Fig. 9From: A dosing strategy model of deep deterministic policy gradient algorithm for sepsis patientsChange trend of TD-error after each training episodeBack to article page