Hafner et al.: discrete world models for Atari behavior learning

What changed

Used discrete latent dynamics and separately trained imagined behavior learning to obtain strong aggregated Atari performance and extend latent-world-model control to discrete-action tasks.

Scientific scope

Used discrete latent dynamics and separately trained imagined behavior learning to obtain strong aggregated Atari performance and extend latent-world-model control to discrete-action tasks. Human-level aggregate score is not mastery of every game or general intelligence. Environment interactions, model training and hardware budget remain costs. First preprint 2020, conference 2021, pinned revision 2022.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Human-level aggregate score is not mastery of every game or general intelligence. Environment interactions, model training and hardware budget remain costs. First preprint 2020, conference 2021, pinned revision 2022.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Mastering Atari with Discrete World Models
• Hafner et al.: discrete world models for Atari behavior learning — author/company project record

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction