SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2202.01273 · ICML · 2022

Beyond Images: Label Noise Transition Matrix Estimation for Tasks with Lower-Quality Features

Yang Liu, Zhaowei Zhu, Jialu Wang

arXiv · PDF · Open in the Atlas

Code that ran

14 of 25 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 6 failures of our instrument; 11 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/2202.01273.json · badge: /api/ran/2202.01273.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
ucsc-real/est-t-mi canonical (classifier label)
classifier basis not recorded on the link
13 ran · 4 no recorded run
ucsc-real/beyondimages canonical (classifier label)
classifier basis: classifier_2025
1 ran · 7 no recorded run
FunctionStatusWhere it lives
cosDistance Ran
ran; our contract draft was wrong, not the code
ucsc-real/est-t-mi/run_nn.py
pointer only (licence: MIT); the stored copy get_code serves, from docta-ai/docta, is pointer only (licence: NOASSERTION) · get_code("d1f5c578a50e0457")
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 docta-ai/docta, and GitHub could not classify the licence of that repository (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.

cosDistance Ran
ran; our contract draft was wrong, not the code
ucsc-real/est-t-mi/myutils.py
pointer only (licence: MIT); the stored copy get_code serves, from Docta-ai/docta, is pointer only (licence: NOASSERTION) · get_code("1aefb5a2a8344cf3")
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 Docta-ai/docta, and GitHub could not classify the licence of that repository (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.

count_2nn_acc Ran
ran on a synthesized input
ucsc-real/est-t-mi/run_nn.py
code served (permissive licence) · get_code("4ae9e39e851444d5")
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. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Parameters for count_2nn_acc
KINDS = 3  # number of classes
cluster_sum = 5  # number of samples

# Create feature coordinates: (cluster_sum, feature_dim)
feat_cord = torch.randn(cluster_sum, 4)

# Create labels: each sample gets a label from 0 to KINDS-1
label = torch.tensor([0, 1, 2, 0, 1], dtype=torch.long)

ARGS = {
    "KINDS": KINDS,
    "feat_cord": feat_cord,
    "label": label,
    "cluster_sum": cluster_sum,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list 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
entropy Ran
ran on a synthesized input
ucsc-real/est-t-mi/utils_mi.py
code served (permissive licence) · get_code("0b72d6d48231f171")
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 scipy.special import gamma, psi
from numpy import pi
from sklearn.neighbors import NearestNeighbors

# Create a small random dataset
np.random.seed(0)
X = np.random.randn(8, 3).astype(np.float64)

# Define ARGS mapping every parameter of entropy to a value
ARGS = {
    'X': X,
    'k': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, 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
extract_feature Ran
ran on a synthesized input
ucsc-real/est-t-mi/hoc.py
code served (permissive licence) · get_code("c2f96fe6354bb1fb")
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 dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Define a minimal model that returns (representation, prediction) tuple
class DummyExtractorModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 3)
    
    def forward(self, *args):
        # Handle both feature input and (text, offsets) input
        if len(args) == 1:
            feature = args[0]
        else:
            feature = args[0]  # text argument, we'll just use it as-is
        
        representation = self.fc(feature)
        prediction = torch.zeros(feature.shape[0], 2)  # dummy prediction
        return representation, prediction

# Create model instance
model_extract = DummyExtractorModel()

# Create a simple dataloader with 3-element tuples: (noisy_label, clean_label, features)
# Batch size 2, feature dim 4
batch_size = 2
feature_dim = 4
num_batches = 3

dataloader = []
for i in range(num_batches):
    noisy_labels = torch.tensor([0, 1], dtype=torch.long)
    clean_labels = torch.tensor([0, 1], dtype=torch.long)
    features = torch.randn(batch_size, feature_dim)
    dataloader.append((noisy_labels, clean_labels, features))

ARGS = {
    'model_extract': model_extract,
    'dataloader': dataloader,
    'start_from_1': False
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
generate_noise_matrix_from_diagonal Ran
ran on a synthesized input
ucsc-real/est-t-mi/myutils.py
code served (permissive licence) · get_code("a802dce412324d52")
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 (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

np.random.seed(0)

# The function expects a 1D numpy array 'diag' with diagonal values
# K = diag.shape[0] determines the size of the noise matrix
# The values should be between 0 and 1 (they represent diagonal probabilities)
diag = np.array([0.7, 0.8, 0.6, 0.75])

ARGS = {
    "diag": diag
}
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 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
make_arg_list Ran
ran on a synthesized input
ucsc-real/est-t-mi/utils/misc.py
code served (permissive licence) · get_code("a014abf3410a9ac7")
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 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

