SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.09564 · 2020

Towards Robust and Reproducible Active Learning Using Neural Networks

arXiv · PDF · Open in the Atlas

Code that ran

11 of 12 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, 2 violated it and 4 have no contract verdict; 5 failures of our instrument; 1 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/2002.09564.json · badge: /api/ran/2002.09564.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
prateekmunjal/torchal canonical (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
acl21/deep-active-learning-pytorch community (classifier label)
classifier basis not recorded on the link
10 ran · 1 no recorded run
FunctionStatusWhere it lives
alexnet Ran
ran; our contract draft was wrong, not the code
acl21/deep-active-learning-pytorch/pycls/models/alexnet.py
code served (permissive licence) · get_code("1d3e648204ca8f7f")
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

torch.manual_seed(0)

ARGS = {
    'pretrained': False,
    'progress': True,
}
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 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
construct_optimizer Ran
ran on a synthesized input
acl21/deep-active-learning-pytorch/pycls/core/optimizer.py
code served (permissive licence) · get_code("08ca33c7fa25c77b")
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 SGD. 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

# Set seed for reproducibility
torch.manual_seed(0)

# Create a minimal model with batch norm layers
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc1 = nn.Linear(4, 8)
        self.bn1 = nn.BatchNorm1d(8)
        self.fc2 = nn.Linear(8, 3)
        self.bn2 = nn.BatchNorm1d(3)
    
    def forward(self, x):
        x = self.fc1(x)
        x = self.bn1(x)
        x = self.fc2(x)
        x = self.bn2(x)
        return x

# Create a minimal config class
class Config:
    def __init__(self):
        self.BN = BNConfig()
        self.OPTIM = OptimConfig()

class BNConfig:
    def __init__(self):
        self.USE_CUSTOM_WEIGHT_DECAY = True
        self.CUSTOM_WEIGHT_DECAY = 0.0001

class OptimConfig:
    def __init__(self):
        self.TYPE = 'sgd'
        self.BASE_LR = 0.01
        self.MOMENTUM = 0.9
        self.WEIGHT_DECAY = 0.0005
        self.DAMPENING = 0.0
        self.NESTEROV = True

# Instantiate config and model
cfg = Config()
model = SimpleModel()

ARGS = {
    'cfg': cfg,
    'model': model
}
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 SGD — a SGD output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SGD 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
conv1x1 Ran
ran; our contract draft was wrong, not the code
acl21/deep-active-learning-pytorch/pycls/models/resnet.py
pointer only (licence: MIT); 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 (MIT), 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
acl21/deep-active-learning-pytorch/pycls/models/resnet.py
pointer only (licence: MIT); 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 (MIT), 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.

drop_connect Ran
ran; our contract draft was wrong, not the code
acl21/deep-active-learning-pytorch/pycls/utils/net.py
code served (permissive licence) · get_code("79583b4154444ddf")
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 Tensor, ndim 4, dtype family float, shape (2, 4, 3, 5) 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)

# Create a small 4D tensor (batch_size=2, channels=4, height=3, width=5)
x = torch.randn(2, 4, 3, 5, dtype=torch.float32)

# drop_ratio parameter - a float between 0 and 1
drop_ratio = 0.5

ARGS = {
    "x": x,
    "drop_ratio": drop_ratio,
}
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 Tensor, ndim 4, dtype family float, shape (2, 4, 3, 5) — shape (2, 4, 3, 5) 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_flat_weights Ran
ran on a synthesized input
acl21/deep-active-learning-pytorch/pycls/utils/net.py
code served (permissive licence) · get_code("b26a6d70c4c1b982")
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 (23, 1) 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)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        return self.linear2(self.linear1(x))

model = SimpleModel()

ARGS = {
    "model": model
}
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 (23, 1) — shape (23, 1) 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
is_eval_epoch Ran
ran, violated the contract we drafted
prateekmunjal/torchal/tools/train_al.py
code served (permissive licence) · get_code("24d376b7f841ed4c")
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 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
import torch.nn as nn

class Config:
    def __init__(self):
        self.TRAIN = TrainConfig()
        self.OPTIM = OptimConfig()

class TrainConfig:
    def __init__(self):
        self.EVAL_PERIOD = 5

class OptimConfig:
    def __init__(self):
        self.MAX_EPOCH = 20

cfg = Config()
cur_epoch = 4

ARGS = {
    "cfg": cfg,
    "cur_epoch": cur_epoch
}
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 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
lr_fun_exp Ran
ran on a synthesized input
acl21/deep-active-learning-pytorch/pycls/core/optimizer.py
code served (permissive licence) · get_code("2ea7d3298243fe68")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

class ConfigStandIn:
    def __init__(self):
        self.OPTIM = OptimConfig()

class OptimConfig:
    def __init__(self):
        self.LR_POLICY = 'exp'
        self.MIN_LR = 0.1
        self.MAX_EPOCH = 100

cfg = ConfigStandIn()
cur_epoch = 50

ARGS = {
    'cfg': cfg,
    'cur_epoch': cur_epoch
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
lr_fun_steps Ran
ran on a synthesized input
acl21/deep-active-learning-pytorch/pycls/core/optimizer.py
code served (permissive licence) · get_code("89e97f7341666d97")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

class Config:
    def __init__(self):
        self.OPTIM = OptimConfig()

class OptimConfig:
    def __init__(self):
        self.STEPS = [0, 10, 20, 30]
        self.LR_MULT = 0.1

cfg = Config()
cur_epoch = 15

ARGS = {
    "cfg": cfg,
    "cur_epoch": cur_epoch
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_layers Ran
ran; our contract draft was wrong, not the code
acl21/deep-active-learning-pytorch/pycls/models/vgg.py
code served (permissive licence) · get_code("0bac0ac5d129be80")
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
from typing import Union, List

torch.manual_seed(0)

