SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.08746 · 2021

Real-time Adversarial Perturbations against Deep Reinforcement Learning Policies: Attacks and Defenses

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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 4 have no contract verdict; 1 failure 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/2106.08746.json · badge: /api/ran/2106.08746.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
ssg-research/ad3-action-distribution-divergence-detector canonical (classifier label)
classifier basis: census
5 ran · 0 no recorded run
FunctionStatusWhere it lives
add_paddings Ran
ran on a synthesized input
ssg-research/ad3-action-distribution-divergence-detector/src/detection_module/DetectionModule.py
code served (permissive licence) · get_code("4c0653e7958792a3")
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 list, ndim 1, dtype family int; 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)

# Analyzing the function:
# - l: a list that gets extended with repeated content
# - content: a value to repeat and add to the list
# - width: an integer representing the target width

# The function extends list l with (width - len(l)) copies of content
# to make the list reach the desired width

l = [1, 2, 3]
content = 0
width = 7

ARGS = {
    "l": l,
    "content": content,
    "width": width
}
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 list, ndim 1, dtype family int — a list 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
init Ran
ran; our contract draft was wrong, not the code
ssg-research/ad3-action-distribution-divergence-detector/src/agents/models.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ArnaudFickinger/adversarial-surprise, is pointer only (licence: NONE) · get_code("9de355e93051e4ad")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from ArnaudFickinger/adversarial-surprise, 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.

parse_key Ran
ran on a synthesized input
ssg-research/ad3-action-distribution-divergence-detector/src/create_detection_script.py
code served (permissive licence) · get_code("4cadff05eab7dfcd")
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 1, 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

def parse_key(key_str):
    game, frames, percentile, alar_perc, queue_size = key_str.split('-')
    return (game, frames, percentile, alar_perc, queue_size)

key_str = "pong-1000-95-0.5-32"

ARGS = {
    "key_str": key_str
}
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 1, 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
post_process Ran
ran on a synthesized input
ssg-research/ad3-action-distribution-divergence-detector/src/agents/action_conditional_video_prediction.py
code served (permissive licence) · get_code("3ba7c8da77c3ab15")
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 Tensor, 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

torch.manual_seed(0)

# y is used in multiplication (y * 255.0), so it should be a tensor
y = torch.randn(2, 4)

# mean_obs is used in addition (... + mean_obs), so it should be a tensor
mean_obs = torch.randn(2, 4)

ARGS = {
    "y": y,
    "mean_obs": mean_obs
}
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 Tensor, 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
pre_process Ran
ran on a synthesized input
ssg-research/ad3-action-distribution-divergence-detector/src/agents/action_conditional_video_prediction.py
code served (permissive licence) · get_code("a39e3b63ff5b97d2")
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 Tensor, 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

torch.manual_seed(0)

# x: tensor to be normalized
x = torch.randn(2, 4)

# mean_obs: mean observation value to subtract
mean_obs = torch.randn(2, 4)

ARGS = {
    "x": x,
    "mean_obs": mean_obs,
}
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 Tensor, 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

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 1 later) and 2 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
init 34 other papers
14 dated earlier, 1 the same month, 19 later · 5 only as an assembly brick
1705.05363 (2017-05)
Curiosity-driven Exploration by Self-supervised Prediction
1706.02275 (2017-06)
Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
1810.12894 (2018-10)
Exploration by Random Network Distillation
and 31 more not named here
post_process 1 other paper
1 dated earlier
1907.09470 (2019-07)
Characterizing Attacks on Deep Reinforcement Learning
pre_process 1 other paper
1 dated earlier
1907.09470 (2019-07)
Characterizing Attacks on Deep Reinforcement Learning

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

Deep reinforcement learning (DRL) is vulnerable to adversarial perturbations. Adversaries can mislead the policies of DRL agents by perturbing the state of the environment observed by the agents. Existing attacks are feasible in principle, but face challenges in practice, either by being too slow to fool DRL policies in real time or by modifying past observations stored in the agent's memory. We show that Universal Adversarial Perturbations (UAP), independent of the individual inputs to which they are applied, can fool DRL policies effectively and in real time. We introduce three attack variants leveraging UAP. Via an extensive evaluation using three Atari 2600 games, we show that our attacks are effective, as they fully degrade the performance of three different DRL agents (up to 100%, even when the $l_\infty$ bound on the perturbation is as small as 0.01). It is faster than the frame rate (60 Hz) of image capture and considerably faster than prior attacks ($\approx 1.8$ms). Our attack technique is also efficient, incurring an online computational cost of $\approx 0.027$ms. Using two tasks involving robotic movement, we confirm that our results generalize to complex DRL tasks. Furthermore, we demonstrate that the effectiveness of known defenses diminishes against universal perturbations. We introduce an effective technique that detects all known adversarial perturbations against DRL policies, including all universal perturbations presented in this paper.

For agents

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

get_harvested_code_for_paper("2106.08746")
get_code_for_paper("2106.08746")
have("2106.08746")

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

curl https://syntology.ai/api/ran/2106.08746.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.08746.svg)](https://syntology.ai/paper/2106.08746)

Connect an agent — have() is free.