SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2301.01197 · 2023

Backdoor Attacks Against Dataset Distillation

arXiv · PDF · Open in the Atlas

Code that ran

13 of 15 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; 1 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 5 failures of our instrument; 2 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/2301.01197.json · badge: /api/ran/2301.01197.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
liuyugeng/baadd canonical (classifier label)
classifier basis not recorded on the link
13 ran · 2 no recorded run
FunctionStatusWhere it lives
VGG11 Ran
ran; our contract draft was wrong, not the code
liuyugeng/baadd/DC/networks.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from BigML-CS-UCLA/MKDT, is pointer only (licence: NONE) · get_code("190dfb67aacf4767")
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 BigML-CS-UCLA/MKDT, 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.

VGG11BN Ran
ran; our contract draft was wrong, not the code
liuyugeng/baadd/DC/networks.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from BigML-CS-UCLA/MKDT, is pointer only (licence: NONE) · get_code("6fa9b9b022f6dbfa")
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 BigML-CS-UCLA/MKDT, 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.

VGG13 Ran
ran; our contract draft was wrong, not the code
liuyugeng/baadd/DC/networks.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from BigML-CS-UCLA/MKDT, is pointer only (licence: NONE) · get_code("3da7c8c828c4f0af")
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 BigML-CS-UCLA/MKDT, 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.

clone_tuple Ran
ran on a synthesized input
liuyugeng/baadd/DD/networks/utils.py
code served (permissive licence) · get_code("9ff25651ae35371b")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create a tuple of tensors with different requires_grad settings
tensor1 = torch.randn(2, 4, requires_grad=True)
tensor2 = torch.randn(2, 4, requires_grad=False)
tensor3 = torch.randn(2, 4, requires_grad=True)

tensors = (tensor1, tensor2, tensor3)

ARGS = {
    "tensors": tensors,
    "requires_grad": 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 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
conv1x1 Ran
ran; our contract draft was wrong, not the code
liuyugeng/baadd/DD/networks/ResNet.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from medaitech/scrd4ad, is pointer only (licence: NONE) · get_code("2a80220dabcb742a")
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 medaitech/scrd4ad, 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.

conv3x3 Ran
ran; our contract draft was wrong, not the code
liuyugeng/baadd/DD/networks/ResNet.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from medaitech/scrd4ad, is pointer only (licence: NONE) · get_code("600ff2c45e0de056")
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 medaitech/scrd4ad, 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.

distance_wb Ran
ran, honoured the contract we drafted
liuyugeng/baadd/DC/utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from saehyung-lee/dcc, is pointer only (licence: NONE) · get_code("6e9c65aaa7c1b5da")
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 saehyung-lee/dcc, 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.

final_objective_loss Ran
ran on a synthesized input
liuyugeng/baadd/DD/basics.py
code served (permissive licence) · get_code("6fc97d2f9170a6d7")
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. 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
import torch.nn.functional as F

torch.manual_seed(0)

# Create a minimal state object with required attributes
class State:
    def __init__(self, num_classes, mode, attack_class=None, target_class=None):
        self.num_classes = num_classes
        self.mode = mode
        self.attack_class = attack_class
        self.target_class = target_class

# Test case 1: Binary classification with distill_basic mode
state_binary = State(num_classes=2, mode='distill_basic')
output_binary = torch.randn(2, 1)  # batch_size=2, single output for binary
label_binary = torch.randint(0, 2, (2, 1)).float()  # binary labels

# Test case 2: Multi-class with distill_basic mode
state_multiclass = State(num_classes=5, mode='distill_basic')
output_multiclass = torch.randn(2, 5)  # batch_size=2, 5 classes
label_multiclass = torch.randint(0, 5, (2,))  # class labels

# Test case 3: Multi-class with distill_attack mode
state_attack = State(num_classes=5, mode='distill_attack', attack_class=1, target_class=2)
output_attack = torch.randn(2, 5)
label_attack = torch.randint(0, 5, (2,))

ARGS = {
    'state': state_multiclass,
    'output': output_multiclass,
    'label': label_multiclass
}
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
init_weights Ran
ran on a synthesized input
liuyugeng/baadd/DD/networks/utils.py
code served (permissive licence) · get_code("074d873dd1b4ae2b")
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 AlexNet. 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
from torch.nn import init

torch.manual_seed(0)

