SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2206.07912 · ICML · 2022

Double Sampling Randomized Smoothing

Jiawei Zhang, Linyi Li, Bo Li, Tao Xie

arXiv · PDF · Open in the Atlas

Code that ran

15 of 17 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; 14 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 0 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2206.07912.json · badge: /api/ran/2206.07912.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
llylly/dsrs canonical (classifier label)
classifier basis: classifier_2025
0 ran · 1 no recorded run
dsgiitr/re_dsrs community
Papers with Code archive
15 ran · 1 no recorded run
FunctionStatusWhere it lives
P_double_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("f06d8dd1d486d93b")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
import scipy

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

ARGS = {
    "disttype": "gaussian",
    "d": 10,
    "k": 2,
    "sigma": 1.5,
    "beta": 0.8,
    "lamb1": 0.5,
    "lamb2": -0.3,
    "r": 2.0,
    "bisearch_precision": 0.01,
}
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, 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
P_single_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("2e6af63ad6bfdd87")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
from scipy.stats import norm
import scipy

ARGS = {
    'disttype': 'gaussian',
    'd': 10,
    'k': 2,
    'sigma': 1.5,
    'logK': 2.0,
    'r': 3.0
}
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, 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
Pprime_single_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("8ea8528158788615")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
import scipy

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

# Define ARGS dictionary with valid arguments for Pprime_single_var_numerical
ARGS = {
    'disttype': 'gaussian',
    'd': 5.0,
    'k': 1.5,
    'sigma': 2.0,
    'beta': 3.0,
    'logK': 1.5,
    'r': 2.5
}
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, 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
Pshift_double_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("7db131201488099d")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta

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

# Define ARGS dictionary with valid arguments for Pshift_double_var_numerical
ARGS = {
    'disttype': 'gaussian',  # Either 'gaussian' or another type (e.g., 'chi')
    'd': 10,                  # Dimension parameter (positive integer)
    'k': 2,                   # Shape parameter (positive integer, k < d/2)
    'sigma': 1.5,             # Scale parameter (positive float)
    'beta': 1.2,              # Scale parameter (positive float)
    'lamb1': 0.5,             # Lambda 1 parameter (can be positive or negative)
    'lamb2': -0.3,            # Lambda 2 parameter (can be positive or negative)
    'rad': 2.0,               # Radius parameter (positive float)
    'bisearch_precision': 0.01,  # Binary search precision (small positive float)
    'eps': 1e-08              # Tolerance parameter (small positive float)
}
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, 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
Pshift_single_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("fb9da182e97479f9")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
from scipy.stats import norm
import scipy

ARGS = {
    'disttype': 'gaussian',
    'd': 10,
    'k': 2.0,
    'sigma': 1.5,
    'logK': 0.5,
    'r': 2.0,
    'limit': 50
}
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, 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
Q_double_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("435090fa4bff4fa8")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
import scipy

# Set up arguments for Q_double_var_numerical
ARGS = {
    'disttype': 'gaussian',  # Can be 'gaussian' or other types
    'd': 10,  # dimension parameter
    'k': 2,  # k parameter
    'sigma': 1.5,  # sigma parameter
    'beta': 2.0,  # beta parameter
    'lamb1': 0.5,  # lambda1 parameter
    'lamb2': -0.3,  # lambda2 parameter
    'r': 1.0,  # r parameter
    'bisearch_precision': 0.01,  # binary search precision
    'limit': 50  # integration limit
}
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, 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
Q_single_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("9540a262bfdc86e7")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
from scipy.stats import norm
import scipy

ARGS = {
    'disttype': 'gaussian',
    'd': 5.0,
    'k': 1.0,
    'sigma': 0.5,
    'beta': 0.3,
    'logK': 2.0,
    'r': 1.5
}
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, 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
Qprime_single_var_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("6814ef870c488dc0")
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, 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
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
import scipy

# Analyze the function to determine what parameters are needed
# The function takes: disttype, d, k, sigma, beta, logK, r
# All appear to be numeric scalar values based on usage

ARGS = {
    'disttype': 'gaussian',  # Can be 'gaussian', 'general-gaussian', or 'general-gaussian-th'
    'd': 4.0,  # dimension parameter, used in Gamma(d/2.0)
    'k': 0.5,  # shape parameter for general-gaussian variants
    'sigma': 1.0,  # scale parameter
    'beta': 0.8,  # beta parameter, used in various calculations
    'logK': 2.0,  # log of threshold K
    'r': 1.5,  # radius or range parameter
}
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, 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
bin_search_K_on_P_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("81a519ede3a6fd98")
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 numpy as np
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
from scipy.stats import norm
import scipy

ARGS = {
    "disttype": "gaussian",
    "d": 10,
    "k": 2,
    "sigma": 1.5,
    "rad": 3.0,
    "P": 0.7,
    "range": 20.0,
    "EPS": 1e-06
}
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 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
bin_search_K_on_Qprime_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("efb5077cd6701432")
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 numpy as np
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
import scipy

ARGS = {
    'disttype': 'gaussian',
    'd': 5,
    'k': 1.0,
    'sigma': 2.0,
    'beta': 1.5,
    'rad': 0.8,
    'Q': 0.5,
    'range': 20.0,
    'EPS': 1e-6
}
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 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
bin_search_lambda_2_on_P_numerical Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("fda2190670c06dcb")
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 numpy as np
from scipy.stats import gamma as Gamma
from scipy.stats import beta as Beta
import scipy

