SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.12920 · 2020

Automatic Perturbation Analysis for Scalable Certified Robustness and Beyond

arXiv · PDF · Open in the Atlas

Code that ran

14 of 18 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, 3 violated it and 9 have no contract verdict; 2 failures of our instrument; 4 no recorded run), measured 2026-11-21; 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/2002.12920.json · badge: /api/ran/2002.12920.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
huanzhang12/CROWN-IBP canonical (classifier label)
classifier basis: census
14 ran · 4 no recorded run
FunctionStatusWhere it lives
argparser Ran
ran; our contract draft was wrong, not the code
huanzhang12/CROWN-IBP/argparser.py
pointer only (licence: BSD-2-Clause); the stored copy get_code serves, from SliencerX/Belief-enriched-robust-Q-learning, is pointer only (licence: NONE) · get_code("62000ff4e804fe8d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-2-Clause), but get_code serves the stored copy, first harvested from SliencerX/Belief-enriched-robust-Q-learning, 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.

batch Ran
ran; our contract draft was wrong, not the code
huanzhang12/CROWN-IBP/convex_adversarial/dual_layers.py
code served (permissive licence) · get_code("778274b3fb463e7e")
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 4, dtype family float, shape (2, 4, 4, 3) 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)

# The function batch(A, n) reshapes tensor A into shape (n, -1, *A.size()[1:])
# It needs:
# - A: a torch tensor with at least 1 dimension
# - n: an integer that divides the first dimension of A

# Create a small tensor with distinct dimensions
# Shape: (8, 4, 3) - batch size 8, feature dim 4, another dim 3
A = torch.randn(8, 4, 3)

# n should divide the first dimension (8), let's use n=2
# This will reshape to (2, -1, 4, 3) = (2, 4, 4, 3)
n = 2

ARGS = {
    "A": A,
    "n": n
}
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 4, dtype family float, shape (2, 4, 4, 3) — shape (2, 4, 4, 3) 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
get_file_close Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/config.py
code served (permissive licence) · get_code("03ba17f5b2a279b2")
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 str. 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 and a test file for the function to find
temp_dir = tempfile.mkdtemp()
test_filename = os.path.join(temp_dir, 'testfile')

# Create an actual file with the extension so glob can find it
test_file_path = test_filename + '.pt'
with open(test_file_path, 'w') as f:
    f.write('test')

ARGS = {
    'filename': test_filename,
    'ext': 'pt',
    'load': True
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
get_model_meta_real Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/converter/mnist_cifar_models.py
code served (permissive licence) · get_code("63f0738b7f84b19d")
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. 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 tensorflow as tf
import json

# Create a minimal TensorFlow model that can be serialized and deserialized
def create_test_model():
    """Create a simple Sequential model with Dense layers and activations"""
    model = tf.keras.Sequential([
        tf.keras.layers.Flatten(input_shape=(2, 4)),
        tf.keras.layers.Dense(8, activation='relu'),
        tf.keras.layers.Dense(5, activation='relu'),
        tf.keras.layers.Dense(3, activation='linear')
    ])
    return model

# Create the model and save its JSON representation
test_model = create_test_model()

# Define a simple model_loader function that loads from JSON
def model_loader(filename, custom_objects=None):
    """Mock model loader that returns our test model"""
    return test_model

# Prepare filename (not actually used since we're not doing file I/O)
filename = "dummy_model.h5"

# Construct ARGS dictionary
ARGS = {
    'filename': filename,
    'model_loader': model_loader
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — 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
get_stats Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/datasets.py
code served (permissive licence) · get_code("1692f72216960bf3")
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 2, 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)

# Create a simple data loader
# The function expects an iterable of (images, labels) tuples
# It accesses: images.size(0), images.size(1), images.size(2), images.size(3)
# and iterates twice over loader, accessing len(loader.dataset)

class SimpleDataset:
    def __init__(self, num_samples=2):
        self.num_samples = num_samples
        # Create images: (batch_size, channels, height, width) = (2, 3, 4, 5)
        self.images = torch.randn(num_samples, 3, 4, 5)
        self.labels = torch.zeros(num_samples)
    
    def __len__(self):
        return self.num_samples
    
    def __getitem__(self, idx):
        return self.images[idx:idx+1], self.labels[idx:idx+1]

class SimpleDataLoader:
    def __init__(self, dataset, batch_size=1):
        self.dataset = dataset
        self.batch_size = batch_size
    
    def __iter__(self):
        for i in range(len(self.dataset)):
            yield self.dataset[i]
    
    def __len__(self):
        return len(self.dataset)

# Create dataset and loader
dataset = SimpleDataset(num_samples=2)
loader = SimpleDataLoader(dataset, batch_size=1)

ARGS = {
    "loader": loader
}
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 2, 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
isfloat Ran
ran, violated the contract we drafted
huanzhang12/CROWN-IBP/argparser.py
code served (permissive licence) · get_code("2f8e524a282482b9")
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 bool, 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

torch.manual_seed(0)

