SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1906.03563 · 2019

Adversarial Attack Generation Empowered by Min-Max Optimization

arXiv · PDF · Open in the Atlas

Code that ran

11 of 25 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, 1 violated it and 10 have no contract verdict; 0 failures of our instrument; 14 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/1906.03563.json · badge: /api/ran/1906.03563.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
wangjksjtu/minmax-adv canonical (classifier label)
classifier basis: census
11 ran · 14 no recorded run
FunctionStatusWhere it lives
cw_loss Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/attacks/minmax.py
code served (permissive licence) · get_code("936c00f08008f28d")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

# Set random seeds for reproducibility
np.random.seed(0)
tf.random.set_seed(0)

# Create small example tensors
batch_size = 2
num_classes = 4

# labels: one-hot encoded labels of shape (batch_size, num_classes)
labels = tf.constant([
    [1.0, 0.0, 0.0, 0.0],
    [0.0, 1.0, 0.0, 0.0]
], dtype=tf.float32)

# logits: raw model outputs of shape (batch_size, num_classes)
logits = tf.constant([
    [2.5, 1.0, 0.5, 0.2],
    [0.1, 3.2, 1.5, 0.8]
], dtype=tf.float32)

# confidence: margin parameter
confidence = 50

ARGS = {
    "labels": labels,
    "logits": logits,
    "confidence": confidence
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 1, dtype family float — no output shape is stated: observe_signature.py recorded SymbolicTensor ndim 1, which disagrees with the fixture run
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_interval Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/utils.py
code served (permissive licence) · get_code("f572fa6a26eff62d")
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 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)

ARGS = {
    "eps_l": 0.1,
    "eps_r": 1.0,
    "num": 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 list, ndim 1, dtype family float — 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
get_models Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/ens_attack.py
code served (permissive licence) · get_code("8d68ad2ec3439b5c")
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 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

torch.manual_seed(0)

# Create a mock modelzoo object with model attributes
class MockModelZoo:
    def __init__(self):
        self.modelResNet = nn.Linear(4, 3)
        self.modelVGG = nn.Linear(4, 3)
        self.modelMobileNet = nn.Linear(4, 3)

modelzoo = MockModelZoo()

archs = ['ResNet', 'VGG', 'MobileNet']

ARGS = {
    'modelzoo': modelzoo,
    'archs': archs
}
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 float — 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
identity Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/utils.py
code served (permissive licence) · get_code("22ad888be68bfa1b")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 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 tensorflow as tf
import numpy as np

# The function takes images as input and converts it to a tensor, then returns its identity
# We need to provide a valid image input - can be numpy array, list, or tensor

# Create a small batch of images (batch_size=2, height=4, width=3, channels=3)
images_data = np.random.randn(2, 4, 3, 3).astype(np.float32)

ARGS = {
    "images": images_data
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 3) — shape (2, 4, 3, 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
image_folder_custom_label Ran
ran on a synthesized input
wangjksjtu/minmax-adv/pytorch/utils.py
code served (permissive licence) · get_code("83e7640948d1e7ea")
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 ImageFolder. 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 torchvision.transforms as transforms
import tempfile
import os
from PIL import Image

torch.manual_seed(0)

# Create a temporary directory structure with dummy images
temp_dir = tempfile.mkdtemp()
class_dirs = ['class_a', 'class_b', 'class_c']

for class_name in class_dirs:
    class_path = os.path.join(temp_dir, class_name)
    os.makedirs(class_path, exist_ok=True)
    # Create a small dummy image
    img = Image.new('RGB', (4, 3), color='red')
    img.save(os.path.join(class_path, 'dummy_0.jpg'))
    img = Image.new('RGB', (4, 3), color='blue')
    img.save(os.path.join(class_path, 'dummy_1.jpg'))

# Define the transform
transform = transforms.Compose([
    transforms.Resize((4, 3)),
    transforms.ToTensor(),
])

# Define idx2label mapping - maps indices to custom label names
idx2label = ['custom_label_0', 'custom_label_1', 'custom_label_2']

ARGS = {
    'root': temp_dir,
    'transform': transform,
    'idx2label': idx2label,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ImageFolder — a ImageFolder 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
l2_distance Ran
ran on a synthesized input
wangjksjtu/minmax-adv/pytorch/utils.py
code served (permissive licence) · get_code("e9de120c9fccad8d")
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 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# Build a minimal model that takes images and outputs logits
class SimpleModel(nn.Module):
    def __init__(self, num_classes=3):
        super().__init__()
        self.linear = nn.Linear(4, num_classes)
    
    def forward(self, x):
        # x shape: (batch, 4)
        return self.linear(x)

