SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.03471 · NeurIPS · 2024

Learning Action and Reasoning-Centric Image Editing from Videos and Simulations

Christopher Pal, Siva Reddy, Benno Krojer, Dheeraj Vattikonda, Luis Lara, Mila Jampani, Eva Portelance

arXiv · PDF · Open in the Atlas

Code that ran

6 of 7 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; 4 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 1 failure of our instrument; 1 no recorded run), measured 2026-11-21; 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/2407.03471.json · badge: /api/ran/2407.03471.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
McGill-NLP/AURORA canonical (classifier label)
classifier basis: census
4 ran · 1 no recorded run
OSU-NLP-Group/MagicBrush canonical (classifier label)
classifier basis: census
2 ran · 0 no recorded run
FunctionStatusWhere it lives
calculate_accuracy Ran
ran, honoured the contract we drafted
McGill-NLP/AURORA/disc_edit.py
code served (permissive licence) · get_code("92cfb056b8e8465c")
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 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
import numpy as np

torch.manual_seed(0)

# The function expects 'losses' which is iterable of sequences
# Each sequence should have at least 2 elements (first element compared to min of rest)
# Creating a list of lists where each inner list represents a loss vector

losses = [
    [0.5, 1.2, 0.8, 1.5],    # First element 0.5 < min(1.2, 0.8, 1.5) = 0.8, so correct
    [2.0, 0.3, 0.4, 0.2],    # First element 2.0 NOT < min(0.3, 0.4, 0.2) = 0.2, so incorrect
    [0.1, 0.9, 0.7, 1.1],    # First element 0.1 < min(0.9, 0.7, 1.1) = 0.7, so correct
]

ARGS = {
    "losses": losses
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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 int — 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
create_scene Ran
ran, honoured the contract we drafted
McGill-NLP/AURORA/eq-kubric/create_scene.py
code served (permissive licence) · get_code("111eff482132b5d8")
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 subprocess
import os
import tempfile

# Create a temporary directory for logs
log_dir = tempfile.mkdtemp()
log_file_path = os.path.join(log_dir, 'error_log.txt')

# Ensure the log file exists
with open(log_file_path, 'w') as f:
    f.write('')

ARGS = {
    'index': 0,
    'dataset_type': 'location',
    'max_num_objects': 1,
    'retries': 10,
    'log_file': log_file_path,
    'verbose': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
eval_clip_i Ran
ran, honoured the contract we drafted
OSU-NLP-Group/MagicBrush/evaluation/image_eval.py
pointer only (licence: NOASSERTION) · get_code("7068874c55737e99")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, OSU-NLP-Group/MagicBrush (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.

eval_distance Ran
ran, honoured the contract we drafted
OSU-NLP-Group/MagicBrush/evaluation/image_eval.py
pointer only (licence: NOASSERTION) · get_code("8d23b7e0bcc26acc")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, OSU-NLP-Group/MagicBrush (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_obj_x_left Ran
ran, violated the contract we drafted
McGill-NLP/AURORA/eq-kubric/my_kubric_twoframe_attribute.py
code served (permissive licence) · get_code("f0a030a3b1efb81b")
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 float, ndim 0, dtype family float, 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 numpy as np

torch.manual_seed(0)

# The function accesses bound[0][0] and scale[0]
# bound appears to be a 2D array-like structure
# scale appears to be a 1D array-like structure

bound = [[5.0, 3.0], [2.0, 1.0]]  # 2D list, accessing [0][0] gives 5.0
scale = [2.0, 1.5]  # 1D list, accessing [0] gives 2.0

ARGS = {
    "bound": bound,
    "scale": scale
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, 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
update_progress_bar Ran
ran; our contract draft was wrong, not the code
McGill-NLP/AURORA/eq-kubric/create_dataset.py
code served (permissive licence) · get_code("61da83f7deb73152")
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 function. 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
from unittest.mock import MagicMock

torch.manual_seed(0)

# Create a mock progress bar object that has an update() method
mock_bar = MagicMock()
mock_bar.update = MagicMock()

ARGS = {
    "bar": mock_bar
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
calculate_clip_similarity Not yet run
unverified (harvested, no recorded run)
McGill-NLP/AURORA/disc_edit.py
code served (permissive licence) · get_code("99305f6b90dd8d54")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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 7 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 2026-11-21.

Repositories linked to this paper

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

Abstract

An image editing model should be able to perform diverse edits, ranging from object replacement, changing attributes or style, to performing actions or movement, which require many forms of reasoning. Current general instruction-guided editing models have significant shortcomings with action and reasoning-centric edits. Object, attribute or stylistic changes can be learned from visually static datasets. On the other hand, high-quality data for action and reasoning-centric edits is scarce and has to come from entirely different sources that cover e.g. physical dynamics, temporality and spatial reasoning. To this end, we meticulously curate the AURORA Dataset (Action-Reasoning-Object-Attribute), a collection of highquality training data, human-annotated and curated from videos and simulation engines. We focus on a key aspect of quality training data: triplets (source image, prompt, target image) contain a single meaningful visual change described by the prompt, i.e., truly minimal changes between source and target images. To demonstrate the value of our dataset, we evaluate an AURORA-finetuned model on a new expert-curated benchmark (AURORA-BENCH) covering 8 diverse editing tasks. Our model significantly outperforms previous editing models as judged by human raters. For automatic evaluations, we find important flaws in previous metrics and caution their use for semantically hard editing tasks. Instead, we propose a new automatic metric that focuses on discriminative understanding. We hope that our efforts : (1) curating a quality training dataset and an evaluation benchmark, (2) developing critical evaluations, and (3) releasing a state-of-the-art model 1 , will fuel further progress on general image editing.

For agents

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

get_harvested_code_for_paper("2407.03471")
get_code_for_paper("2407.03471")
have("2407.03471")

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

curl https://syntology.ai/api/ran/2407.03471.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.03471.svg)](https://syntology.ai/paper/2407.03471)

Connect an agent — have() is free.