Decoupling feature extraction from policy learning: assessing benefits of state representation learning in goal based robotics

Open in new window