ARGS = {
    'arg': {
        'seeds': 3
    }
}
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 list, ndim 1, 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
make_noisy_data Ran
ran on a synthesized input
ucsc-real/est-t-mi/utils/dataloader.py
code served (permissive licence) · get_code("7f8fa58e6835ca57")
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 int, shape (10,) 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 simple binary classification label array
# with some 0s and 1s for testing
np.random.seed(0)
y = np.array([0, 0, 0, 1, 1, 1, 0, 1, 0, 1])

# Define noise rates for class 0 and class 1
e0 = 0.2  # 20% of class 0 samples will be flipped
e1 = 0.3  # 30% of class 1 samples will be flipped

ARGS = {
    'y': y,
    'e0': e0,
    'e1': e1
}
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 ndarray, ndim 1, dtype family int, shape (10,) — shape (10,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_output_format Ran
ran; our contract draft was wrong, not the code
ucsc-real/est-t-mi/utils/logger.py
code served (permissive licence) · get_code("5eee517402b390d5")
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 HumanOutputFormat. 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 os
import sys
import tempfile
import shutil

# Create a temporary directory for output
temp_dir = tempfile.mkdtemp()

# Define arguments for make_output_format function
ARGS = {
    'format': 'stdout',
    'ev_dir': temp_dir,
    'log_suffix': '_test'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type HumanOutputFormat — a HumanOutputFormat output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a HumanOutputFormat output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mutual_information Ran
ran on a synthesized input
ucsc-real/est-t-mi/utils_mi.py
code served (permissive licence) · get_code("9cf29a68b1e2cbf5")
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 seed for reproducibility
np.random.seed(0)

# Create sample data for mutual information calculation
# Variables should be tuples of arrays, each with shape (n_samples, n_features)
# Using small but distinct dimensions for clarity

# Variable 1: 10 samples, 2 features
X = np.random.randn(10, 2)

# Variable 2: 10 samples, 3 features
Y = np.random.randn(10, 3)

# Variable 3: 10 samples, 4 features (optional, for testing with 3 variables)
Z = np.random.randn(10, 4)

# Create the variables tuple - at least 2 variables required
variables = (X, Y)

# Parameter k for nearest neighbors
k = 1

ARGS = {
    'variables': variables,
    'k': k
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, 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
nearest_distances Ran
ran on a synthesized input
ucsc-real/est-t-mi/utils_mi.py
code served (permissive licence) · get_code("149d5dc30e60227f")
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 (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
from sklearn.neighbors import NearestNeighbors

# Create a small sample dataset
# N=2 points, M=4 dimensions
np.random.seed(0)
X = np.random.randn(2, 4).astype(np.float32)

# Define arguments for the function
ARGS = {
    "X": X,
    "k": 1
}
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 ndarray, ndim 1, dtype family float, shape (2,) — shape (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
onehot Ran
ran; our contract draft was wrong, not the code
ucsc-real/beyondimages/utils/dataloader.py
code served (permissive licence) · get_code("8ea81c57e23ddc38")
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 numpy as np
import pandas as pd

# Create a small test dataframe with some categorical columns
df = pd.DataFrame({
    'numeric_col': [1.5, 2.5, 3.5, 4.5],
    'cat_col1': ['a', 'b', 'a', 'c'],
    'cat_col2': ['x', 'y', 'x', 'y'],
    'cat_col3': ['p', 'q', 'r', 'p']
})

cols = ['cat_col1', 'cat_col2']

ARGS = {
    'df': df,
    'cols': cols
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas, sklearn — 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
onehot Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ucsc-real/est-t-mi/utils/dataloader.py
pointer only (licence: MIT); the stored copy get_code serves, from Docta-ai/docta, is pointer only (licence: NOASSERTION) · get_code("ef874310d3110242")
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 Docta-ai/docta, and GitHub could not classify the licence of that repository (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.

profile Ran
ran; our contract draft was wrong, not the code
ucsc-real/est-t-mi/utils/logger.py
pointer only (licence: MIT); the stored copy get_code serves, from facebookresearch/gen_dgrl, is pointer only (licence: NOASSERTION) · get_code("98948fae5be7e92e")
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 facebookresearch/gen_dgrl, and GitHub could not classify the licence of that repository (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.

calc_func Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/hoc.py
code served (permissive licence) · get_code("95864ec98b0352e1")
How it ran

no invocation recorded.

count_y Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/hoc.py
code served (permissive licence) · get_code("3c35d31a8ca77746")
How it ran

no invocation recorded.

