SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1910.01741 · 2019

Improving Sample Efficiency in Model-Free Reinforcement Learning from Images

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 4 failures of our instrument; 0 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/1910.01741.json · badge: /api/ran/1910.01741.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
denisyarats/pytorch_sac_ae canonical (classifier label)
classifier basis: plan_row
5 ran · 0 no recorded run
KarlXing/RL-Visual-Continuous-Control reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
gaussian_logprob Ran
ran; our contract draft was wrong, not the code
denisyarats/pytorch_sac_ae/sac_ae.py
code served (permissive licence) · get_code("daa7b3a0355eb7c3")
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) and returned Tensor, ndim 2, dtype family float, shape (2, 1) 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 numpy as np

torch.manual_seed(0)

# Create a small noise tensor with shape (batch_size=2, feature_dim=4)
noise = torch.randn(2, 4)

# Create a small log_std tensor with shape (batch_size=2, feature_dim=4)
log_std = torch.randn(2, 4)

ARGS = {
    "noise": noise,
    "log_std": log_std,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 1) — shape (2, 1) 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
make_dir Ran
ran; our contract draft was wrong, not the code
denisyarats/pytorch_sac_ae/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from silihuang-ai/idt, is pointer only (licence: NONE) · get_code("42958997018c3581")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from silihuang-ai/idt, and GitHub detected no licence in that repository when it was checked. 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.

make_encoder Ran
ran on a synthesized input
denisyarats/pytorch_sac_ae/encoder.py
code served (permissive licence) · get_code("b41b083f7956cdf2")
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 PixelEncoder. 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 torch
import torch.nn as nn

torch.manual_seed(0)

# Test with pixel encoder
ARGS = {
    'encoder_type': 'pixel',
    'obs_shape': (3, 84, 84),  # 3 channels, 84x84 image
    'feature_dim': 50,
    'num_layers': 2,
    'num_filters': 32
}
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 PixelEncoder — a PixelEncoder output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PixelEncoder 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: none by the harness; any seed is in the setup block
module_hash Ran
ran, honoured the contract we drafted
denisyarats/pytorch_sac_ae/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from agentification/RP_PGM, is pointer only (licence: NONE) · get_code("63cc0b17d508c362")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from agentification/RP_PGM, and GitHub detected no licence in that repository when it was checked. 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.

preprocess_obs Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
KarlXing/RL-Visual-Continuous-Control/src/agent/sac_ae.py
code served (permissive licence) · get_code("975f23eeea267e55")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) 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

torch.manual_seed(0)

# Create a small image tensor with shape (batch=2, channels=3, height=4, width=5)
# Values should be in [0, 255] range as typical for image preprocessing
obs = torch.randint(0, 256, (2, 3, 4, 5), dtype=torch.float32)

# bits parameter - using default value of 5
bits = 5

ARGS = {
    'obs': obs,
    'bits': bits
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
squash Ran
ran; our contract draft was wrong, not the code
denisyarats/pytorch_sac_ae/sac_ae.py
code served (permissive licence) · get_code("e4f442d1770eacd1")
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) and returned tuple. 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 torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for the function arguments
batch_size = 2
dim = 4

# mu: input tensor that will be passed through tanh
mu = torch.randn(batch_size, dim)

# pi: tensor that will be passed through tanh (can be None)
pi = torch.randn(batch_size, dim)

# log_pi: log probability tensor (can be None)
log_pi = torch.randn(batch_size, 1)

ARGS = {
    "mu": mu,
    "pi": pi,
    "log_pi": log_pi,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 5 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 5 later) and 1 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.

FunctionAlso present inEarliest-datedOthers named
gaussian_logprob 15 other papers
15 later · 3 only as an assembly brick
2004.04136 (2020-04)
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
2004.14990 (2020-04)
Reinforcement Learning with Augmented Data
2007.04309 (2020-07)
Self-Supervised Policy Adaptation during Deployment
and 12 more not named here
squash 14 other papers
14 later · 3 only as an assembly brick
2004.04136 (2020-04)
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
2004.14990 (2020-04)
Reinforcement Learning with Augmented Data
2007.04309 (2020-07)
Self-Supervised Policy Adaptation during Deployment
and 11 more not named here
make_dir 8 other papers
8 later
2004.04136 (2020-04)
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
2010.09163 (2020-10)
D2RL: Deep Dense Architectures in Reinforcement Learning
2202.06258 (2022-02)
Flowformer: Linearizing Transformers with Conservation Flows
and 5 more not named here
module_hash 5 other papers
5 later
2004.04136 (2020-04)
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
2010.09163 (2020-10)
D2RL: Deep Dense Architectures in Reinforcement Learning
2203.12759 (2022-03)
Asynchronous Reinforcement Learning for Real-Time Control of Physical Robots
and 2 more not named here
preprocess_obs 4 other papers
4 later
2004.04136 (2020-04)
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
2010.09163 (2020-10)
D2RL: Deep Dense Architectures in Reinforcement Learning
2203.12759 (2022-03)
Asynchronous Reinforcement Learning for Real-Time Control of Physical Robots
and 1 more not named here

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

Training an agent to solve control tasks directly from high-dimensional images with model-free reinforcement learning (RL) has proven difficult. A promising approach is to learn a latent representation together with the control policy. However, fitting a high-capacity encoder using a scarce reward signal is sample inefficient and leads to poor performance. Prior work has shown that auxiliary losses, such as image reconstruction, can aid efficient representation learning. However, incorporating reconstruction loss into an off-policy learning algorithm often leads to training instability. We explore the underlying reasons and identify variational autoencoders, used by previous investigations, as the cause of the divergence. Following these findings, we propose effective techniques to improve training stability. This results in a simple approach capable of matching state-of-the-art model-free and model-based algorithms on MuJoCo control tasks. Furthermore, our approach demonstrates robustness to observational noise, surpassing existing approaches in this setting. Code, results, and videos are anonymously available at https://sites.google.com/view/sac-ae/home.

For agents

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

get_harvested_code_for_paper("1910.01741")
get_code_for_paper("1910.01741")
have("1910.01741")

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

curl https://syntology.ai/api/ran/1910.01741.json

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

[![Syntology run record](https://syntology.ai/api/ran/1910.01741.svg)](https://syntology.ai/paper/1910.01741)

Connect an agent — have() is free.