Dream to Control: Learning Behaviors by Latent Imagination
Danijar Hafner, Timothy Lillicrap, Jimmy Ba, Mohammad Norouzi
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/danijar/dreamerOfficialtf★ 0
- github.com/Eclectic-Sheep/sheeprlpytorch★ 436
- github.com/kc-ml2/SimpleDreamerpytorch★ 152
- github.com/facebookresearch/denoised_mdppytorch★ 137
- github.com/agiachris/STAPpytorch★ 50
- github.com/adityabingi/Dreamerpytorch★ 49
- github.com/zdhnarsil/stochastic-marginal-actor-criticpytorch★ 24
- github.com/minhphd/PyDreamerV1pytorch★ 7
- github.com/juliusfrost/dreamer-pytorchpytorch★ 0
- github.com/chamorajg/pl-dreamerpytorch★ 0
Abstract
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.