Papers › Model-Based Reinforcement Learning for Atari

Model-Based Reinforcement Learning for Atari

1 Mar 2019arXiv:1903.00374archive 2025-07-28

Lukasz Kaiser, Mohammad Babaeizadeh, Piotr Milos, Blazej Osinski, Roy H. Campbell, Konrad Czechowski, Dumitru Erhan, Chelsea Finn, Piotr Kozakowski, Sergey Levine, Afroz Mohiuddin, Ryan Sepassi, George Tucker, Henryk Michalewski

Model-free reinforcement learning (RL) can be used to learn effective policies for complex tasks, such as Atari games, even from image observations. However, this typically requires very large amounts of interaction -- substantially more, in fact, than a human would need to learn the same games. How can people learn so quickly? Part of the answer may be that people can learn how the game works and predict which actions will lead to desirable outcomes. In this paper, we explore how video prediction models can similarly enable agents to solve Atari games with fewer interactions than model-free methods. We describe Simulated Policy Learning (SimPLe), a complete model-based deep RL algorithm based on video prediction models and present a comparison of several model architectures, including a novel architecture that yields the best results in our setting. Our experiments evaluate SimPLe on a range of Atari games in low data regime of 100k interactions between the agent and the environment, which corresponds to two hours of real-time play. In most games SimPLe outperforms state-of-the-art model-free algorithms, in some games by over an order of magnitude.

PaperPDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 21 code samples harvested from 2 repositories linked to this paper: 16 ran (of which 8 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 6 where Syntology's instrument failed) · 5 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 4 samples from 1 repository, 2 ran (2 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 17 samples from 1 repository, 14 ran (8 with no instrument failure; 6 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

tensorflow/tensor2tensor officialmentioned in papermentioned on GitHubtfSyntology: 4 samples, 2 ran (0 instrument-failed) · 2 unverified report
thomas-schillaci/SimPLe mentioned on GitHubpytorchSyntology: 17 samples, 14 ran (6 instrument-failed) · 3 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

21 samples harvested: 16 ran (of which 8 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 6 where Syntology's instrument failed) · 5 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

5ran · our draft was wrong
1ran; the contract check could not run
10ran
5unverified

Licence: 17 of the 21 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. For 3 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

tensorflow/tensor2tensor official repository 4 samples, 2 ran (0 instrument-failed) · 2 unverified
Restarter tensor2tensor/rl/trainer_model_based.py ran Apache-2.0 (permissive) · 44bf0671a31eeade · report fixture: returned _GeneratorContextManager · level recorded 2026-09-24
world_model_step_increment tensor2tensor/rl/trainer_model_based.py ran Apache-2.0 (permissive) · 7f4e91b128910dac · report fixture: signature hparams:HParams/0/?, epoch:int/0/int · returned int · level recorded 2026-09-25
train_supervised tensor2tensor/rl/trainer_model_based.py unverified Apache-2.0 (permissive) · a43b6bf6738c00ab · report
train_world_model tensor2tensor/rl/trainer_model_based.py unverified Apache-2.0 (permissive) · 0697e8a4c6c8e81e · report
thomas-schillaci/SimPLe community (archive-listed) 17 samples, 14 ran (6 instrument-failed) · 3 unverified
ActionInjector simple/next_frame_predictor.py ran licence not identified · pointer only · 46f5f7efa2488ce5 · report fixture: returned ActionInjector · level recorded 2026-09-22
BitsPredictor simple/next_frame_predictor.py ran licence not identified · pointer only · eb5b747f01be0865 · report fixture: returned BitsPredictor · level recorded 2026-09-22
Container simple/next_frame_predictor.py ran licence not identified · pointer only · 3f3786cc58d216a6 · report fixture: returned Container · level recorded 2026-09-22
MeanAttention simple/next_frame_predictor.py ran fingerprinted licence not identified · pointer only · 46579683ae077d93 · report fixture: returned MeanAttention · level recorded 2026-09-22
MiddleNetwork simple/next_frame_predictor.py ran fingerprinted licence not identified · pointer only · daf905c29d5ce6db · report fixture: returned MiddleNetwork · level recorded 2026-09-22
ParameterSealer simple/next_frame_predictor.py ran licence not identified · pointer only · 466db63e34381d96 · report fixture: returned ParameterSealer · level recorded 2026-09-22
RewardEstimator simple/next_frame_predictor.py ran licence not identified · pointer only · b32c79f2cdecb3f4 · report fixture: returned RewardEstimator · level recorded 2026-09-22
ValueEstimator simple/next_frame_predictor.py ran fingerprinted licence not identified · pointer only · f443555aff8be5de · report fixture: returned ValueEstimator · level recorded 2026-09-22
bit_to_int simple/next_frame_predictor.py ran · our draft was wrong licence not identified · pointer only · de8fa36badf28fa1 · report fixture: signature x_bit:Tensor/2/float, num_bits:int/0/int · returned Tensor · level recorded 2026-09-22
get_timing_signal_nd simple/next_frame_predictor.py ran · our draft was wrong licence not identified · pointer only · ba50aefcc31495b3 · report fixture: signature shape:tuple/1/int, min_timescale:float/0/float, max_timescale:float/0/float · returned Tensor · level recorded 2026-09-22
int_to_bit simple/next_frame_predictor.py ran · our draft was wrong licence not identified · pointer only · 6e9a001a9d751bff · report fixture: signature x_int:Tensor/1/int, num_bits:int/0/int · returned Tensor · level recorded 2026-09-22
mix in 6 papers simple/next_frame_predictor.py ran · our draft was wrong fingerprinted licence not identified · pointer only · 3ed887b5cb35096c · report fixture: signature x1:Tensor/2/float, x2:Tensor/2/float, epsilon:float/0/float · returned Tensor · level recorded 2026-09-22
one_hot_encode in 18 papers simple/next_frame_predictor.py ran · our draft was wrong licence not identified · pointer only · e2e902f8ee14d5c6 · report fixture: signature action:Tensor/1/int, n:int/0/int, dtype:dtype/0/? · returned Tensor · level recorded 2026-09-22
sample_with_temperature simple/next_frame_predictor.py ran; the contract check could not run (RAISES) fingerprinted licence not identified · pointer only · e71f7c4a3dcf9c53 · report fixture: signature logits:Tensor/2/float, temperature:float/0/float · returned Tensor · level recorded 2026-09-22
NextFramePredictor simple/next_frame_predictor.py unverified licence not identified · pointer only · 93ead6112daae1d2 · report
StochasticModel simple/next_frame_predictor.py unverified licence not identified · pointer only · 7be2f31a986731f4 · report
standardize_frame simple/next_frame_predictor.py unverified licence not identified · pointer only · c321a59458ae8d05 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. Siblings are assemblies lifted from this paper in the same batch for the same method: alternative readings kept side by side; none replaces or supersedes another.