# Create model
model = SimpleModel(num_classes=3)

# Create tiny tensors
batch_size = 2
image_dim = 4

images = torch.randn(batch_size, image_dim)
adv_images = torch.randn(batch_size, image_dim)
labels = torch.tensor([0, 2])

ARGS = {
    'model': model,
    'images': images,
    'adv_images': adv_images,
    'labels': labels,
    'device': 'cpu'
}
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 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
make_ens_minmax Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/ens_attack.py
code served (permissive licence) · get_code("ff4af237f270b266")
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 numpy as np

class MockSession:
    """Mock TensorFlow session that simulates sess.run()"""
    def __init__(self, K):
        self.K = K
    
    def run(self, fetches, feed_dict):
        # Extract batch size from feed_dict
        x_batch = feed_dict.get('env.x')
        batch_size = x_batch.shape[0] if x_batch is not None else 2
        
        # Return mock adversarial examples and weights
        x_adv = x_batch + np.random.randn(*x_batch.shape) * 0.01
        W = np.random.randn(batch_size, self.K)
        
        return (x_adv, W)

class MockEnv:
    """Mock environment with required attributes for the function"""
    def __init__(self, K):
        self.K = K
        # Symbolic names for tensors (used in sess.run as fetches)
        self.x_mm_ens = 'x_mm_ens_tensor'
        self.W = 'W_tensor'
        # Placeholder names for feed_dict keys
        self.x = 'env.x'
        self.ens_mm_eps = 'env.ens_mm_eps'
        self.ens_mm_epochs = 'env.ens_mm_epochs'

# Setup parameters
K = 3
n_samples = 10
X_data = np.random.randn(n_samples, 4).astype(np.float32)
mock_env = MockEnv(K)
mock_sess = MockSession(K)

ARGS = {
    'sess': mock_sess,
    'env': mock_env,
    'X_data': X_data,
    'K': K,
    'epochs': 2,
    'eps': 0.01,
    'batch_size': 5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
make_minmax_eot Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/trans_attack.py
code served (permissive licence) · get_code("445cb63b8dd7ebb7")
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 (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 numpy as np

class MockEnv:
    def __init__(self):
        self.x = "x_placeholder"
        self.trans_mm_eps = "eps_placeholder"
        self.trans_mm_epochs = "epochs_placeholder"
        self.x_mm_trans = np.random.randn(2, 4)
        self.delta = np.random.randn(2, 4)
        self.W = np.random.randn(10, 4)

class MockSession:
    def __init__(self, env):
        self.env = env
    
    def run(self, fetches, feed_dict=None):
        x_mm_trans = self.env.x_mm_trans
        delta = self.env.delta
        W = self.env.W
        return [x_mm_trans, delta, W]

X_data = np.random.randn(6, 4).astype(np.float32)
env = MockEnv()
sess = MockSession(env)

ARGS = {
    "sess": sess,
    "env": env,
    "X_data": X_data,
    "epochs": 1,
    "eps": 0.01,
    "batch_size": 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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 (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
proj_box_appro Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/attacks/utils.py
code served (permissive licence) · get_code("24a5cabdc0b7b91c")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 2) 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 numpy as np
import tensorflow as tf

# Create a small 4D tensor with shape (batch=2, height=4, width=3, channels=2)
# This matches the expected input format for the function
batch_size = 2
height = 4
width = 3
channels = 2

v = tf.random.normal([batch_size, height, width, channels])

# Maximum norm allowed
eps = 1.5

# L_p norm to use (testing with p=2)
p = 2

ARGS = {
    "v": v,
    "eps": eps,
    "p": p
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 2) — shape (2, 4, 3, 2) 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
proj_prob_simplex Ran
ran on a synthesized input
wangjksjtu/minmax-adv/neurips21/attacks/utils.py
code served (permissive licence) · get_code("964d413db2fdd014")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)

# Create small test tensors
batch_size = 2
K = 4

# Create a small W tensor with shape (batch_size, K)
# Initialize with random values that make sense for the simplex projection
W = tf.constant([
    [0.5, 0.3, 0.2, 0.1],
    [0.4, 0.2, 0.3, 0.15]
], dtype=tf.float32)

ARGS = {
    'W': W,
    'batch_size': batch_size,
    'K': K
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
str2bool Ran
ran, violated the contract we drafted
wangjksjtu/minmax-adv/neurips21/utils.py
code served (permissive licence) · get_code("f017532fc389cbfe")
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 argparse

def str2bool(v):
    if v.lower() in ('yes', 'true', 't', 'y', '1'):
        return True
    elif v.lower() in ('no', 'false', 'f', 'n', '0'):
        return False
    else:
        raise argparse.ArgumentTypeError('Boolean value expected.')

