SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1903.00374 · 2019

Model-Based Reinforcement Learning for Atari

Łukasz Kaiser, Sergey Levine, Chelsea Finn, Afroz Mohiuddin, George Tucker, Konrad Czechowski, Henryk Michalewski, Mohammad Babaeizadeh, Deepsense Ai, Ryan Sepassi, Piotr Miłos, Błażej Osi Ński, and 3 more

arXiv · PDF · Open in the Atlas

Code that ran

16 of 21 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (8 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 10 have no contract verdict; 6 failures of our instrument; 5 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1903.00374.json · badge: /api/ran/1903.00374.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
thomas-schillaci/SimPLe community
Papers with Code archive
14 ran · 3 no recorded run
tensorflow/tensor2tensor marked official in Papers with Code
Papers with Code archive
2 ran · 2 no recorded run
FunctionStatusWhere it lives
ActionInjector Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("46f5f7efa2488ce5")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

BitsPredictor Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("eb5b747f01be0865")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Container Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("3f3786cc58d216a6")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

MeanAttention Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("46579683ae077d93")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

MiddleNetwork Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("daf905c29d5ce6db")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

ParameterSealer Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("466db63e34381d96")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Restarter Ran
ran on a synthesized input
tensorflow/tensor2tensor/tensor2tensor/rl/trainer_model_based.py
code served (permissive licence) · get_code("44bf0671a31eeade")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned _GeneratorContextManager. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create a temporary directory for checkpoints
temp_dir = tempfile.mkdtemp()

# Create a fake checkpoint file to simulate an existing checkpoint at step 100
checkpoint_path = os.path.join(temp_dir, "model.ckpt-100")
with open(checkpoint_path, 'w') as f:
    f.write("fake checkpoint")

# Create checkpoint state file
checkpoint_state = os.path.join(temp_dir, "checkpoint")
with open(checkpoint_state, 'w') as f:
    f.write(f'model_checkpoint_path: "{checkpoint_path}"\n')
    f.write(f'all_model_checkpoint_paths: "{checkpoint_path}"\n')

INIT_ARGS = {
    "model_mode": "train",
    "checkpoint_dir": temp_dir,
    "target_local_step": 150
}

# The main computation method is training_loop, which is a context manager
METHOD = "training_loop"

# training_loop takes no arguments (it's a context manager)
ARGS = {}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type _GeneratorContextManager — a _GeneratorContextManager output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _GeneratorContextManager output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
RewardEstimator Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("b32c79f2cdecb3f4")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

ValueEstimator Ran
ran on a synthesized input · constructed an object; no result was computed
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("f443555aff8be5de")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

bit_to_int Ran
ran; our contract draft was wrong, not the code
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("de8fa36badf28fa1")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_timing_signal_nd Ran
ran; our contract draft was wrong, not the code
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("ba50aefcc31495b3")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

int_to_bit Ran
ran; our contract draft was wrong, not the code
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("6e9a001a9d751bff")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

mix Ran
ran; our contract draft was wrong, not the code
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("3ed887b5cb35096c")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

one_hot_encode Ran
ran; our contract draft was wrong, not the code
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("e2e902f8ee14d5c6")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

sample_with_temperature Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("e71f7c4a3dcf9c53")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

world_model_step_increment Ran
ran on a synthesized input
tensorflow/tensor2tensor/tensor2tensor/rl/trainer_model_based.py
code served (permissive licence) · get_code("7f4e91b128910dac")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

class HParams:
    def __init__(self):
        self.initial_epoch_train_steps_multiplier = 2
        self.model_train_steps = 100

hparams = HParams()
epoch = 0

ARGS = {
    "hparams": hparams,
    "epoch": epoch
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
NextFramePredictor Not yet run
unverified (harvested, no recorded run)
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("93ead6112daae1d2")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

StochasticModel Not yet run
unverified (harvested, no recorded run)
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("7be2f31a986731f4")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

standardize_frame Not yet run
unverified (harvested, no recorded run)
thomas-schillaci/SimPLe/simple/next_frame_predictor.py
pointer only (licence: NOASSERTION) · get_code("c321a59458ae8d05")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, thomas-schillaci/SimPLe (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

train_supervised Not yet run
unverified (harvested, no recorded run)
tensorflow/tensor2tensor/tensor2tensor/rl/trainer_model_based.py
code served (permissive licence) · get_code("a43b6bf6738c00ab")
How it ran

no invocation recorded.

train_world_model Not yet run
unverified (harvested, no recorded run)
tensorflow/tensor2tensor/tensor2tensor/rl/trainer_model_based.py
code served (permissive licence) · get_code("0697e8a4c6c8e81e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 21 samples lifted from this paper's linked repositories, 21 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 21 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

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.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1903.00374")
get_code_for_paper("1903.00374")
have("1903.00374")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/1903.00374.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/1903.00374.svg)](https://syntology.ai/paper/1903.00374)

Connect an agent — have() is free.