ARGS = {
    'disttype': 'gaussian',
    'd': 10,
    'k': 2,
    'sigma': 1.5,
    'beta': 0.8,
    'lamb1': 0.5,
    'rad': 2.0,
    'P': 0.5,
    'bisearch_precision': 0.001,
    'bisearch_boundary': 10.0,
    'eps': 1e-4,
    'L': -5.0,
    'R': 5.0,
}
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 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
fast_beta Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("cc016e2d63decb65")
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, 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
from scipy.stats import beta as Beta

torch.manual_seed(0)

# Analyze the function signature and body:
# - p: used in comparisons (== 1, <= 300) and passed to Beta(p, p)
#   So p should be a numeric type (int or float)
# - x: used in comparisons (min/max with 1.0 and 0.0, <= comparisons)
#   and passed to Beta.cdf(), so x should be a float in [0, 1]

# Test cases covering different branches:
# 1. p == 1 case
# 2. p <= 300 or x in [FAST_BETA_TH, 1.0 - FAST_BETA_TH] case
# 3. x < FAST_BETA_TH case
# 4. else case (x > 1.0 - FAST_BETA_TH and p > 300)

# For a single execution, pick one scenario
p_value = 2  # satisfies p <= 300, will use Beta CDF
x_value = 0.5  # in the valid range [0, 1]

ARGS = {
    "p": p_value,
    "x": x_value,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — 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, 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
lambertWlog Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("331144e4c91f00b9")
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, 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
import scipy

ARGS = {
    'logx': 100.0,
    'mode': 'lowerbound'
}
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, 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
ln_exp_plus_exp Ran
ran, honoured the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("431e5dacd9481349")
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, 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

# Construct arguments for ln_exp_plus_exp
# The function takes: sgn1 (int), pow1 (float), sgn2 (int), pow2 (float), EPS (float)

ARGS = {
    'sgn1': 1,
    'pow1': 5.0,
    'sgn2': 1,
    'pow2': 3.0,
    'EPS': 1e-06
}
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, 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
sum_exp_greater_than_one Ran
ran, violated the contract we drafted
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("781b5c0335c1dea9")
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 numpy as np

# The function takes 4 scalar parameters: sgn1, pow1, sgn2, pow2
# sgn1 and sgn2 appear to be sign multipliers (scalar numbers)
# pow1 and pow2 appear to be power/exponent values (scalar numbers)

# Let's construct valid arguments for different test cases

# Test case 1: Both exponents in safe range [-20, 20]
sgn1 = 1.0
pow1 = 0.5
sgn2 = 1.0
pow2 = 1.5

ARGS = {
    "sgn1": sgn1,
    "pow1": pow1,
    "sgn2": sgn2,
    "pow2": pow2
}
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 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
check Not yet run
unverified (harvested, no recorded run)
dsgiitr/re_dsrs/dsrs/algo/algo.py
code served (permissive licence) · get_code("20f04d1a583d6fc1")
How it ran

no invocation recorded.

check Not yet run
unverified (harvested, no recorded run)
llylly/dsrs/algo/algo.py
code served (permissive licence) · get_code("e2025ff88e86c006")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 4 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 4 later) and 13 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
fast_beta
brick of this paper's assembly: helper
1 other paper
1 later
2406.02309 (2024-06)
Effects of Exponential Gaussian Distribution on (Double Sampling) Randomized Smoothing
lambertWlog
brick of this paper's assembly: helper
1 other paper
1 later
2406.02309 (2024-06)
Effects of Exponential Gaussian Distribution on (Double Sampling) Randomized Smoothing
ln_exp_plus_exp
brick of this paper's assembly: helper
1 other paper
1 later
2406.02309 (2024-06)
Effects of Exponential Gaussian Distribution on (Double Sampling) Randomized Smoothing
sum_exp_greater_than_one
brick of this paper's assembly: helper
1 other paper
1 later
2406.02309 (2024-06)
Effects of Exponential Gaussian Distribution on (Double Sampling) Randomized Smoothing

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

Neural networks (NNs) are known to be vulnerable against adversarial perturbations, and thus there is a line of work aiming to provide robustness certification for NNs, such as randomized smoothing, which samples smoothing noises from a certain distribution to certify the robustness for a smoothed classifier. However, as shown by previous work, the certified robust radius in randomized smoothing suffers from scaling to large datasets ("curse of dimensionality"). To overcome this hurdle, we propose a Double Sampling Randomized Smoothing (DSRS) framework, which exploits the sampled probability from an additional smoothing distribution to tighten the robustness certification of the previous smoothed classifier. Theoretically, under mild assumptions, we prove that DSRS can certify Θ( √ d) robust radius under 2 norm where d is the input dimension, implying that DSRS may be able to break the curse of dimensionality of randomized smoothing. We instantiate DSRS for a generalized family of Gaussian smoothing and propose an efficient and sound computing method based on customized dual optimization considering sampling error. Extensive experiments on MNIST, CIFAR-10, and ImageNet verify our theory and show that DSRS certifies larger robust radii than existing baselines consistently under different settings. Code is available at https: //github.com/llylly/DSRS.

For agents

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

get_harvested_code_for_paper("2206.07912")
get_code_for_paper("2206.07912")
have("2206.07912")

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

curl https://syntology.ai/api/ran/2206.07912.json

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

[![Syntology run record](https://syntology.ai/api/ran/2206.07912.svg)](https://syntology.ai/paper/2206.07912)

Connect an agent — have() is free.