# Create a minimal state object with required attributes
class State:
    def __init__(self, init_type, init_param, shadow=False):
        self.init = init_type
        self.init_param = init_param
        self.shadow = shadow

# Create a minimal AlexNet-like model for testing
class AlexNet(nn.Module):
    def __init__(self):
        super(AlexNet, self).__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 8, kernel_size=3),
            nn.ReLU(inplace=True),
        )
        self.classifier = nn.Sequential(
            nn.Linear(8 * 30 * 30, 64),
            nn.ReLU(inplace=True),
            nn.Linear(64, 32),
            nn.ReLU(inplace=True),
            nn.Linear(32, 16),
            nn.ReLU(inplace=True),
            nn.Linear(16, 8),
            nn.ReLU(inplace=True),
            nn.Linear(8, 4),
            nn.ReLU(inplace=True),
            nn.Linear(4, 10),
        )
    
    def forward(self, x):
        x = self.features(x)
        x = x.view(x.size(0), -1)
        x = self.classifier(x)
        return x

# Create a minimal ResNet18-like model
class ResNet18(nn.Module):
    def __init__(self):
        super(ResNet18, self).__init__()
        self.conv1 = nn.Conv2d(3, 8, kernel_size=3, padding=1)
        self.bn1 = nn.BatchNorm2d(8)
        self.fc = nn.Linear(8 * 32 * 32, 10)
    
    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.fc(x.view(x.size(0), -1))
        return x

# Create a minimal VGG11-like model
class VGG11(nn.Module):
    def __init__(self):
        super(VGG11, self).__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 8, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
        )
        self.classifier = nn.Sequential(
            nn.Linear(8 * 32 * 32, 64),
            nn.ReLU(inplace=True),
            nn.Linear(64, 32),
            nn.ReLU(inplace=True),
            nn.Linear(32, 16),
            nn.ReLU(inplace=True),
            nn.Linear(16, 8),
            nn.ReLU(inplace=True),
            nn.Linear(8, 4),
            nn.ReLU(inplace=True),
            nn.Linear(4, 10),
        )
    
    def forward(self, x):
        x = self.features(x)
        x = x.view(x.size(0), -1)
        x = self.classifier(x)
        return x

# Create a minimal AlexCifarNet-like model
class AlexCifarNet(nn.Module):
    def __init__(self):
        super(AlexCifarNet, self).__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 8, kernel_size=3),
            nn.ReLU(inplace=True),
        )
        self.classifier = nn.Sequential(
            nn.Linear(8 * 30 * 30, 64),
            nn.ReLU(inplace=True),
            nn.Linear(64, 32),
            nn.ReLU(inplace=True),
            nn.Linear(32, 16),
            nn.ReLU(inplace=True),
            nn.Linear(16, 8),
            nn.ReLU(inplace=True),
            nn.Linear(8, 10),
        )
    
    def forward(self, x):
        x = self.features(x)
        x = x.view(x.size(0), -1)
        x = self.classifier(x)
        return x

# Test with 'normal' init_type which doesn't load pretrained models
net = AlexNet()
state = State(init_type='normal', init_param=0.02, shadow=False)

ARGS = {
    'net': net,
    'state': state,
}
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 AlexNet — a AlexNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a AlexNet 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
permute_list Ran
ran on a synthesized input
liuyugeng/baadd/DD/train_distilled_image.py
code served (permissive licence) · get_code("12b94db554dfecb4")
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 numpy as np

# The function permute_list takes a list parameter and uses:
# - len(list) to get the length
# - list[i] indexing to access elements
# So we just need a regular Python list with some elements

test_list = [10, 20, 30, 40, 50]

ARGS = {
    "list": test_list
}
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 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
read_bndbox Ran
ran on a synthesized input
liuyugeng/baadd/DD/datasets/pascal_voc.py
code served (permissive licence) · get_code("fdee81a9e734615a")
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. 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 os
import os.path
import tempfile
import xml.etree.ElementTree as ET

# Create a temporary directory structure mimicking VOC dataset
temp_root = tempfile.mkdtemp()
voc_dir = os.path.join(temp_root, 'VOCdevkit', 'VOC2012', 'Annotations')
os.makedirs(voc_dir, exist_ok=True)

