Dream to Control: Learning Behaviors by Latent Imagination
This addresses the challenge of deriving behaviors from world models for visual control tasks, offering a novel approach that improves efficiency and performance.
The paper tackles the problem of learning complex behaviors from high-dimensional sensory inputs by introducing Dreamer, a reinforcement learning agent that solves long-horizon tasks from images using latent imagination, achieving superior data-efficiency, computation time, and final performance on 20 challenging visual control tasks.
Learned world models summarize an agent's experience to facilitate learning complex behaviors. While learning world models from high-dimensional sensory inputs is becoming feasible through deep learning, there are many potential ways for deriving behaviors from them. We present Dreamer, a reinforcement learning agent that solves long-horizon tasks from images purely by latent imagination. We efficiently learn behaviors by propagating analytic gradients of learned state values back through trajectories imagined in the compact state space of a learned world model. On 20 challenging visual control tasks, Dreamer exceeds existing approaches in data-efficiency, computation time, and final performance.