Fig. 4From: Incorporating causal factors into reinforcement learning for dynamic treatment regimes in HIVThe evolution of reward of a the first patient; and b the 300th patientBack to article page