# Create sample XML annotation files
def create_xml_file(filepath, image_name, objects):
    root = ET.Element('annotation')
    
    filename_elem = ET.SubElement(root, 'filename')
    filename_elem.text = image_name
    
    for obj_name, xmin, ymin, xmax, ymax in objects:
        obj_elem = ET.SubElement(root, 'object')
        
        name_elem = ET.SubElement(obj_elem, 'name')
        name_elem.text = obj_name
        
        bndbox_elem = ET.SubElement(obj_elem, 'bndbox')
        
        xmin_elem = ET.SubElement(bndbox_elem, 'xmin')
        xmin_elem.text = str(xmin)
        
        ymin_elem = ET.SubElement(bndbox_elem, 'ymin')
        ymin_elem.text = str(ymin)
        
        xmax_elem = ET.SubElement(bndbox_elem, 'xmax')
        xmax_elem.text = str(xmax)
        
        ymax_elem = ET.SubElement(bndbox_elem, 'ymax')
        ymax_elem.text = str(ymax)
    
    tree = ET.ElementTree(root)
    tree.write(filepath)

# Create first XML file
create_xml_file(
    os.path.join(voc_dir, '000001.xml'),
    '000001.jpg',
    [('cat', 10, 20, 100, 150), ('dog', 50, 60, 120, 180)]
)

# Create second XML file
create_xml_file(
    os.path.join(voc_dir, '000002.xml'),
    '000002.jpg',
    [('person', 5, 10, 200, 300)]
)

ARGS = {
    'root': temp_root,
    'dataset': 'VOC2012',
    'paths': ['000001', '000002']
}
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 list — a list 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
read_split Ran
ran on a synthesized input
liuyugeng/baadd/DD/datasets/pascal_voc.py
code served (permissive licence) · get_code("26cdf3966b9542d6")
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 os
import tempfile
import shutil

# Create a temporary directory structure that mimics VOC dataset layout
temp_dir = tempfile.mkdtemp()
voc_root = os.path.join(temp_dir, 'VOCdevkit', 'VOC2012', 'ImageSets', 'Main')
os.makedirs(voc_root, exist_ok=True)

# Create a sample split file for the aeroplane category
split_file = os.path.join(voc_root, 'aeroplane_train.txt')
with open(split_file, 'w') as f:
    f.write('000001 1\n')
    f.write('000002 -1\n')
    f.write('000003 1\n')

ARGS = {
    'root': temp_dir,
    'dataset': 'VOC2012',
    'split': 'train'
}
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 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
task_loss Ran
ran on a synthesized input
liuyugeng/baadd/DD/basics.py
code served (permissive licence) · get_code("2a1340040e1606e7")
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. 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
import torch.nn.functional as F

torch.manual_seed(0)

# Stand-in class for state object that has num_classes attribute
class State:
    def __init__(self, num_classes):
        self.num_classes = num_classes

# Test case 1: Binary classification (num_classes == 2)
state_binary = State(num_classes=2)
output_binary = torch.randn(2, 1)  # batch_size=2, single output for binary
label_binary = torch.randint(0, 2, (2, 1)).float()  # binary labels

# Test case 2: Multi-class classification (num_classes > 2)
state_multiclass = State(num_classes=4)
output_multiclass = torch.randn(2, 4)  # batch_size=2, 4 classes
label_multiclass = torch.randint(0, 4, (2,))  # class indices for each sample

# Create ARGS dictionary for binary classification case
ARGS = {
    'state': state_binary,
    'output': output_binary,
    'label': label_binary
}
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
get_dataset Not yet run
unverified (harvested, no recorded run)
liuyugeng/baadd/DC/utils.py
code served (permissive licence) · get_code("d7bb9b31f21b0b51")
How it ran

no invocation recorded.

