SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.09578 · 2018

Why Comparing Single Performance Scores Does Not Allow to Draw Conclusions About Machine Learning Approaches

arXiv · PDF · Open in the Atlas

Code that ran

2 of 6 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 1 has no contract verdict; 1 failure of our instrument; 4 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/1803.09578.json · badge: /api/ran/1803.09578.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
UKPLab/emnlp2017-bilstm-cnn-crf canonical (classifier label)
classifier basis: census
2 ran · 4 no recorded run
FunctionStatusWhere it lives
compute_f1 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
UKPLab/emnlp2017-bilstm-cnn-crf/util/BIOF1Validation.py
code served (permissive licence) · get_code("2561978b0fd1e5d3")
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
import numpy as np

torch.manual_seed(0)

# Create small prediction and correct label sequences
# predictions: list of sentences, each sentence is a list of label indices
predictions = [
    [0, 1, 2, 0],  # sentence 1: 4 tokens with label indices
    [1, 1, 0, 2],  # sentence 2: 4 tokens with label indices
]

# correct: list of sentences, each sentence is a list of label indices
correct = [
    [0, 1, 2, 0],  # sentence 1: 4 tokens with label indices
    [1, 1, 0, 2],  # sentence 2: 4 tokens with label indices
]

# idx2Label: mapping from label index to label string
# Must include BIO encoding tags
idx2Label = {
    0: 'O',
    1: 'B-PER',
    2: 'I-PER',
}

# Test with different encoding schemes
correctBIOErrors = 'No'
encodingScheme = 'BIO'

ARGS = {
    'predictions': predictions,
    'correct': correct,
    'idx2Label': idx2Label,
    'correctBIOErrors': correctBIOErrors,
    'encodingScheme': encodingScheme,
}
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 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
compute_precision_token_basis Ran
ran on a synthesized input
UKPLab/emnlp2017-bilstm-cnn-crf/util/BIOF1Validation.py
code served (permissive licence) · get_code("6e5cd3fa57bb3c06")
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 torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - guessed_sentences: list of sentences, each sentence is a list of labels
# - correct_sentences: list of sentences, each sentence is a list of labels
# - O_Label: a label value to filter out (typically "O" for outside in NER)
# The function compares tokens between guessed and correct sentences,
# counting precision for non-O labels

# Create sample data
guessed_sentences = [
    ["B-PER", "I-PER", "O", "B-LOC"],
    ["O", "B-ORG", "I-ORG", "O"]
]

correct_sentences = [
    ["B-PER", "I-PER", "O", "B-LOC"],
    ["O", "B-ORG", "B-ORG", "O"]
]

O_Label = "O"

ARGS = {
    "guessed_sentences": guessed_sentences,
    "correct_sentences": correct_sentences,
    "O_Label": O_Label
}
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 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
compute_f1_token_basis Not yet run
unverified (harvested, no recorded run)
UKPLab/emnlp2017-bilstm-cnn-crf/util/BIOF1Validation.py
code served (permissive licence) · get_code("1b66320c71babcd2")
How it ran

no invocation recorded.

path_energy Not yet run
unverified (harvested, no recorded run)
UKPLab/emnlp2017-bilstm-cnn-crf/neuralnets/keraslayers/ChainCRF.py
code served (permissive licence) · get_code("6f5b259e1a881655")
How it ran

no invocation recorded.

path_energy0 Not yet run
unverified (harvested, no recorded run)
UKPLab/emnlp2017-bilstm-cnn-crf/neuralnets/keraslayers/ChainCRF.py
code served (permissive licence) · get_code("e13517b914ec8cfe")
How it ran

no invocation recorded.

sparse_chain_crf_loss Not yet run
unverified (harvested, no recorded run)
UKPLab/emnlp2017-bilstm-cnn-crf/neuralnets/keraslayers/ChainCRF.py
code served (permissive licence) · get_code("29092b6b1ee0e8b7")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 0 later) and 0 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
path_energy 2 other papers
2 dated earlier
1603.01360 (2016-03)
Neural Architectures for Named Entity Recognition
1802.05365 (2018-02)
Deep contextualized word representations
path_energy0 2 other papers
2 dated earlier
1603.01360 (2016-03)
Neural Architectures for Named Entity Recognition
1802.05365 (2018-02)
Deep contextualized word representations
sparse_chain_crf_loss 2 other papers
2 dated earlier
1603.01360 (2016-03)
Neural Architectures for Named Entity Recognition
1802.05365 (2018-02)
Deep contextualized word representations
compute_f1 1 other paper
1 dated earlier
1802.05365 (2018-02)
Deep contextualized word representations
compute_f1_token_basis 1 other paper
1 dated earlier
1802.05365 (2018-02)
Deep contextualized word representations
compute_precision_token_basis 1 other paper
1 dated earlier
1802.05365 (2018-02)
Deep contextualized word representations

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

Developing state-of-the-art approaches for specific tasks is a major driving force in our research community. Depending on the prestige of the task, publishing it can come along with a lot of visibility. The question arises how reliable are our evaluation methodologies to compare approaches? One common methodology to identify the state-of-the-art is to partition data into a train, a development and a test set. Researchers can train and tune their approach on some part of the dataset and then select the model that worked best on the development set for a final evaluation on unseen test data. Test scores from different approaches are compared, and performance differences are tested for statistical significance. In this publication, we show that there is a high risk that a statistical significance in this type of evaluation is not due to a superior learning approach. Instead, there is a high risk that the difference is due to chance. For example for the CoNLL 2003 NER dataset we observed in up to 26% of the cases type I errors (false positives) with a threshold of p < 0.05, i.e., falsely concluding a statistically significant difference between two identical approaches. We prove that this evaluation setup is unsuitable to compare learning approaches. We formalize alternative evaluation setups based on score distributions.

For agents

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

get_harvested_code_for_paper("1803.09578")
get_code_for_paper("1803.09578")
have("1803.09578")

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

curl https://syntology.ai/api/ran/1803.09578.json

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

[![Syntology run record](https://syntology.ai/api/ran/1803.09578.svg)](https://syntology.ai/paper/1803.09578)

Connect an agent — have() is free.