// ◐ World Models & Self-Supervised Learning

Dreamer (Latent Imagination)

Trains behavior by imagining trajectories in a learned latent world model.

Dreamer（隐空间想象） 2019 Advanced core

IN PLAIN WORDS

Learn a recurrent state-space model; train actor–critic on imagined rollouts. DreamerV3 solves 150+ tasks with fixed hyperparameters and first collected diamonds in Minecraft from scratch.

Used in: Model-based RL in robotics and games

TIMELINE · prerequisites → this → what builds on it

2018 World Models2018 Soft Actor-Cr…2019 Dreamer (Latent I…

01Original paper

Dream to Control: Learning Behaviors by Latent Imagination

Danijar Hafner, Timothy Lillicrap, Jimmy Ba, et al. · 2019

arXiv ↗ ✓ Verified · arXiv API

Note: DreamerV3：arXiv:2301.04104（2023）

02Where it sits

Learn the left column first; the right column builds on this one.

Learn first

Dreamer (Latent Imagination)

Then you can learn

Nothing in the atlas builds on it yet

Left: prerequisites · right: builds on this · dashed: related. Full map →

03Learning resources

Official docs, explainers, courses and reference code. All links checked.

Reference code

github.com ↗EN

04Where the course teaches it

U73 Speech, Video & World Models Whisper, video diffusion (Sora-style DiT), 3-D reconstruction, world models, Dreamer and Genie. 8 h10 algorithms4 papers

On learning paths:Generative Models & Frontier Research

For AI / data

Verification: verified, checked 2026-10-21.
JSON: /api/learn/algorithms.json (id: dreamer)