ARGS = {
    "value": "3.14"
}
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 bool, 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
isint Ran
ran, violated the contract we drafted
huanzhang12/CROWN-IBP/argparser.py
pointer only (licence: BSD-2-Clause); the stored copy get_code serves, from SliencerX/Belief-enriched-robust-Q-learning, is pointer only (licence: NONE) · get_code("10119e52ad9b6cf6")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-2-Clause), but get_code serves the stored copy, first harvested from SliencerX/Belief-enriched-robust-Q-learning, 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.

load_config Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/config.py
code served (permissive licence) · get_code("6fd34470e7a5aa5d")
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 dict. 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 json
import tempfile
import os

# Create temporary directory for test files
tmpdir = tempfile.mkdtemp()

# Create defaults.json
defaults_config = {
    "models": [
        {"name": "model1", "param": "value1"},
        {"name": "model2", "param": "value2"}
    ],
    "training": {"lr": 0.001},
    "path_prefix": "/default/path"
}

defaults_path = os.path.join(tmpdir, "defaults.json")
with open(defaults_path, "w") as f:
    json.dump(defaults_config, f)

# Create a test config file
test_config = {
    "models": [
        {"name": "model1", "param": "custom_value"},
        {"name": "model3", "repeats": 2}
    ],
    "training": {"lr": 0.01}
}

config_path = os.path.join(tmpdir, "test_config.json")
with open(config_path, "w") as f:
    json.dump(test_config, f)

# Change to tmpdir so defaults.json can be found
original_dir = os.getcwd()
os.chdir(tmpdir)

# Create a minimal args object
class Args:
    def __init__(self, config, overrides_dict=None, model_subset=None, path_prefix=None):
        self.config = config
        self.overrides_dict = overrides_dict
        self.model_subset = model_subset
        self.path_prefix = path_prefix

ARGS = {
    "args": Args(
        config=config_path,
        overrides_dict={"training": {"lr": 0.02}},
        model_subset=[0, 1],
        path_prefix="/custom/path"
    )
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
mnist_loaders Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/datasets.py
code served (permissive licence) · get_code("5a918baac3f47e20")
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
import torch.nn as nn
from torch.utils import data
import torchvision.transforms as transforms

torch.manual_seed(0)

class MockDataset:
    """Mock dataset that mimics torchvision MNIST dataset interface"""
    def __init__(self, root, train=True, download=False, transform=None):
        self.root = root
        self.train = train
        self.download = download
        self.transform = transform
        self.data = torch.randn(10, 28, 28)
        self.targets = torch.randint(0, 10, (10,))
    
    def __len__(self):
        return len(self.data)
    
    def __getitem__(self, idx):
        x = self.data[idx].unsqueeze(0)
        y = self.targets[idx].item()
        if self.transform:
            x = self.transform(x.squeeze(0))
        return x, y

ARGS = {
    'dataset': MockDataset,
    'batch_size': 2,
    'shuffle_train': True,
    'shuffle_test': False,
    'normalize_input': False,
    'num_examples': 4,
    'test_batch_size': 2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — 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
model_cnn_1layer Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/model_defs.py
code served (permissive licence) · get_code("c04566214aef8894")
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)

class Flatten(nn.Module):
    def forward(self, x):
        return x.view(x.size(0), -1)

ARGS = {
    "in_ch": 3,
    "in_dim": 16,
    "width": 2,
}
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
model_mlp_any Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/model_defs.py
code served (permissive licence) · get_code("536b44e648824359")
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 = {
    "in_dim": 16,
    "neurons": [32, 64, 128],
    "out_dim": 10
}
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
model_mlp_uniform Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/model_defs.py
code served (permissive licence) · get_code("1d4d589798bf1c33")
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 = {
    "in_dim": 16,
    "layer": 3,
    "neurons": 32,
    "out_dim": 10
}
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
unbatch Ran
ran, violated the contract we drafted
huanzhang12/CROWN-IBP/convex_adversarial/dual_layers.py
code served (permissive licence) · get_code("9efb427bc4cdf3df")
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 (6, 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)

# The unbatch function takes a tensor A and reshapes it
# It flattens the first dimension and keeps all subsequent dimensions
# A.view(-1, *A.size()[2:]) takes batch and sequence dims and flattens them

# Create a small tensor with shape (batch=2, seq=3, features=4)
# This way we can see the unbatching effect clearly
A = torch.randn(2, 3, 4)

ARGS = {
    "A": A
}
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 (6, 4) — shape (6, 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
update_dict Ran
ran on a synthesized input
huanzhang12/CROWN-IBP/config.py
code served (permissive licence) · get_code("346b1abd4961b091")
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 dict. 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

# Base dictionary with some config options
d = {
    'learning_rate': 0.001,
    'batch_size': 32,
    'model': {
        'hidden_dim': 128,
        'num_layers': 2
    },
    'optimizer': 'adam'
}

# Update dictionary with new values, including nested dicts and new keys
u = {
    'learning_rate': 0.0001,
    'model': {
        'hidden_dim': 256,
        'dropout': 0.5
    },
    'new_key': 'new_value'
}

