Antoine Neuraz, rf.ph1732361199pa@za1732361199ruen.1732361199eniot1732361199na1732361199, HeKA team and Necker Hospital, AP-HP
Adrien Coulet, rf.ai1732361199rni@t1732361199eluoc1732361199.neir1732361199da1732361199, HeKA team