SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2110.15073 · 2021

MMD Aggregated Two-Sample Test

arXiv · PDF · Open in the Atlas

Code that ran

14 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; 2 honoured the contract we drafted, 0 violated it and 9 have no contract verdict; 3 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/2110.15073.json · badge: /api/ran/2110.15073.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
antoninschrab/mmdagg-paper community (classifier label)
classifier basis not recorded on the link
14 ran · 1 no recorded run
FunctionStatusWhere it lives
G Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/sampling.py
code served (permissive licence) · get_code("74f5342c2c79173c")
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

def G(x):
    """
    Function G defined in Section 5.4 of our paper.
    input: x: real number
    output: G(x): real number
    """
    if -1 < x and x < -0.5:
        return np.exp(-1 / (1 - (4 * x + 3) ** 2))
    if -0.5 < x and x < 0:
        return -np.exp(-1 / (1 - (4 * x + 1) ** 2))
    return 0

# Test with various real numbers
test_values = [
    -0.75,  # Falls in first condition: -1 < x < -0.5
    -0.25,  # Falls in second condition: -0.5 < x < 0
    0.5,    # Falls in return 0 case
    -1.5,   # Falls in return 0 case
]

ARGS = {
    "x": -0.75
}
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 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
compute_median_bandwidth Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/median.py
code served (permissive licence) · get_code("7b05376c6740b314")
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 1, dtype family float, shape (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 numpy as np

# Create a small sample array Z with shape (m+n, d)
# Using shape (5, 3) - 5 pooled samples, 3 dimensions
Z = np.array([
    [1.0, 2.0, 3.0],
    [1.5, 2.5, 3.5],
    [2.0, 3.0, 4.0],
    [2.5, 3.5, 4.5],
    [3.0, 4.0, 5.0]
], dtype=np.float64)

ARGS = {
    "Z": Z
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (3,) — shape (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
compute_median_bandwidth_subset Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/median.py
code served (permissive licence) · get_code("de5787afebcb8a7c")
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 1, dtype family float, shape (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

# Set random seed for reproducibility
np.random.seed(0)

# Create small sample arrays with distinct dimensions
# X: (m, d) - 5 samples, 4 dimensions
X = np.random.randn(5, 4).astype(np.float64)

# Y: (n, d) - 3 samples, 4 dimensions (different from X to be distinct)
Y = np.random.randn(3, 4).astype(np.float64)

# Define arguments for compute_median_bandwidth_subset
ARGS = {
    "seed": 42,
    "X": X,
    "Y": Y,
    "max_samples": 2000,
    "min_value": 0.0001
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (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
compute_pairwise_matrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
antoninschrab/mmdagg-paper/mmdagg/np.py
code served (permissive licence) · get_code("139c7796806ebcb2")
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 ndarray, ndim 2, dtype family float, shape (6, 6) 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 scipy.spatial

np.random.seed(0)

# Create sample data
# X and Y should be (m, d) arrays
m = 3  # number of samples
d = 4  # dimension

X = np.random.randn(m, d).astype(np.float32)
Y = np.random.randn(m, d).astype(np.float32)

# l parameter should be 'l1' or 'l2'
l = 'l2'

ARGS = {
    'X': X,
    'Y': Y,
    'l': l
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy — 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, 6) — shape (6, 6) 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
create_weights Ran
ran, honoured the contract we drafted
antoninschrab/mmdagg-paper/mmdagg/jax.py
code served (permissive licence) · get_code("ade4e10146ee1a5c")
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 ArrayImpl, ndim 1, dtype family float, shape (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 jax.numpy as jnp

# Define arguments for create_weights function
ARGS = {
    'N': 5,
    'weights_type': 'uniform'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 1, dtype family float, shape (5,) — shape (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
f0 Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/sampling.py
code served (permissive licence) · get_code("3de090d02c055924")
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 int, ndim 0, dtype family int, 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 numpy as np

# The function f0 takes a single parameter 'x' which should be an array
# It checks if all values are between 0 and 1 (inclusive)
# and returns 1 if true, 0 otherwise

# Create a test array that satisfies the condition (all values in [0, 1])
x_valid = np.array([0.2, 0.5, 0.8, 0.1])

ARGS = {
    "x": x_valid
}
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 int, ndim 0, dtype family int, 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
f_theta Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/sampling.py
code served (permissive licence) · get_code("8f3b20cf3cb686c2")
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

# Set random seed for reproducibility
np.random.seed(42)

# Define ARGS dict with all parameters for f_theta
ARGS = {
    'x': np.array([0.3, 0.5, 0.2, 0.7]),  # (d,) array, d=4
    'p': 3,  # non-negative integer (number of perturbations)
    's': 1.5,  # positive number (smoothness parameter)
    'perturbation_multiplier': 0.5,  # positive number (c_d in Eq. (17))
    'seed': 123  # integer random seed
}
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 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
kernel_matrices Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/kernel.py
code served (permissive licence) · get_code("43b365de0cc59f50")
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 3, 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

# Set random seed for reproducibility
np.random.seed(0)

# Define small, distinct-sized inputs
m = 2  # number of X points
n = 3  # number of Y points
d = 4  # dimensionality

# Create sample data
X = np.random.randn(m, d).astype(np.float64)
Y = np.random.randn(n, d).astype(np.float64)

# Define bandwidth array (one per dimension)
bandwidth = np.ones(d, dtype=np.float64)

# Define bandwidth multipliers
bandwidth_multipliers = np.array([0.5, 1.0, 2.0], dtype=np.float64)

# Define kernel type
kernel_type = "gaussian"

# Construct ARGS dictionary
ARGS = {
    "X": X,
    "Y": Y,
    "kernel_type": kernel_type,
    "bandwidth": bandwidth,
    "bandwidth_multipliers": bandwidth_multipliers,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 3, 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
kernel_matrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
antoninschrab/mmdagg-paper/mmdagg/jax.py
code served (permissive licence) · get_code("2c48c4eb0714eef3")
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 ArrayImpl, ndim 2, dtype family float, shape (4, 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 jax.numpy as jnp
import numpy as np

# Create a small pairwise distance matrix (4x4)
# This represents pairwise distances between 2 points (2m=4, so m=2)
pairwise_matrix = jnp.array([
    [0.0, 1.5, 2.0, 2.5],
    [1.5, 0.0, 1.2, 1.8],
    [2.0, 1.2, 0.0, 0.9],
    [2.5, 1.8, 0.9, 0.0]
], dtype=jnp.float32)

# Valid combinations of l and kernel based on the function logic:
# - gaussian with l2
# - imq with l2
# - matern_0.5_l1 with l1, or matern_0.5_l2 with l2, or laplace with l1
# - matern_1.5_l1 with l1, or matern_1.5_l2 with l2
# - etc.

l = "l2"
kernel = "gaussian"
bandwidth = 1.0

ARGS = {
    "pairwise_matrix": pairwise_matrix,
    "l": l,
    "kernel": kernel,
    "bandwidth": bandwidth
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 2, dtype family float, shape (4, 4) — shape (4, 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
kernel_matrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
antoninschrab/mmdagg-paper/mmdagg/np.py
code served (permissive licence) · get_code("4e7beb2199214856")
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 ndarray, ndim 2, dtype family float, shape (4, 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

# Create a small pairwise distance matrix (4x4)
pairwise_matrix = np.array([
    [0.0, 0.5, 1.0, 1.5],
    [0.5, 0.0, 0.8, 1.2],
    [1.0, 0.8, 0.0, 0.6],
    [1.5, 1.2, 0.6, 0.0]
])

# Valid combinations of l and kernel_type
l = "l2"
kernel_type = "gaussian"
bandwidth = 1.0

ARGS = {
    "pairwise_matrix": pairwise_matrix,
    "l": l,
    "kernel_type": kernel_type,
    "bandwidth": bandwidth
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 (4, 4) — shape (4, 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
pairwise_l1_distance Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/kernel.py
code served (permissive licence) · get_code("dac4409612e19fc2")
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 (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 numpy as np
from numba import njit

@njit
def pairwise_l1_distance(Z):
    """
    Compute the pairwise L^1-distance matrix between all points in Z.
    inputs: Z is (mn,d) array
    output: (mn,mn) array of pairwise squared distances (L^1)
    """
    mn, d = Z.shape
    output = np.zeros((mn, mn))
    for i in range(mn):
        for j in range(mn):
            temp = 0.0
            for u in range(d):
                temp += np.abs(Z[i, u] - Z[j, u])
            output[i, j] = temp
    return output

# Create a small test array with shape (mn, d)
# Using distinct dimensions: 3 points, 4 dimensions each
Z = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0],
    [2.0, 3.0, 4.0, 5.0]
], dtype=np.float64)

ARGS = {
    "Z": Z
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, 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 (3, 3) — shape (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
pairwise_square_l2_distance Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/kernel.py
code served (permissive licence) · get_code("a73df06f2cbc35e5")
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 (2, 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

# Create a small sample point cloud: 2 points in 4D space
Z = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float64)

ARGS = {
    "Z": Z
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, 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 (2, 2) — shape (2, 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
truncated_gaussian Ran
ran, honoured the contract we drafted
antoninschrab/mmdagg-paper/ost.py
code served (permissive licence) · get_code("7da2a914d0ee3daf")
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