# show_warning flag
show_warning = True

ARGS = {
    'd': d,
    'u': u,
    'show_warning': show_warning
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
cifar_loaders Not yet run
unverified (harvested, no recorded run)
huanzhang12/CROWN-IBP/datasets.py
code served (permissive licence) · get_code("c41b33e06e9006fb")
How it ran

no invocation recorded.

get_model_meta Not yet run
unverified (harvested, no recorded run)
huanzhang12/CROWN-IBP/converter/mnist_cifar_models.py
code served (permissive licence) · get_code("368ba58e4dcc5a2a")
How it ran

no invocation recorded.

robust_loss Not yet run
unverified (harvested, no recorded run)
huanzhang12/CROWN-IBP/convex_adversarial/dual_network.py
code served (permissive licence) · get_code("23b6aa5bec1fc235")
How it ran

no invocation recorded.

robust_loss_parallel Not yet run
unverified (harvested, no recorded run)
huanzhang12/CROWN-IBP/convex_adversarial/dual_network.py
code served (permissive licence) · get_code("1cc8bc10c156fe29")
How it ran

no invocation recorded.

Also present in other papers' code

Of 18 samples lifted from this paper's linked repositories, 18 carry a function identity and 0 do not. Of the 18 distinct functions with an identity, 13 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 9 later) and 5 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
isfloat 4 other papers
1 dated earlier, 3 later
1712.07629 (2017-12)
SuperPoint: Self-Supervised Interest Point Detection and Description
2008.06081 (2020-08)
Adversarial Training and Provable Robustness: A Tale of Two Objectives
2105.02486 (2021-05)
Towards General Natural Language Understanding with Probabilistic Worldbuilding
and 1 more not named here
batch 2 other papers
2 dated earlier
1711.00851 (2017-11)
Provable Defenses against Adversarial Examples via the Convex Outer Adversarial Polytope
1912.01329 (2019-12)
Neural Network Branching for Neural Network Verification
isint 2 other papers
2 later
2008.06081 (2020-08)
Adversarial Training and Provable Robustness: A Tale of Two Objectives
2403.04050 (2024-03)
Belief-Enriched Pessimistic Q-Learning against Adversarial State Perturbations
unbatch 2 other papers
2 dated earlier
1711.00851 (2017-11)
Provable Defenses against Adversarial Examples via the Convex Outer Adversarial Polytope
1912.01329 (2019-12)
Neural Network Branching for Neural Network Verification
argparser 1 other paper
1 later
2403.04050 (2024-03)
Belief-Enriched Pessimistic Q-Learning against Adversarial State Perturbations
cifar_loaders 1 other paper
1 later
2008.06081 (2020-08)
Adversarial Training and Provable Robustness: A Tale of Two Objectives
get_model_meta_real 1 other paper
1 dated earlier
1811.00866 (2018-11)
Efficient Neural Network Robustness Certification with General Activation Functions
get_stats 1 other paper
1 later
2008.06081 (2020-08)
Adversarial Training and Provable Robustness: A Tale of Two Objectives
mnist_loaders 1 other paper
1 later
2008.06081 (2020-08)
Adversarial Training and Provable Robustness: A Tale of Two Objectives
model_cnn_1layer 1 other paper
1 later
2006.04924 (2020-06)
A Self-supervised Approach for Adversarial Robustness
model_mlp_any 1 other paper
1 later
2006.04924 (2020-06)
A Self-supervised Approach for Adversarial Robustness
model_mlp_uniform 1 other paper
1 later
2006.04924 (2020-06)
A Self-supervised Approach for Adversarial Robustness
robust_loss_parallel 1 other paper
1 dated earlier
1912.01329 (2019-12)
Neural Network Branching for Neural Network Verification

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

Linear relaxation based perturbation analysis (LiRPA) for neural networks, which computes provable linear bounds of output neurons given a certain amount of input perturbation, has become a core component in robustness verification and certified defense. The majority of LiRPA-based methods focus on simple feed-forward networks and need particular manual derivations and implementations when extended to other architectures. In this paper, we develop an automatic framework to enable perturbation analysis on any neural network structures, by generalizing existing LiRPA algorithms such as CROWN to operate on general computational graphs. The flexibility, differentiability and ease of use of our framework allow us to obtain state-of-the-art results on LiRPA based certified defense on fairly complicated networks like DenseNet, ResNeXt and Transformer that are not supported by prior works. Our framework also enables loss fusion, a technique that significantly reduces the computational complexity of LiRPA for certified defense. For the first time, we demonstrate LiRPA based certified defense on Tiny ImageNet and Downscaled ImageNet where previous approaches cannot scale to due to the relatively large number of classes. Our work also yields an open-source library for the community to apply LiRPA to areas beyond certified defense without much LiRPA expertise, e.g., we create a neural network with a probably flat optimization landscape by applying LiRPA to network parameters. Our opensource library is available at https://github.com/KaidiXu/auto_LiRPA.

For agents

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

get_harvested_code_for_paper("2002.12920")
get_code_for_paper("2002.12920")
have("2002.12920")

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

curl https://syntology.ai/api/ran/2002.12920.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.12920.svg)](https://syntology.ai/paper/2002.12920)

Connect an agent — have() is free.