# cfg: List[Union[str, int]] - a configuration list with integers (channel counts) and 'M' (MaxPool markers)
# batch_norm: bool - whether to use batch normalization
cfg = [64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512]
batch_norm_flag = True

ARGS = {
    "cfg": cfg,
    "batch_norm": batch_norm_flag,
}
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
unwrap_model Ran
ran, violated the contract we drafted
acl21/deep-active-learning-pytorch/pycls/core/net.py
code served (permissive licence) · get_code("8ba24560996e7d76")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned SimpleModel. 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 simple model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Create the base model
base_model = SimpleModel()

# The function checks if the model is wrapped in DistributedDataParallel
# and returns model.module if it is, otherwise returns the model itself
# We'll test with a non-wrapped model (the wrapped case would require
# distributed setup which isn't available in this environment)

ARGS = {
    "model": base_model
}
Setup block drafted by
a language model, Claude Sonnet 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 SimpleModel — a SimpleModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModel 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
resnet18 Not yet run
unverified (harvested, no recorded run)
acl21/deep-active-learning-pytorch/pycls/models/resnet.py
code served (permissive licence) · get_code("5f276f6de9406cd1")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 7 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 1 the same month, 7 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
conv1x1 112 other papers
4 dated earlier, 108 later · 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
3 dated earlier, 108 later · 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
1 dated earlier, 35 later
1905.12794 (2019-05)
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback
2003.04297 (2020-03)
Improved Baselines with Momentum Contrastive Learning
2103.13859 (2021-03)
Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networks
and 33 more not named here
drop_connect 29 other papers
2 dated earlier, 1 the same month, 26 later
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
2002.11318 (2020-02)
Can we have it all? On the Trade-off between Spatial and Adversarial Robustness of Neural Networks
and 26 more not named here
make_layers 9 other papers
9 later
2306.12685 (2023-06)
Rethinking the Backward Propagation for Adversarial Transferability
2306.13856 (2023-06)
Learning-to-Rank Meets Language: Boosting Language-Driven Ordering Alignment for Ordinal Classification
2308.06248 (2023-08)
FunnyBirds: A Synthetic Vision Dataset for a Part-Based Analysis of Explainable AI Methods
and 6 more not named here
alexnet 3 other papers
3 later
2111.13998 (2021-11)
Targeted Supervised Contrastive Learning for Long-Tailed Recognition
2306.13856 (2023-06)
Learning-to-Rank Meets Language: Boosting Language-Driven Ordering Alignment for Ordinal Classification
2411.02853 (2024-11)
ADOPT: Modified Adam Can Converge with Any β 2 with the Optimal Rate
unwrap_model 3 other papers
1 dated earlier, 2 later
1905.13214 (2019-05)
On Network Design Spaces for Visual Recognition
2405.05808 (2024-05)
Fast and Controllable Post-training Sparsity: Learning Optimal Sparsity Allocation with Global Constraint in Minutes
2607.06982 (2026-07)
EdgeCompress: Coupling Multidimensional Model Compression and Dynamic Inference for EdgeAI

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

Active learning (AL) is a promising ML paradigm that has the potential to parse through large unlabeled data and help reduce annotation cost in domains where labeling data can be prohibitive. Recently proposed neural network based AL methods use different heuristics to accomplish this goal. In this study, we demonstrate that under identical experimental settings, different types of AL algorithms (uncertainty based, diversity based, and committee based) produce an inconsistent gain over random sampling baseline. Through a variety of experiments, controlling for sources of stochasticity, we show that variance in performance metrics achieved by AL algorithms can lead to results that are not consistent with the previously reported results. We also found that under strong regularization, AL methods show marginal or no advantage over the random sampling baseline under a variety of experimental conditions. Finally, we conclude with a set of recommendations on how to assess the results using a new AL algorithm to ensure results are reproducible and robust under changes in experimental conditions. We share our codes to facilitate AL evaluations. We believe our findings and recommendations will help advance reproducible research in AL using neural networks. We open source our code at https://github.com/PrateekMunjal/TorchAL

For agents

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

get_harvested_code_for_paper("2002.09564")
get_code_for_paper("2002.09564")
have("2002.09564")

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

curl https://syntology.ai/api/ran/2002.09564.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.09564.svg)](https://syntology.ai/paper/2002.09564)

Connect an agent — have() is free.