SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2205.12644 · 2022

LingMess: Linguistically Informed Multi Expert Scorers for Coreference Resolution

arXiv · PDF · Open in the Atlas

Code that ran

11 of 11 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; 3 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 1 failure of our instrument; 0 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, 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/2205.12644.json · badge: /api/ran/2205.12644.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
shon-otmazgin/lingmess-coref canonical (classifier label)
classifier basis not recorded on the link
6 ran · 0 no recorded run
shon-otmazgin/fastcoref community (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
FunctionStatusWhere it lives
flatten Ran
ran, honoured the contract we drafted
shon-otmazgin/fastcoref/fastcoref/utilities/util.py
code served (permissive licence) · get_code("20922c79a0929bbf")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T22:50:02Z). 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 flatten takes a list of lists and flattens it into a single list
# It iterates through sublists and extracts items from each

# Create a simple nested list structure
nested_list = [
    [1, 2, 3],
    [4, 5],
    [6, 7, 8, 9]
]

ARGS = {
    "l": nested_list
}
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 int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
align_clusters Ran
ran on a synthesized input
shon-otmazgin/lingmess-coref/util.py
code served (permissive licence) · get_code("a2f69b15b51e5874")
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

torch.manual_seed(0)

# clusters: list of clusters, where each cluster is a list of [start, end] pairs
clusters = [
    [[0, 1], [2, 3]],
    [[1, 2]],
    [[0, 3], [1, 2]]
]

# subtoken_maps: list/dict mapping subtoken indices to new indices (or None)
# The function accesses subtoken_maps[start] and subtoken_maps[end]
# Creating a list that maps indices to new indices or None
subtoken_maps = [0, 1, 2, 3, None, 1]

# new_word_maps: list/dict mapping indices to final word indices
# The function accesses new_word_maps[start] and new_word_maps[end]
new_word_maps = [10, 20, 30, 40, 50]

ARGS = {
    "clusters": clusters,
    "subtoken_maps": subtoken_maps,
    "new_word_maps": new_word_maps
}
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 — 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
b_cubed Ran
ran on a synthesized input
shon-otmazgin/fastcoref/fastcoref/utilities/metrics.py
code served (permissive licence) · get_code("3643e9b2dc2db0a3")
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 tuple, 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.

from collections import Counter

def b_cubed(clusters, mention_to_gold):
    num, dem = (0, 0)
    for c in clusters:
        if len(c) == 1:
            continue
        gold_counts = Counter()
        correct = 0
        for m in c:
            if m in mention_to_gold:
                gold_counts[tuple(mention_to_gold[m])] += 1
        for c2, count in gold_counts.items():
            if len(c2) != 1:
                correct += count * count
        num += correct / float(len(c))
        dem += len(c)
    return (num, dem)


# Setup code
ARGS = {
    "clusters": [
        ["mention_0", "mention_1", "mention_2"],
        ["mention_3", "mention_4"],
        ["mention_5"]
    ],
    "mention_to_gold": {
        "mention_0": ("entity_A", "entity_B"),
        "mention_1": ("entity_A", "entity_B"),
        "mention_2": ("entity_A", "entity_B"),
        "mention_3": ("entity_C", "entity_D"),
        "mention_4": ("entity_C", "entity_D"),
        "mention_5": ("entity_E",),
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 1, 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
create_batches Ran
ran on a synthesized input
shon-otmazgin/lingmess-coref/coref_dataset.py
code served (permissive licence) · get_code("1419e13af2d888ce")
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 Dataset. 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 logging
from collections import defaultdict
from datasets import Dataset
from torch.utils.data import DataLoader, IterableDataset

torch.manual_seed(0)

logging.basicConfig(level=logging.INFO)

class MockDataset(IterableDataset):
    def __init__(self, num_samples=2):
        self.num_samples = num_samples
        self.dataset = self
        
    def __iter__(self):
        for i in range(self.num_samples):
            yield {
                'input_ids': torch.tensor([1, 2, 3, 4]),
                'attention_mask': torch.tensor([1, 1, 1, 0]),
                'label': torch.tensor([i % 2])
            }
    
    def __len__(self):
        return self.num_samples

class MockSampler:
    def __init__(self, dataset):
        self.dataset = dataset
    
    def __iter__(self):
        for i in range(len(self.dataset)):
            yield {
                'input_ids': torch.tensor([1, 2, 3, 4]),
                'attention_mask': torch.tensor([1, 1, 1, 0]),
                'label': torch.tensor([i % 2])
            }

mock_dataset = MockDataset(num_samples=2)
mock_sampler = MockSampler(mock_dataset)

ARGS = {
    'sampler': mock_sampler,
    'path_to_save': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
datasets, torch, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Dataset — a Dataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Dataset output the recorded ndim 1 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
f1 Ran
ran, honoured the contract we drafted
shon-otmazgin/fastcoref/fastcoref/utilities/metrics.py
pointer only (licence: MIT); the stored copy get_code serves, from sapienzanlp/bookcoref, is pointer only (licence: NONE) · get_code("5dc251a1bb16c94f")
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 sapienzanlp/bookcoref, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_doc_key Ran
ran; our contract draft was wrong, not the code
shon-otmazgin/lingmess-coref/prepare_ontonotes/conll.py
code served (permissive licence) · get_code("f300a33fe05fc609")
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 str. 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

def get_doc_key(doc_id, part):
    return '{}_{}'.format(doc_id, int(part))

ARGS = {
    'doc_id': 'doc_123',
    'part': 5
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
muc Ran
ran, honoured the contract we drafted
shon-otmazgin/fastcoref/fastcoref/utilities/metrics.py
pointer only (licence: MIT); the stored copy get_code serves, from sapienzanlp/bookcoref, is pointer only (licence: NONE) · get_code("8b30464b3ec47d85")
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 sapienzanlp/bookcoref, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

normalize_word Ran
ran on a synthesized input
shon-otmazgin/lingmess-coref/prepare_ontonotes/minimize.py
code served (permissive licence) · get_code("3b625164c1147fcf")
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 str. 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.

ARGS = {
    'word': '/.',
    'language': 'arabic'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
output_evaluation_metrics Ran
ran on a synthesized input
shon-otmazgin/lingmess-coref/util.py
code served (permissive licence) · get_code("190a77cd99d3e249")
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 logging
import os
import tempfile

logger = logging.getLogger(__name__)

class MockMetricCalculator:
    def __init__(self, precision, recall, f1):
        self.precision = precision
        self.recall = recall
        self.f1 = f1
    
    def get_prf(self):
        return self.precision, self.recall, self.f1

class MockCorefCategories:
    def get_stats(self):
        return {
            'category_a': 0.85,
            'category_b': 0.92,
            'category_info': {'nested': 'value'}
        }

metrics_dict = {
    'loss': 0.456,
    'post_pruning': MockMetricCalculator(0.78, 0.82, 0.80),
    'mentions': MockMetricCalculator(0.85, 0.88, 0.865),
    'coref': MockMetricCalculator(0.91, 0.89, 0.90),
    'coref_categories': MockCorefCategories()
}

output_dir = tempfile.mkdtemp()

prefix = "validation"

ARGS = {
    'metrics_dict': metrics_dict,
    'output_dir': output_dir,
    'prefix': prefix
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 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
pad_clusters_outside Ran
ran on a synthesized input
shon-otmazgin/fastcoref/fastcoref/utilities/util.py
code served (permissive licence) · get_code("616fe9a061b0fa0c")
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)

# Analyze the function:
# - clusters: a list of lists (each element can be a list, likely representing clusters)
# - max_num_clusters: an integer specifying the maximum number of clusters
# The function pads clusters with empty lists to reach max_num_clusters length

