SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.03985 · 2024

A Bias-Variance Decomposition for Ensembles over Multiple Synthetic Datasets

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; 4 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 5 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/2402.03985.json · badge: /api/ran/2402.03985.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
dpbayes/generative-ensemble-bias-variance-decomposition canonical (classifier label)
classifier basis: census
11 ran · 1 no recorded run
FunctionStatusWhere it lives
adult_reduced_discretisation Ran
ran; our contract draft was wrong, not the code
dpbayes/generative-ensemble-bias-variance-decomposition/lib/experiment_config.py
code served (permissive licence) · get_code("131d9e48e28818cf")
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 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 numpy as np

# Create a minimal DataFrame with the required columns
df = pd.DataFrame({
    'age': [25, 35, 45, 55, 65, 75],
    'hours_per_week': [20, 30, 40, 50, 60, 70]
})

ARGS = {
    'df': df
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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
cdp_delta Ran
ran, honoured the contract we drafted
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/cdp2adp.py
pointer only (licence: MIT); the stored copy get_code serves, from miguel-fuentes/jam_aistats, is pointer only (licence: NONE) · get_code("12513a27a1865536")
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 miguel-fuentes/jam_aistats, 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.

cdp_delta_standard Ran
ran, honoured the contract we drafted
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/cdp2adp.py
code served (permissive licence) · get_code("6026d9da4638c121")
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 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 math

def cdp_delta_standard(rho, eps):
    assert rho >= 0
    assert eps >= 0
    if rho == 0:
        return 0
    return math.exp(-(eps - rho) ** 2 / (4 * rho))

ARGS = {
    "rho": 0.5,
    "eps": 1.0,
}
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 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
cdp_eps Ran
ran, honoured the contract we drafted
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/cdp2adp.py
pointer only (licence: MIT); the stored copy get_code serves, from miguel-fuentes/jam_aistats, is pointer only (licence: NONE) · get_code("438c48170078d9e5")
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 miguel-fuentes/jam_aistats, 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.

downward_closure Ran
ran; our contract draft was wrong, not the code
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/aim.py
pointer only (licence: MIT); the stored copy get_code serves, from miguel-fuentes/jam_aistats, is pointer only (licence: NONE) · get_code("ef1ef33de7a9f882")
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 miguel-fuentes/jam_aistats, 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.

generalized_em_scores Ran
ran, violated the contract we drafted
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/mechanism.py
pointer only (licence: MIT); the stored copy get_code serves, from miguel-fuentes/jam_aistats, is pointer only (licence: NONE) · get_code("a0f36b2b02e60d5c")
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 miguel-fuentes/jam_aistats, 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.

is_classification Ran
ran, violated the contract we drafted
dpbayes/generative-ensemble-bias-variance-decomposition/lib/experiment_config.py
code served (permissive licence) · get_code("73be1750da2ee3a5")
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

torch.manual_seed(0)

classification_datasets = ['german-credit', 'adult', 'breast-cancer']

def is_classification(dataset):
    return dataset in classification_datasets

ARGS = {
    'dataset': 'adult'
}
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
pareto_efficient Ran
ran, honoured the contract we drafted
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/mechanism.py
pointer only (licence: MIT); the stored copy get_code serves, from miguel-fuentes/jam_aistats, is pointer only (licence: NONE) · get_code("87f72fb6f142740c")
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 miguel-fuentes/jam_aistats, 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.

powerset Ran
ran; our contract draft was wrong, not the code
dpbayes/generative-ensemble-bias-variance-decomposition/lib/aim/aim.py
pointer only (licence: MIT); the stored copy get_code serves, from miguel-fuentes/jam_aistats, is pointer only (licence: NONE) · get_code("802df9e4d1beab55")
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 miguel-fuentes/jam_aistats, 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.

tqdm_choice Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
dpbayes/generative-ensemble-bias-variance-decomposition/lib/util.py
code served (permissive licence) · get_code("92d5ddfb9acb81c2")
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 torch
from tqdm import tqdm

# Create a simple iterator
torch.manual_seed(0)
iterator = [1, 2, 3, 4, 5]

# Define ARGS with all parameters
ARGS = {
    "iterator": iterator,
    "show_progress": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, tqdm — 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
train_model_return_mse Ran
ran; our contract draft was wrong, not the code
dpbayes/generative-ensemble-bias-variance-decomposition/lib/util.py
code served (permissive licence) · get_code("e3c6fd4fec348bcd")
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 float64, ndim 0, 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 numpy as np
from sklearn.linear_model import LinearRegression

np.random.seed(0)

# Create small training and test datasets
X_train = np.random.randn(2, 4)  # 2 samples, 4 features
y_train = np.random.randn(2)      # 2 targets
X_test = np.random.randn(3, 4)    # 3 samples, 4 features
y_test = np.random.randn(3)        # 3 targets

# Create a model that supports clone(), fit(), and predict()
model = LinearRegression()

ARGS = {
    "model": model,
    "X_train": X_train,
    "y_train": y_train,
    "X_test": X_test,
    "y_test": y_test,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, 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_oh_df Not yet run
unverified (harvested, no recorded run)
dpbayes/generative-ensemble-bias-variance-decomposition/lib/util.py
code served (permissive licence) · get_code("1e0c03c46fb5b6f0")
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 (1 with at least one paper dated earlier, 0 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
cdp_delta_standard 4 other papers
2 dated earlier, 2 later
2004.00010 (2020-04)
The Discrete Gaussian for Differential Privacy
2401.17823 (2024-01)
Privacy-preserving data release leveraging optimal transport and particle gradient descent
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data
and 1 more not named here
cdp_delta 1 other paper
1 later
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data
cdp_eps 1 other paper
1 later
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data
downward_closure 1 other paper
1 later
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data
generalized_em_scores 1 other paper
1 later
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data
pareto_efficient 1 other paper
1 later
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data
powerset 1 other paper
1 later
2403.07797 (2024-03)
Joint Selection: Adaptively Incorporating Public Information for Private Synthetic Data

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

Recent studies have highlighted the benefits of generating multiple synthetic datasets for supervised learning, from increased accuracy to more effective model selection and uncertainty estimation. These benefits have clear empirical support, but the theoretical understanding of them is currently very light. We seek to increase the theoretical understanding by deriving bias-variance decompositions for several settings of using multiple synthetic datasets, including differentially private synthetic data. Our theory yields a simple rule of thumb to select the appropriate number of synthetic datasets in the case of mean-squared error and Brier score. We investigate how our theory works in practice with several real datasets, downstream predictors and error metrics. As our theory predicts, multiple synthetic datasets often improve accuracy, while a single large synthetic dataset gives at best minimal improvement, showing that our insights are practically relevant.

For agents

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

get_harvested_code_for_paper("2402.03985")
get_code_for_paper("2402.03985")
have("2402.03985")

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

curl https://syntology.ai/api/ran/2402.03985.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.03985.svg)](https://syntology.ai/paper/2402.03985)

Connect an agent — have() is free.