New: the Timeline is live. Track world model releases, papers, and benchmark updates in real time.
world-models.io
The Knowledge Hub for AI World Models

World Models (Ha & Schmidhuber)

The seminal 2018 paper by Ha & Schmidhuber that introduced the modern concept of world models for reinforcement learning, combining VAE, MDN-RNN, and controller components.

robotics model-based-rl simulation embodied-ai

Key Attributes

Static HTML snapshot of the model record for crawlers and no-JS readers.

AttributeValue
ModelWorld Models (Ha & Schmidhuber)
Lab / OrganizationDeepMind
CategoryModel-Based RL
SubtypeVAE + MDN-RNN
World Model TypeGenerative latent dynamics model
Primary DomainGame environments
ArchitectureVAE (visual encoder) + MDN-RNN (dynamics) + linear controller
ModalityVisual
Training MethodUnsupervised VAE training + RNN dynamics learning + evolutionary policy search
Statusfoundational
Year2018
Performance Index48/100 (high confidence, v1.1)

About World Models (Ha & Schmidhuber)

Main editorial body preserved directly in static HTML.

Ha and Schmidhuber's World Models paper demonstrated that an agent can learn a compressed spatial and temporal representation of the environment, then train a policy entirely within this learned 'dream' world. The architecture uses a VAE for spatial encoding and an MDN-RNN for temporal dynamics, creating a generative model of the environment that the agent can imagine in.

World Models (Ha & Schmidhuber) is a generative latent dynamics model developed by Google Brain / IDSIA in 2018 for game environments.

Editorial Snapshot

Short extractable facts for answer engines and no-JS readers.

SignalValue
DefinitionWorld Models (Ha & Schmidhuber) is a generative latent dynamics model developed by Google Brain / IDSIA in 2018 for game environments.
Short DescriptionThe seminal paper that popularized the concept of world models: learning to imagine environments and training policies entirely in dreams.
Benchmark Rows1
FAQ Entries1
Related Models3
Related Guides3
Related Research Topics3
Last Updated2026-02-10

Notable Features

Key capabilities associated with this model.

  • Coined the modern usage of 'world models' in RL
  • Training policies entirely in dreams
  • VAE + RNN architecture
  • Interactive web-based demonstrations

Use Cases

Representative applications attached to this model record.

Game playing (VizDoom, CarRacing)Imagination-based policy learningConceptual demonstration of world models

Strengths and Limitations

Balanced assessment surfaced in static HTML.

Strengths

  • Conceptually foundational
  • Elegant and interpretable
  • Demonstrated dream-based training
  • Highly influential

Limitations

  • Simple environments only
  • VAE reconstruction loss limitations
  • Limited scalability

Benchmarks

Published benchmark evidence attached to this model record.

BenchmarkMetricResultSource
CarRacing-v0Avg Score 906 / 1000Competitive (dream-trained)Source

References and Citations

Primary references preserved in static HTML for citation extraction.

ReferenceLink
Ha, D. & Schmidhuber, J., 2018. World Models. NeurIPS 2018.Open source

Related Models

Nearby models linked from the current editorial record.

ModelCategoryWorld Model TypeIndex v1.1
PlaNetModel-Based RLLatent space planning model57/100
DreamerV2Model-Based RLImagination-based dynamics model72/100
RSSMLatent DynamicsCore dynamics architecture64/100

Direct Comparisons

Side-by-side comparisons already connected to this model.

ComparisonMatchupSummary
Ha & Schmidhuber World Model vs DreamerV3Ha & Schmidhuber World Model vs DreamerV3The original 2018 'World Models' paper vs. the current state-of-the-art: how five years of research transformed a foundational concept into a domain-general world model agent.
AMI vs Ha World ModelAMI vs Ha World ModelTwo pioneering cognitive-inspired world models: Ha's 2018 World Model introduced the VAE+RNN+Controller architecture, while AMI proposes an autonomous machine intelligence framework inspired by biological cognition.

Guides Referencing This Model

Crawler-readable guide links tied to this model.

GuideSummary
World Models for BeginnersA comprehensive introduction to AI world models: what they are, how they work, and why they matter for the future of robotics, reinforcement learning, and embodied AI.
Building World Models: A Practical GuideA practical guide to implementing world models: from choosing architectures and training setups to debugging dynamics learning and policy optimization.
How to Read World Models PapersA practical reading path through world-model research, from foundational concepts to latent dynamics, planning, simulators, and self-supervised approaches.

Research Topics Referencing This Model

Connected research areas surfaced directly in static HTML.

TopicSummary
Model-Based Reinforcement LearningWhat model-based reinforcement learning is, how world models enable imagination-based planning, and why Dreamer, MuZero, PlaNet, and TD-MPC2 matter.
Self-Supervised World ModelsHow self-supervised world models learn environment dynamics without rewards, from JEPA and V-JEPA to predictive latent representations.
World Models: A Comprehensive SurveyA survey of AI world models covering taxonomy, leading architectures, landmark systems, open challenges, and future research directions.

Frequently Asked Questions

FAQ answers rendered directly into static HTML for extractable responses.

Is this where the term 'world model' comes from?

The concept existed earlier (e.g., Schmidhuber 1990), but Ha & Schmidhuber's 2018 paper popularized the term in the modern deep RL context and inspired the current wave of world model research.

Quick Answer

Short extractable summary preserved directly in static HTML.

  • World Models (Ha & Schmidhuber) is a generative latent dynamics model developed by Google Brain / IDSIA in 2018 for game environments.
  • Use this page when you need a fast read on how World Models (Ha & Schmidhuber) fits into the model-based rl landscape, then validate the details in the benchmarks, citations, and related pages.
  • A key strength surfaced in the editorial record is conceptually foundational.

Editorial Trust Signals

Editorial provenance and refresh policy preserved directly in static HTML.

Published by world-models.io editorial board.

Lead editor Tyler D. - Technical editor, methodology and benchmark analysis.

This model page synthesizes primary papers, official model pages, benchmark evidence, and related world-models.io context into a reference resource.

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: 2026-02-10.

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.

Primary sources onlyLast reviewed date visibleMethodology documentedSource links included

External Sources

Primary model and lab sources embedded in static HTML.

References

  1. [1] Ha, D. & Schmidhuber, J., 2018. World Models. NeurIPS 2018.