# Setup code to construct valid arguments
ARGS = {
    'v': 'true'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 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
evaluate Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/ens_attack.py
code served (permissive licence) · get_code("213267d02b7820b8")
How it ran

no invocation recorded.

evaluate Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/trans_attack.py
code served (permissive licence) · get_code("e607dfc06f869a5d")
How it ran

no invocation recorded.

evaluate Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/uni_attack.py
code served (permissive licence) · get_code("247982bef8192964")
How it ran

no invocation recorded.

fgm Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/attacks/fast_gradient.py
code served (permissive licence) · get_code("3ed5f9dfac95521e")
How it ran

no invocation recorded.

fgmt Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/attacks/fast_gradient.py
code served (permissive licence) · get_code("cd6509a1aade295c")
How it ran

no invocation recorded.

get_model Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/pytorch/ens_attack.py
code served (permissive licence) · get_code("1d6df60e31ef00c5")
How it ran

no invocation recorded.

make_uni_minmax Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/uni_attack.py
code served (permissive licence) · get_code("232ff7a5fe7a603e")
How it ran

no invocation recorded.

minmax_ens Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/attacks/minmax.py
code served (permissive licence) · get_code("647f2c2e24c1966b")
How it ran

no invocation recorded.

minmax_uni Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/attacks/minmax.py
code served (permissive licence) · get_code("8b57098358df5568")
How it ran

no invocation recorded.

modelA Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/models.py
code served (permissive licence) · get_code("d8f5f464cb0707b2")
How it ran

no invocation recorded.

modelB Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/models.py
code served (permissive licence) · get_code("d9e789eb6c5505bc")
How it ran

no invocation recorded.

modelC Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/models.py
code served (permissive licence) · get_code("4e8ba12a7a587e76")
How it ran

no invocation recorded.

pgd Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/attacks/pgd.py
code served (permissive licence) · get_code("fed551f077ea431d")
How it ran

no invocation recorded.

proj_box Not yet run
unverified (harvested, no recorded run)
wangjksjtu/minmax-adv/neurips21/attacks/utils.py
code served (permissive licence) · get_code("8bf68b71ad3a0000")
How it ran

no invocation recorded.

Also present in other papers' code

Of 25 samples lifted from this paper's linked repositories, 25 carry a function identity and 0 do not. Of the 25 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 2 later) and 23 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
str2bool 82 other papers
16 dated earlier, 1 the same month, 65 later
1510.03055 (2015-10)
A Diversity-Promoting Objective Function for Neural Conversation Models
1703.06490 (2017-03)
Generating Multi-label Discrete Patient Records using Generative Adversarial Networks
1710.10467 (2017-10)
Generalized End-to-End Loss for Speaker Verification
and 79 more not named here
image_folder_custom_label 1 other paper
1 later
2110.10444 (2021-10)
Moiré Attack (MA): A New Potential Risk of Screen Photos

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

The worst-case training principle that minimizes the maximal adversarial loss, also known as adversarial training (AT), has shown to be a state-of-the-art approach for enhancing adversarial robustness. Nevertheless, min-max optimization beyond the purpose of AT has not been rigorously explored in the adversarial context. In this paper, we show how a general framework of min-max optimization over multiple domains can be leveraged to advance the design of different types of adversarial attacks. In particular, given a set of risk sources, minimizing the worst-case attack loss can be reformulated as a min-max problem by introducing domain weights that are maximized over the probability simplex of the domain set. We showcase this unified framework in three attack generation problems -- attacking model ensembles, devising universal perturbation under multiple inputs, and crafting attacks resilient to data transformations. Extensive experiments demonstrate that our approach leads to substantial attack improvement over the existing heuristic strategies as well as robustness improvement over state-of-the-art defense methods trained to be robust against multiple perturbation types. Furthermore, we find that the self-adjusted domain weights learned from our min-max framework can provide a holistic tool to explain the difficulty level of attack across domains. Code is available at https://github.com/wangjksjtu/minmax-adv.

For agents

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

get_harvested_code_for_paper("1906.03563")
get_code_for_paper("1906.03563")
have("1906.03563")

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

curl https://syntology.ai/api/ran/1906.03563.json

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

[![Syntology run record](https://syntology.ai/api/ran/1906.03563.svg)](https://syntology.ai/paper/1906.03563)

Connect an agent — have() is free.