resnet18 Not yet run
unverified (harvested, no recorded run)
liuyugeng/baadd/DD/networks/ResNet.py
code served (permissive licence) · get_code("3f04618a5504c3ae")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 12 are also present in other papers whose code we hold (12 with at least one paper dated earlier, 3 the same month, 7 later) and 3 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
conv1x1 112 other papers
35 dated earlier, 1 the same month, 74 later, 2 same year, order not determinable · 6 only as an assembly brick
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1806.01531 (2018-06)
Deep Mixture of Experts via Shallow Embedding
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
and 109 more not named here
conv3x3 111 other papers
36 dated earlier, 1 the same month, 72 later, 2 same year, order not determinable · 6 only as an assembly brick
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1806.01531 (2018-06)
Deep Mixture of Experts via Shallow Embedding
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
and 108 more not named here
resnet18 36 other papers
14 dated earlier, 1 the same month, 21 later
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
2002.09564 (2020-02)
Towards Robust and Reproducible Active Learning Using Neural Networks
2003.04297 (2020-03)
Improved Baselines with Momentum Contrastive Learning
and 33 more not named here
VGG11BN 28 other papers
4 dated earlier, 24 later
2203.01531 (2022-03)
CAFE: Learning to Condense Dataset by Aligning Features
2204.07513 (2022-04)
Synthesizing Informative Training Samples with GAN
2210.16774 (2022-10)
Dataset Distillation via Factorization
and 25 more not named here
VGG11 27 other papers
4 dated earlier, 23 later
2203.01531 (2022-03)
CAFE: Learning to Condense Dataset by Aligning Features
2204.07513 (2022-04)
Synthesizing Informative Training Samples with GAN
2210.16774 (2022-10)
Dataset Distillation via Factorization
and 24 more not named here
VGG13 19 other papers
3 dated earlier, 16 later
2204.07513 (2022-04)
Synthesizing Informative Training Samples with GAN
2210.16774 (2022-10)
Dataset Distillation via Factorization
2211.04446 (2022-11)
Private Set Generation with Discriminative Information
and 16 more not named here
distance_wb 11 other papers
3 dated earlier, 8 later
2202.02916 (2022-02)
Dataset Condensation with Contrastive Signals
2204.07513 (2022-04)
Synthesizing Informative Training Samples with GAN
2211.04446 (2022-11)
Private Set Generation with Discriminative Information
and 8 more not named here
clone_tuple 2 other papers
2 dated earlier
1811.10959 (2018-11)
Dataset Distillation
1910.02551 (2019-10)
Soft-Label Dataset Distillation and Text Dataset Distillation
read_bndbox 2 other papers
2 dated earlier
1811.10959 (2018-11)
Dataset Distillation
1910.02551 (2019-10)
Soft-Label Dataset Distillation and Text Dataset Distillation
permute_list 1 other paper
1 dated earlier
1811.10959 (2018-11)
Dataset Distillation
read_split 1 other paper
1 dated earlier
1811.10959 (2018-11)
Dataset Distillation
task_loss 1 other paper
1 dated earlier
1811.10959 (2018-11)
Dataset Distillation

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

Dataset distillation has emerged as a prominent technique to improve data efficiency when training machine learning models. It encapsulates the knowledge from a large dataset into a smaller synthetic dataset. A model trained on this smaller distilled dataset can attain comparable performance to a model trained on the original training dataset. However, the existing dataset distillation techniques mainly aim at achieving the best trade-off between resource usage efficiency and model utility. The security risks stemming from them have not been explored. This study performs the first backdoor attack against the models trained on the data distilled by dataset distillation models in the image domain. Concretely, we inject triggers into the synthetic data during the distillation procedure rather than during the model training stage, where all previous attacks are performed. We propose two types of backdoor attacks, namely NAIVEATTACK and DOORPING. NAIVEATTACK simply adds triggers to the raw data at the initial distillation phase, while DOORPING iteratively updates the triggers during the entire distillation procedure. We conduct extensive evaluations on multiple datasets, architectures, and dataset distillation techniques. Empirical evaluation shows that NAIVEATTACK achieves decent attack success rate (ASR) scores in some cases, while DOORPING reaches higher ASR scores (close to 1.0) in all cases. Furthermore, we conduct a comprehensive ablation study to analyze the factors that may affect the attack performance. Finally, we evaluate multiple defense mechanisms against our backdoor attacks and show that our attacks can practically circumvent these defense mechanisms.

For agents

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

get_harvested_code_for_paper("2301.01197")
get_code_for_paper("2301.01197")
have("2301.01197")

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

curl https://syntology.ai/api/ran/2301.01197.json

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

[![Syntology run record](https://syntology.ai/api/ran/2301.01197.svg)](https://syntology.ai/paper/2301.01197)

Connect an agent — have() is free.