2 assemblies on this paper: 2 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).

2 sibling assemblies for SimPLe (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.

harvested Assembly for SimPLe (graph method: SimPLe). 17 bricks; individually on a fixture: 14 ran (of which 8 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 6 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.

one of 2 siblings

  1. NextFramePredictor entry · level 0 · unverified · 93ead6112daae1d2
  2. ActionInjector helper · level 1 · ran · 46f5f7efa2488ce5
  3. Container helper · level 1 · ran · 3f3786cc58d216a6
  4. MiddleNetwork model · level 1 · ran · daf905c29d5ce6db
  5. RewardEstimator helper · level 1 · ran · b32c79f2cdecb3f4
  6. StochasticModel model · level 0 · unverified · 7be2f31a986731f4
  7. ValueEstimator helper · level 1 · ran · f443555aff8be5de
  8. get_timing_signal_nd helper · level 1 · ran · our draft was wrong · ba50aefcc31495b3
  9. standardize_frame data_transform · level 0 · unverified · c321a59458ae8d05
  10. ParameterSealer helper · level 1 · ran · 466db63e34381d96
  11. BitsPredictor model · level 1 · ran · eb5b747f01be0865
  12. MeanAttention helper · level 1 · ran · 46579683ae077d93
  13. mix helper · level 1 · ran · our draft was wrong · 3ed887b5cb35096c
  14. bit_to_int data_transform · level 1 · ran · our draft was wrong · de8fa36badf28fa1
  15. int_to_bit data_transform · level 1 · ran · our draft was wrong · 6e9a001a9d751bff
  16. one_hot_encode helper · level 1 · ran · our draft was wrong · e2e902f8ee14d5c6
  17. sample_with_temperature helper · level 1 · ran; the contract check could not run (RAISES) · e71f7c4a3dcf9c53

harvested Assembly for SimPLe (graph method: SimPLe). 4 bricks; individually on a fixture: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

one of 2 siblings

  1. train_world_model entry · level 0 · unverified · 0697e8a4c6c8e81e
  2. Restarter helper · level 1 · ran · 44bf0671a31eeade
  3. train_supervised helper · level 0 · unverified · a43b6bf6738c00ab
  4. world_model_step_increment scheduler · level 1 · ran · 7f4e91b128910dac

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="1903.00374") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Atari GamesAtari Games 100kModel-based Reinforcement LearningReinforcement LearningReinforcement Learning (RL)Video Predictionmodelreinforcement-learning

Datasets

Introduced by this paper, per the archive.

Atari 100k

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Atari Games Atari games SimPLe Mean Human Normalized Score 25.3% #12 of 12 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections