SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1802.04422 · 2018

A comparative study of fairness-enhancing interventions in machine learning

arXiv · PDF · Open in the Atlas

Code that ran

10 of 11 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 0 have no contract verdict; 9 failures of our instrument; 1 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/1802.04422.json · badge: /api/ran/1802.04422.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
algofairness/fairness-comparison canonical (classifier label)
classifier basis: classifier_2025
4 ran · 0 no recorded run
charan223/FairDeepLearning reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 1 no recorded run
JSGoedhart/fairness-comparison reimplementation (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
huanglx12/Stable-Fair-Classification extension (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
all_possible_graphs Ran
ran; our contract draft was wrong, not the code
JSGoedhart/fairness-comparison/fairness/analysis.py
pointer only (licence: NOASSERTION) · get_code("a8410405c788323a")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, JSGoedhart/fairness-comparison (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

compute_diff_nums Ran
ran, honoured the contract we drafted
huanglx12/Stable-Fair-Classification/fairness/benchmark.py
pointer only (licence: NONE) · get_code("47a0278c9fef3265")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, huanglx12/Stable-Fair-Classification 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.

filter_table Ran
ran; our contract draft was wrong, not the code
charan223/FairDeepLearning/scripts/plots_tables/csvs/parse_csv.py
code served (permissive licence) · get_code("7da718e0807f10c7")
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 DataFrame. 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 pandas as pd
import torch

torch.manual_seed(0)

# Create a sample DataFrame with metric columns and grouping columns
data_table = pd.DataFrame({
    'metric_a': [0.5, 1.5],
    'metric_b': [0.8, 1.2],
    'group_x': ['A', 'B'],
    'group_y': [1, 2]
})

# metric_cols maps original column names to new column names
metric_cols = {
    'metric_a': 'transformed_a',
    'metric_b': 'transformed_b'
}

# all_grouping_cols lists the grouping column names
all_grouping_cols = ['group_x', 'group_y']

ARGS = {
    'data_table': data_table,
    'metric_cols': metric_cols,
    'all_grouping_cols': all_grouping_cols
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
pandas, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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_dict_sensitive_vals Ran
ran; our contract draft was wrong, not the code
algofairness/fairness-comparison/fairness/benchmark.py
pointer only (licence: NOASSERTION) · get_code("6642c1dab7b8e92b")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, algofairness/fairness-comparison (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_iters Ran
ran; our contract draft was wrong, not the code
charan223/FairDeepLearning/train/execute.py
code served (permissive licence) · get_code("c861a2f90457c10b")
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 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 numpy as np
from unittest.mock import Mock

torch.manual_seed(0)

class MockDataset:
    """Mock dataset that implements get_batch_iterator method"""
    
    def __init__(self):
        self.data = {
            'train': [(torch.randn(2, 4), torch.randint(0, 2, (2,))) for _ in range(3)],
            'valid': [(torch.randn(2, 4), torch.randint(0, 2, (2,))) for _ in range(2)],
            'test': [(torch.randn(2, 4), torch.randint(0, 2, (2,))) for _ in range(2)]
        }
    
    def get_batch_iterator(self, split, batch_size, shuffle=False, keep_remainder=False, seed=None):
        """Returns an iterator over batches for the given split"""
        batches = self.data[split]
        if shuffle and seed is not None:
            rng = np.random.RandomState(seed)
            indices = rng.permutation(len(batches))
            batches = [batches[i] for i in indices]
        
        # Return an iterator-like object
        class BatchIterator:
            def __init__(self, batches):
                self.batches = batches
                self.index = 0
            
            def __iter__(self):
                return iter(self.batches)
            
            def __next__(self):
                if self.index >= len(self.batches):
                    raise StopIteration
                batch = self.batches[self.index]
                self.index += 1
                return batch
        
        return BatchIterator(batches)

class MockArgs:
    """Mock args object with required attributes"""
    def __init__(self):
        self.batch_size = 2
        self.seed = 42

