SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.01801 · 2022

Safe Reinforcement Learning From Pixels Using a Stochastic Latent Representation

arXiv · PDF · Open in the Atlas

Code that ran

3 of 4 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; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 0 failures of our instrument; 1 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/2210.01801.json · badge: /api/ran/2210.01801.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
safe-slac/safe-slac canonical (classifier label)
classifier basis not recorded on the link
3 ran · 1 no recorded run
FunctionStatusWhere it lives
build_mlp Ran
ran on a synthesized input
safe-slac/safe-slac/slac/utils.py
code served (permissive licence) · get_code("bcd2f30df5201848")
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 Sequential. 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)

ARGS = {
    'input_dim': 4,
    'output_dim': 3,
    'hidden_units': [64, 64],
    'hidden_activation': nn.Tanh(),
    'output_activation': None
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
create_feature_actions Ran
ran on a synthesized input
safe-slac/safe-slac/slac/utils.py
code served (permissive licence) · get_code("c17d4065283d4027")
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 tuple, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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)

# The function expects two tensors: feature_ and action_
# Looking at the function body:
# - feature_ is indexed with [:, :-1] and [:, 1:], so it needs at least 2 dimensions
# - action_ is indexed with [:, :-1] and [:, 1:], so it needs at least 2 dimensions
# - Both are reshaped to (N, -1) where N is the batch size
# - They are concatenated along dim=-1

# Create small tensors:
# feature_: shape (2, 5, 4) - batch size 2, sequence length 5, feature dim 4
# action_: shape (2, 5, 3) - batch size 2, sequence length 5, action dim 3

feature_ = torch.randn(2, 5, 4)
action_ = torch.randn(2, 5, 3)

ARGS = {
    "feature_": feature_,
    "action_": action_
}
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 tuple, ndim 3, dtype family float — a tuple output has no array shape of its own; none is stated
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
sample_reproduction Ran
ran on a synthesized input
safe-slac/safe-slac/slac/utils.py
code served (permissive licence) · get_code("f958d15bc44a28f7")
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 ndarray, ndim 2, dtype family float, shape (2, 4) 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

torch.manual_seed(0)

class SimpleEncoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 8)
    
    def forward(self, state):
        return self.linear(state)

class SimpleDecoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(6, 4)
    
    def forward(self, z):
        decoded = self.linear(z)
        return decoded, torch.ones_like(decoded) * 0.1

class LatentModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.encoder = SimpleEncoder()
        self.decoder = SimpleDecoder()
    
    def sample_posterior(self, feature, action):
        batch_size = feature.shape[0]
        z1_mean = torch.randn(batch_size, 3)
        z1_std = torch.ones(batch_size, 3) * 0.5
        z1 = z1_mean + z1_std * torch.randn(batch_size, 3)
        z2 = torch.randn(batch_size, 3)
        return z1_mean, z1_std, z1, z2
    
    def sample_prior(self, action, z2):
        batch_size = z2.shape[0]
        z1_mean = torch.randn(batch_size, 3)
        z1_std = torch.ones(batch_size, 3) * 0.5
        return z1_mean, z1_std

latent_model = LatentModel()
device = 'cpu'
state = [[1, 2, 3, 4], [5, 6, 7, 8]]
action = [[0.1, 0.2], [0.3, 0.4]]

ARGS = {
    'latent_model': latent_model,
    'device': device,
    'state': state,
    'action': action
}
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 ndarray, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) 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_safety Not yet run
unverified (harvested, no recorded run)
safe-slac/safe-slac/slac/env.py
code served (permissive licence) · get_code("c63d0b541a34fef9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 0 later) and 0 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
build_mlp 1 other paper
1 dated earlier
1907.00953 (2019-07)
Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent Variable Model
create_feature_actions 1 other paper
1 dated earlier
1907.00953 (2019-07)
Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent Variable Model
make_safety 1 other paper
1 dated earlier
1907.00953 (2019-07)
Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent Variable Model
sample_reproduction 1 other paper
1 dated earlier
1907.00953 (2019-07)
Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent Variable Model

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

We address the problem of safe reinforcement learning from pixel observations. Inherent challenges in such settings are (1) a trade-off between reward optimization and adhering to safety constraints, (2) partial observability, and (3) high-dimensional observations. We formalize the problem in a constrained, partially observable Markov decision process framework, where an agent obtains distinct reward and safety signals. To address the curse of dimensionality, we employ a novel safety critic using the stochastic latent actor-critic (SLAC) approach. The latent variable model predicts rewards and safety violations, and we use the safety critic to train safe policies. Using well-known benchmark environments, we demonstrate competitive performance over existing approaches with respects to computational requirements, final reward return, and satisfying the safety constraints.

For agents

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

get_harvested_code_for_paper("2210.01801")
get_code_for_paper("2210.01801")
have("2210.01801")

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

curl https://syntology.ai/api/ran/2210.01801.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.01801.svg)](https://syntology.ai/paper/2210.01801)

Connect an agent — have() is free.