# Construct valid arguments
clusters = [
    [1, 2, 3],
    [4, 5]
]

max_num_clusters = 5

ARGS = {
    "clusters": clusters,
    "max_num_clusters": max_num_clusters
}
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
resolve_doc_key Ran
ran on a synthesized input
shon-otmazgin/lingmess-coref/prepare_ontonotes/conll.py
code served (permissive licence) · get_code("4b0a5db027cc39ee")
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 tuple, 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

def resolve_doc_key(doc_key):
    parts = doc_key.split('_')
    return ('_'.join(parts[:-1]), int(parts[-1]))

ARGS = {
    'doc_key': 'document_name_123'
}
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 tuple, ndim 1, 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

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 7 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 1 the same month, 6 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.

FunctionAlso present inEarliest-datedOthers named
flatten 21 other papers
10 dated earlier, 10 later, 1 same year, order not determinable
1609.08675 (2016-09)
YouTube-8M: A Large-Scale Video Classification Benchmark
1802.04821 (2018-02)
Evolved Policy Gradients
1811.05014 (2018-11)
NeXtVLAD: An Efficient Neural Network to Aggregate Frame-level Features for Large-scale Video Classification
and 18 more not named here
f1 12 other papers
5 dated earlier, 1 the same month, 6 later
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1907.10529 (2019-07)
SpanBERT: Improving Pre-training by Representing and Predicting Spans
2021.naacl-main.125 (2021 (year only))
Incorporating Syntax and Semantics in Coreference Resolution with Heterogeneous Graph Attention Network
and 9 more not named here
get_doc_key 8 other papers
5 dated earlier, 3 later
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1907.10529 (2019-07)
SpanBERT: Improving Pre-training by Representing and Predicting Spans
2021.naacl-main.125 (2021 (year only))
Incorporating Syntax and Semantics in Coreference Resolution with Heterogeneous Graph Attention Network
and 5 more not named here
normalize_word 4 other papers
4 dated earlier
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1907.10529 (2019-07)
SpanBERT: Improving Pre-training by Representing and Predicting Spans
2021.naacl-main.125 (2021 (year only))
Incorporating Syntax and Semantics in Coreference Resolution with Heterogeneous Graph Attention Network
and 1 more not named here
b_cubed 3 other papers
2 dated earlier, 1 later
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
2202.10710 (2022-02)
Incorporating Constituent Syntax for Coreference Resolution
2024.acl-srw.56 (2024 (year only))
Z-coref: Thai Coreference and Zero Pronoun Resolution
muc 3 other papers
3 later
2024.acl-srw.56 (2024 (year only))
Z-coref: Thai Coreference and Zero Pronoun Resolution
2407.21489 (2024-07)
Maverick: Efficient and Accurate Coreference Resolution Defying Recent Trends
2507.12075 (2025-07)
BOOKCOREF: Coreference Resolution at Book Scale
pad_clusters_outside 1 other paper
1 later
2024.acl-srw.56 (2024 (year only))
Z-coref: Thai Coreference and Zero Pronoun Resolution

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

While coreference resolution typically involves various linguistic challenges, recent models are based on a single pairwise scorer for all types of pairs. We present LingMess, a new coreference model that defines different categories of coreference cases and optimize multiple pairwise scorers, where each scorer learns a specific set of linguistic challenges. Our model substantially improves pairwise scores for most categories and outperforms cluster-level performance on Ontonotes and 5 additional datasets. Our model is available in https://github.com/shon-otmazgin/lingmess-coref

For agents

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

get_harvested_code_for_paper("2205.12644")
get_code_for_paper("2205.12644")
have("2205.12644")

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

curl https://syntology.ai/api/ran/2205.12644.json

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

[![Syntology run record](https://syntology.ai/api/ran/2205.12644.svg)](https://syntology.ai/paper/2205.12644)

Connect an agent — have() is free.