ARGS = {
    'args': MockArgs(),
    'dataset': MockDataset()
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family 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
get_sensitive_from_filename Ran
ran; our contract draft was wrong, not the code
algofairness/fairness-comparison/analysis/correlation-vis/combine_results_files.py
pointer only (licence: NOASSERTION) · get_code("0f5b147fa50f6183")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, algofairness/fairness-comparison (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

group_and_agg Ran
ran; our contract draft was wrong, not the code
charan223/FairDeepLearning/scripts/plots_tables/csvs/parse_csv.py
code served (permissive licence) · get_code("a6f24b46ffb7aa43")
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 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 pandas as pd
import numpy as np

# Create a sample DataFrame with multiple columns
np.random.seed(0)
data = {
    'group_col1': ['A', 'A', 'B', 'B', 'C', 'C'],
    'group_col2': ['X', 'Y', 'X', 'Y', 'X', 'Y'],
    'metric1': [1.5, 2.3, 3.1, 4.2, 5.0, 6.1],
    'metric2': [10.2, 11.5, 12.3, 13.1, 14.5, 15.2],
    'metric3': [100.1, 102.3, 105.2, 107.1, 110.5, 112.3]
}
tbl = pd.DataFrame(data)

# Define grouping columns
mean_grouping_cols = ['group_col1', 'group_col2']

# Define metric columns as a dict mapping metric names to column names
metric_cols = {
    'metric1': 'metric1',
    'metric2': 'metric2',
    'metric3': 'metric3'
}

ARGS = {
    'tbl': tbl,
    'mean_grouping_cols': mean_grouping_cols,
    'metric_cols': metric_cols
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas — 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_metrics_list Ran
ran; our contract draft was wrong, not the code
algofairness/fairness-comparison/analysis/correlation-vis/combine_results_files.py
pointer only (licence: NOASSERTION) · get_code("91ce47d82c73659b")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, algofairness/fairness-comparison (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

run_alg Ran
ran; our contract draft was wrong, not the code
algofairness/fairness-comparison/fairness/benchmark.py
pointer only (licence: NOASSERTION) · get_code("5bd1ce1fcaa81378")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, algofairness/fairness-comparison (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

run_alg Ran
ran; our contract draft was wrong, not the code
JSGoedhart/fairness-comparison/fairness/benchmark.py
pointer only (licence: NOASSERTION) · get_code("13d270541da56dd7")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, JSGoedhart/fairness-comparison (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

group_and_max Not yet run
unverified (harvested, no recorded run)
charan223/FairDeepLearning/scripts/plots_tables/csvs/parse_csv.py
code served (permissive licence) · get_code("37de72ff09213994")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 11 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

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

Abstract

Computers are increasingly used to make decisions that have significant impact in people's lives. Often, these predictions can affect different population subgroups disproportionately. As a result, the issue of fairness has received much recent interest, and a number of fairness-enhanced classifiers and predictors have appeared in the literature. This paper seeks to study the following questions: how do these different techniques fundamentally compare to one another, and what accounts for the differences? Specifically, we seek to bring attention to many under-appreciated aspects of such fairness-enhancing interventions. Concretely, we present the results of an open benchmark we have developed that lets us compare a number of different algorithms under a variety of fairness measures, and a large number of existing datasets. We find that although different algorithms tend to prefer specific formulations of fairness preservations, many of these measures strongly correlate with one another. In addition, we find that fairness-preserving algorithms tend to be sensitive to fluctuations in dataset composition (simulated in our benchmark by varying training-test splits), indicating that fairness interventions might be more brittle than previously thought.

For agents

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

get_harvested_code_for_paper("1802.04422")
get_code_for_paper("1802.04422")
have("1802.04422")

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

curl https://syntology.ai/api/ran/1802.04422.json

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

[![Syntology run record](https://syntology.ai/api/ran/1802.04422.svg)](https://syntology.ai/paper/1802.04422)

Connect an agent — have() is free.