Static guide overview for crawlers and no-JS readers.
| Attribute | Value |
|---|---|
| Guide | How to Read World Models Papers |
| Summary | A practical reading path through world-model research, from foundational concepts to latent dynamics, planning, simulators, and self-supervised approaches. |
| Related Models | 5 |
| Related Research | 3 |
| References | 4 |
Editorial guide section preserved directly in static HTML.
Start with Ha & Schmidhuber (2018) 'World Models', the paper that popularized the concept. Then read Hafner et al. (2019) 'Learning Latent Dynamics for Planning from Pixels' (PlaNet/RSSM), the foundational architecture for modern world models. Silver et al. (2017) 'The Predictron' introduced abstract world models for planning.
Editorial guide section preserved directly in static HTML.
The Dreamer family represents the most successful line of world model research: DreamerV1 (2020) introduced imagination-based policy learning with RSSM, DreamerV2 (2021) 'Mastering Atari with Discrete World Models' achieved human-level Atari, and DreamerV3 (2023) 'Mastering Diverse Domains through World Models' achieved the first domain-general world model agent.
Editorial guide section preserved directly in static HTML.
Schrittwieser et al. (2020) 'Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model' (MuZero): superhuman game play without rules. Weber et al. (2017) 'Imagination-Augmented Agents': hybrid model-based/model-free agents.
Editorial guide section preserved directly in static HTML.
NVIDIA (2024) 'Cosmos World Foundation Model Platform': industrial-scale world models for physical AI. Google DeepMind (2024) 'Genie 2': interactive 3D environments from single images. Yang et al. (2023) 'Learning Interactive Real-World Simulators' (UniSim).
Editorial guide section preserved directly in static HTML.
LeCun (2022) 'A Path Towards Autonomous Machine Intelligence': the vision for JEPA-based world models. Bardes et al. (2024) 'V-JEPA': self-supervised visual world models without pixel reconstruction.
Editorial guide section preserved directly in static HTML.
For comprehensive overviews, look for survey papers on model-based RL and world models that organize the field's development, taxonomy, and future directions. These provide essential context for understanding how individual contributions fit together.
| Model | Lab | Category | Year |
|---|---|---|---|
| DreamerV3 | Google DeepMind | Model-Based RL | 2023 |
| World Models (Ha & Schmidhuber) | Google Brain / IDSIA | Model-Based RL | 2018 |
| MuZero | Google DeepMind | Model-Based RL | 2020 |
| PlaNet | Model-Based RL | 2019 | |
| V-JEPA | Meta | Self-Supervised World Model | 2024 |
| Topic | Summary |
|---|---|
| Model-Based Reinforcement Learning | What model-based reinforcement learning is, how world models enable imagination-based planning, and why Dreamer, MuZero, PlaNet, and TD-MPC2 matter. |
| Self-Supervised World Models | How self-supervised world models learn environment dynamics without rewards, from JEPA and V-JEPA to predictive latent representations. |
| World Models: A Comprehensive Survey | A survey of AI world models covering taxonomy, leading architectures, landmark systems, open challenges, and future research directions. |
FAQ answers rendered directly into static HTML for extractable responses.
Ha & Schmidhuber (2018) 'World Models' for the concept. For practical impact, Hafner et al. (2023) 'Mastering Diverse Domains through World Models' (DreamerV3) represents the current state of the art.
Start with the abstract and introduction for intuition. Study the architecture diagrams carefully. For DreamerV3 and MuZero, the open-source code is often clearer than the mathematical notation.
Short extractable summary preserved directly in static HTML.
Editorial provenance and refresh policy preserved directly in static HTML.
Published by world-models.io editorial board.
Lead editor Bernard Grenat.
This guide page turns primary-source material into structured workflows, explanations, and linked reading paths.
Each editorial page is assembled from primary sources, normalized into extractable summaries, checked for factual drift, and reviewed before publication or major refreshes. Last reviewed: 2026-06-21.
Pages are refreshed when a new paper, benchmark, release, architecture update, or stronger primary source materially changes the answer a reader or AI system should retrieve.
Each page links back to relevant primary sources and keeps a stable canonical URL so readers can verify claims, trace context, and reference the most up-to-date version. See the editorial policy.