ARGS = {
    'var': 0.5,
    'v_minus': -1.0,
    'level': 0.05
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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
truncation Ran
ran on a synthesized input
antoninschrab/mmdagg-paper/ost.py
code served (permissive licence) · get_code("b8733b899ab4f726")
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

# Set random seed for reproducibility
np.random.seed(0)

# Define dimensions
d = 4

# Create beta_star: optimal projection vector
# Make it sparse (some zeros) to test the non_zero logic
beta_star = np.array([0.5, 0.0, 0.3, 0.0])

# Create tau: vector of test statistics
tau = np.array([1.2, -0.5, 0.8, 0.3])

# Create Sigma: covariance matrix (must be symmetric positive semi-definite)
# Start with a random matrix and make it symmetric positive definite
A = np.random.randn(d, d)
Sigma = A @ A.T + np.eye(d)  # Ensure positive definiteness

# Accuracy threshold
accuracy = 1e-06

ARGS = {
    "beta_star": beta_star,
    "tau": tau,
    "Sigma": Sigma,
    "accuracy": accuracy
}
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 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
mmdagg Not yet run
unverified (harvested, no recorded run)
antoninschrab/mmdagg-paper/mmdagg/jax.py
code served (permissive licence) · get_code("51da116869a49b91")
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, 9 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 7 later) and 6 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
compute_pairwise_matrix 2 other papers
2 later
2206.08843 (2022-06)
AutoML Two-Sample Test
2206.09194 (2022-06)
Efficient Aggregated Kernel Tests using Incomplete $U$-statistics
create_weights 2 other papers
2 later
2202.00824 (2022-02)
KSD Aggregated Goodness-of-fit Test
2206.09194 (2022-06)
Efficient Aggregated Kernel Tests using Incomplete $U$-statistics
kernel_matrix 2 other papers
2 later
2206.08843 (2022-06)
AutoML Two-Sample Test
2206.09194 (2022-06)
Efficient Aggregated Kernel Tests using Incomplete $U$-statistics
G 1 other paper
1 later
2306.08777 (2023-06)
MMD-FUSE: Learning and Combining Kernels for Two-Sample Testing Without Data Splitting
f0 1 other paper
1 later
2306.08777 (2023-06)
MMD-FUSE: Learning and Combining Kernels for Two-Sample Testing Without Data Splitting
f_theta 1 other paper
1 later
2306.08777 (2023-06)
MMD-FUSE: Learning and Combining Kernels for Two-Sample Testing Without Data Splitting
kernel_matrix 1 other paper
1 later
2206.09194 (2022-06)
Efficient Aggregated Kernel Tests using Incomplete $U$-statistics
truncated_gaussian 1 other paper
1 dated earlier
2006.02286 (2020-06)
Learning Kernel Tests Without Data Splitting
truncation 1 other paper
1 dated earlier
2006.02286 (2020-06)
Learning Kernel Tests Without Data Splitting

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

We propose two novel nonparametric two-sample kernel tests based on the Maximum Mean Discrepancy (MMD). First, for a fixed kernel, we construct an MMD test using either permutations or a wild bootstrap, two popular numerical procedures to determine the test threshold. We prove that this test controls the probability of type I error non-asymptotically. Hence, it can be used reliably even in settings with small sample sizes as it remains well-calibrated, which differs from previous MMD tests which only guarantee correct test level asymptotically. When the difference in densities lies in a Sobolev ball, we prove minimax optimality of our MMD test with a specific kernel depending on the smoothness parameter of the Sobolev ball. In practice, this parameter is unknown and, hence, the optimal MMD test with this particular kernel cannot be used. To overcome this issue, we construct an aggregated test, called MMDAgg, which is adaptive to the smoothness parameter. The test power is maximised over the collection of kernels used, without requiring held-out data for kernel selection (which results in a loss of test power), or arbitrary kernel choices such as the median heuristic. We prove that MMDAgg still controls the level non-asymptotically, and achieves the minimax rate over Sobolev balls, up to an iterated logarithmic term. Our guarantees are not restricted to a specific type of kernel, but hold for any product of one-dimensional translation invariant characteristic kernels. We provide a user-friendly parameter-free implementation of MMDAgg using an adaptive collection of bandwidths. We demonstrate that MMDAgg significantly outperforms alternative state-of-the-art MMD-based two-sample tests on synthetic data satisfying the Sobolev smoothness assumption, and that, on real-world image data, MMDAgg closely matches the power of tests leveraging the use of models such as neural networks.

For agents

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

get_harvested_code_for_paper("2110.15073")
get_code_for_paper("2110.15073")
have("2110.15073")

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

curl https://syntology.ai/api/ran/2110.15073.json

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

[![Syntology run record](https://syntology.ai/api/ran/2110.15073.svg)](https://syntology.ai/paper/2110.15073)

Connect an agent — have() is free.