estimate_T_HOC_weight Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/hoc.py
code served (permissive licence) · get_code("7e9d232de2e93484")
How it ran

no invocation recorded.

func Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/hoc.py
code served (permissive licence) · get_code("bd60bff3f7c47a28")
How it ran

no invocation recorded.

get_T_global_min_new Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/hoc.py
code served (permissive licence) · get_code("683f716ca3bcbbb1")
How it ran

no invocation recorded.

get_T_global_min_new Not yet run
unverified (harvested, no recorded run)
ucsc-real/est-t-mi/hoc.py
code served (permissive licence) · get_code("95ed1a9915617f74")
How it ran

no invocation recorded.

get_T_global_min_new Not yet run
unverified (harvested, no recorded run)
ucsc-real/est-t-mi/run_nn.py
code served (permissive licence) · get_code("568a24d3fdc2ebcb")
How it ran

no invocation recorded.

make_noisy_data Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/utils/dataloader.py
code served (permissive licence) · get_code("df4e2a034cb6af96")
How it ran

no invocation recorded. this sample carries JSON test cases instead of a setup block; get_reference_implementation serves them with the level they back.

mutual_information_2d_classification Not yet run
unverified (harvested, no recorded run)
ucsc-real/beyondimages/hoc.py
code served (permissive licence) · get_code("0f3ab4a0f758b991")
How it ran

no invocation recorded.

noisify_multiclass_symmetric Not yet run
unverified (harvested, no recorded run)
ucsc-real/est-t-mi/myutils.py
code served (permissive licence) · get_code("7dddda47015b8708")
How it ran

no invocation recorded.

read_json Not yet run
unverified (harvested, no recorded run)
ucsc-real/est-t-mi/utils/logger.py
code served (permissive licence) · get_code("88ef5a17fc986d9a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 25 samples lifted from this paper's linked repositories, 25 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 5 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 5 later) and 16 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
make_output_format 73 other papers
21 dated earlier, 52 later
1611.05763 (2016-11)
Learning to reinforcement learn
1804.06459 (2018-04)
On Learning Intrinsic Rewards for Policy Gradient Methods
1806.02473 (2018-06)
Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
and 70 more not named here
read_json 22 other papers
16 dated earlier, 6 later
1611.05763 (2016-11)
Learning to reinforcement learn
1711.00123 (2017-11)
Backpropagation through the Void: Optimizing control variates for black-box gradient estimation
1804.06459 (2018-04)
On Learning Intrinsic Rewards for Policy Gradient Methods
and 19 more not named here
profile 13 other papers
9 dated earlier, 4 later
1804.06459 (2018-04)
On Learning Intrinsic Rewards for Policy Gradient Methods
1806.02473 (2018-06)
Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
1806.07857 (2018-06)
RUDDER: Return Decomposition for Delayed Rewards
and 10 more not named here
cosDistance 1 other paper
1 later
2311.11202 (2023-11)
Unmasking and Improving Data Credibility: A Study with Datasets for Training Harmless Language Models
onehot 1 other paper
1 later
2311.11202 (2023-11)
Unmasking and Improving Data Credibility: A Study with Datasets for Training Harmless Language Models

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

The label noise transition matrix, denoting the transition probabilities from clean labels to noisy labels, is crucial for designing statistically robust solutions. Existing estimators for noise transition matrices, e.g., using either anchor points or clusterability, focus on computer vision tasks that are relatively easier to obtain high-quality representations. We observe that tasks with lowerquality features fail to meet the anchor-point or clusterability condition, due to the coexistence of both uninformative and informative representations. To handle this issue, we propose a generic and practical information-theoretic approach to down-weight the less informative parts of the lower-quality features. This improvement is crucial to identifying and estimating the label noise transition matrix. The salient technical challenge is to compute the relevant informationtheoretical metrics using only noisy labels instead of clean ones. We prove that the celebrated f -mutual information measure can often preserve the order when calculated using noisy labels. We then build our transition matrix estimator using this distilled version of features. The necessity and effectiveness of the proposed method are also demonstrated by evaluating the estimation error on a varied set of tabular data and text classification tasks with lower-quality features. Code is available at github.com/UCSC-REAL/BeyondImages.

For agents

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

get_harvested_code_for_paper("2202.01273")
get_code_for_paper("2202.01273")
have("2202.01273")

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

curl https://syntology.ai/api/ran/2202.01273.json

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

[![Syntology run record](https://syntology.ai/api/ran/2202.01273.svg)](https://syntology.ai/paper/2202.01273)

Connect an agent — have() is free.