Fetching the paper…

Dreaming: Model-based Reinforcement Learning by Latent Imagination without Reconstruction · Around