SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1909.01441 · 2019

CrossWeigh: Training Named Entity Tagger from Imperfect Annotations

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 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, 1 violated it and 0 have no contract verdict; 3 failures of our instrument; 0 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/1909.01441.json · badge: /api/ran/1909.01441.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
ZihanWangKi/CrossWeigh canonical (classifier label)
classifier basis: classifier_2025
4 ran · 0 no recorded run
FunctionStatusWhere it lives
get_tokens_and_labels Ran
ran; our contract draft was wrong, not the code
ZihanWangKi/CrossWeigh/flair_scripts/flair_ner.py
code served (permissive licence) · get_code("0b89e60ae07e6da6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned tuple, ndim 2, 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.

class Token:
    def __init__(self, text, ner_tag):
        self.text = text
        self._ner_tag = ner_tag
    
    def get_tag(self, tag_type):
        if tag_type == 'ner':
            return self._ner_tag
        return None

class Tag:
    def __init__(self, value):
        self.value = value

class Sentence:
    def __init__(self, tokens):
        self.tokens = tokens

token1 = Token("John", Tag("PERSON"))
token2 = Token("Smith", Tag("PERSON"))
token3 = Token("works", Tag("O"))
token4 = Token("at", Tag("O"))
token5 = Token("Google", Tag("ORG"))

sentence = Sentence([token1, token2, token3, token4, token5])

ARGS = {
    "sentence": sentence
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
iob2bio Ran
ran; our contract draft was wrong, not the code
ZihanWangKi/CrossWeigh/split.py
code served (permissive licence) · get_code("a585c1cff9abade3")
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 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
import numpy as np

# Analyze the function: iob2bio takes iob_labels as input
# It iterates through pairs of (prev_label, cur_label)
# Each label is indexed with [0] and [2:], so labels are strings like "I-PER", "O", "B-LOC"
# The function converts IOB format to BIO format

# Create sample IOB labels (IOB2 format where I can follow O)
iob_labels = [
    'O',
    'I-PER',      # Should become B-PER (I follows O)
    'I-PER',      # Should stay I-PER (same entity type)
    'I-LOC',      # Should become B-LOC (different entity type)
    'O',
    'I-ORG',      # Should become B-ORG (I follows O)
]

ARGS = {
    'iob_labels': iob_labels
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
iobes2bio Ran
ran, violated the contract we drafted
ZihanWangKi/CrossWeigh/flair_scripts/flair_ner.py
code served (permissive licence) · get_code("acfbfd4e0f76436a")
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 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
import numpy as np

torch.manual_seed(0)

# The function iobes2bio takes a list of IOBES labels (strings)
# and converts them to BIO labels by:
# - Converting 'S' (single) tags to 'B' (begin) tags
# - Converting 'E' (end) tags to 'I' (inside) tags
# - Keeping all other tags unchanged

iobes_labels = [
    'B-PER',
    'I-PER',
    'E-PER',
    'S-LOC',
    'B-ORG',
    'I-ORG',
    'E-ORG',
    'O'
]

ARGS = {
    'iobes_labels': iobes_labels
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
load_dataset_from_column Ran
ran; our contract draft was wrong, not the code
ZihanWangKi/CrossWeigh/split.py
code served (permissive licence) · get_code("982e0d453a27692e")
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 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 tempfile
import os

# Create a temporary file with sample data in column format
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt', encoding='utf-8')

# Write sample BIO format data
sample_data = """John B-PER
Smith I-PER
works O
at O
Google B-ORG
.	O

Mary B-PER
is O
a O
doctor O
.	O

-DOCSTART- O

The O
quick O
brown B-MISC
fox O
.	O
"""

temp_file.write(sample_data)
temp_file.close()

ARGS = {
    'path': temp_file.name,
    'schema': 'bio'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 4 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

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

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

Repositories linked to this paper

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

Abstract

Everyone makes mistakes. So do human annotators when curating labels for named entity recognition (NER). Such label mistakes might hurt model training and interfere model comparison. In this study, we dive deep into one of the widely-adopted NER benchmark datasets, CoNLL03 NER. We are able to identify label mistakes in about 5.38% test sentences, which is a significant ratio considering that the state-of-the-art test F1 score is already around 93%. Therefore, we manually correct these label mistakes and form a cleaner test set. Our re-evaluation of popular models on this corrected test set leads to more accurate assessments, compared to those on the original test set. More importantly, we propose a simple yet effective framework, CrossWeigh, to handle label mistakes during NER model training. Specifically, it partitions the training data into several folds and train independent NER models to identify potential mistakes in each fold. Then it adjusts the weights of training data accordingly to train the final NER model. Extensive experiments demonstrate significant improvements of plugging various NER models into our proposed framework on three datasets. All implementations and corrected test set are available at our Github repo: https://github.com/ZihanWangKi/CrossWeigh.

For agents

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

get_harvested_code_for_paper("1909.01441")
get_code_for_paper("1909.01441")
have("1909.01441")

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

curl https://syntology.ai/api/ran/1909.01441.json

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

[![Syntology run record](https://syntology.ai/api/ran/1909.01441.svg)](https://syntology.ai/paper/1909.01441)

Connect an agent — have() is free.