AAAI
2010
13 years 6 months ago
2010
We consider the problem of incorporating end-user advice into reinforcement learning (RL). In our setting, the learner alternates between practicing, where learning is based on ac...