Fig. 3From: Personalized project recommendations: using reinforcement learningFramework steps of DQN-TBBack to article page