metadata

language:
  - en
pretty_name: CANNOT

Dataset Card for CANNOT

Dataset Description

Homepage: https://github.com/dmlls/cannot-dataset
Repository: https://github.com/dmlls/cannot-dataset
Paper: tba

Dataset Summary

CANNOT is a dataset that focuses on negated textual pairs. It currently contains 77,376 samples, of which roughly of them are negated pairs of sentences, and the other half are not (they are paraphrased versions of each other).

The most frequent negation that appears in the dataset is verbal negation (e.g., will → won't), although it also contains pairs with antonyms (cold → hot).

Languages

CANNOT includes exclusively texts in English.

Dataset Structure

The dataset is given as a .tsv file with the following structure:

premise	hypothesis	label
A sentence.	An equivalent, non-negated sentence (paraphrased).	0
A sentence.	The sentence negated.	1

The dataset can be easily loaded into a Pandas DataFrame by running:

import pandas as pd

dataset = pd.read_csv('negation_dataset_v1.0.tsv', sep='\t')

Dataset Creation

The dataset has been created by cleaning up and merging the following datasets:

Not another Negation Benchmark: The NaN-NLI Test Suite for Sub-clausal Negation (see datasets/nan-nli).
GLUE Diagnostic Dataset (see datasets/glue-diagnostic).
Automated Fact-Checking of Claims from Wikipedia (see datasets/wikifactcheck-english).
From Group to Individual Labels Using Deep Features (see datasets/sentiment-labelled-sentences). In this case, the negated sentences were obtained by using the Python module negate.

Additionally, for each of the negated samples, another pair of non-negated sentences has been added by paraphrasing them with the pre-trained model 🤗tuner007/pegasus_paraphrase.

Furthermore, the dataset from It Is Not Easy To Detect Paraphrases: Analysing Semantic Similarity With Antonyms and Negation Using the New SemAntoNeg Benchmark (see datasets/antonym-substitution) has also been included. This dataset already provides both the paraphrased and negated version for each premise, so no further processing was needed.

Finally, the swapped version of each pair (premise ⇋ hypothesis) has also been included, and any duplicates have been removed.

The contribution of each of these individual datasets to the final CANNOT dataset is:

Dataset	Samples
Not another Negation Benchmark	118
GLUE Diagnostic Dataset	154
Automated Fact-Checking of Claims from Wikipedia	14,970
From Group to Individual Labels Using Deep Features	2,110
It Is Not Easy To Detect Paraphrases	8,597
Total	25,949

Note: The numbers above include only the original queries present in the datasets.

Additional Information

Licensing Information

TODO

Citation Information

tba

Contributions

Contributions to the dataset can be submitted through the project repository.