Dream to Control: Learning Behaviors by Latent Imagination

12/03/2019
by   Danijar Hafner, et al.
0

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.

READ FULL TEXT

page 2

page 5

10/05/2020

Mastering Atari with Discrete World Models

Intelligent agents need to generalize from past experience to achieve go...
02/16/2021

Steadily Learn to Drive with Virtual Memory

Reinforcement learning has shown great potential in developing high-leve...
07/10/2021

LS3: Latent Space Safe Sets for Long-Horizon Visuomotor Control of Iterative Tasks

Reinforcement learning (RL) algorithms have shown impressive success in ...
02/19/2021

Deep Latent Competition: Learning to Race Using Visual Control Policies in Latent Space

Learning competitive behaviors in multi-agent settings such as racing re...
02/19/2021

Learning Composable Behavior Embeddings for Long-horizon Visual Navigation

Learning high-level navigation behaviors has important implications: it ...
05/27/2022

Isolating and Leveraging Controllable and Noncontrollable Visual Dynamics in World Models

World models learn the consequences of actions in vision-based interacti...
03/19/2020

Learning to Fly via Deep Model-Based Reinforcement Learning

Learning to control robots without requiring models has been a long-term...

Code Repositories

dreamer

Dream to Control: Learning Behaviors by Latent Imagination


view repo