Dataset:
assin

Languages: pt
Multilinguality: monolingual
Size Categories: 10K<n<100K
Licenses: unknown
Language Creators: found
Annotations Creators: expert-generated
Source Datasets: original

Dataset Card for ASSIN

Dataset Summary

The ASSIN (Avaliação de Similaridade Semântica e INferência textual) corpus is a corpus annotated with pairs of sentences written in Portuguese that is suitable for the exploration of textual entailment and paraphrasing classifiers. The corpus contains pairs of sentences extracted from news articles written in European Portuguese (EP) and Brazilian Portuguese (BP), obtained from Google News Portugal and Brazil, respectively. To create the corpus, the authors started by collecting a set of news articles describing the same event (one news article from Google News Portugal and another from Google News Brazil) from Google News. Then, they employed Latent Dirichlet Allocation (LDA) models to retrieve pairs of similar sentences between sets of news articles that were grouped together around the same topic. For that, two LDA models were trained (for EP and for BP) on external and large-scale collections of unannotated news articles from Portuguese and Brazilian news providers, respectively. Then, the authors defined a lower and upper threshold for the sentence similarity score of the retrieved pairs of sentences, taking into account that high similarity scores correspond to sentences that contain almost the same content (paraphrase candidates), and low similarity scores correspond to sentences that are very different in content from each other (no-relation candidates). From the collection of pairs of sentences obtained at this stage, the authors performed some manual grammatical corrections and discarded some of the pairs wrongly retrieved. Furthermore, from a preliminary analysis made to the retrieved sentence pairs the authors noticed that the number of contradictions retrieved during the previous stage was very low. Additionally, they also noticed that event though paraphrases are not very frequent, they occur with some frequency in news articles. Consequently, in contrast with the majority of the currently available corpora for other languages, which consider as labels “neutral”, “entailment” and “contradiction” for the task of RTE, the authors of the ASSIN corpus decided to use as labels “none”, “entailment” and “paraphrase”. Finally, the manual annotation of pairs of sentences was performed by human annotators. At least four annotators were randomly selected to annotate each pair of sentences, which is done in two steps: (i) assigning a semantic similarity label (a score between 1 and 5, from unrelated to very similar); and (ii) providing an entailment label (one sentence entails the other, sentences are paraphrases, or no relation). Sentence pairs where at least three annotators do not agree on the entailment label were considered controversial and thus discarded from the gold standard annotations. The full dataset has 10,000 sentence pairs, half of which in Brazilian Portuguese (ptbr) and half in European Portuguese (ptpt). Either language variant has 2,500 pairs for training, 500 for validation and 2,000 for testing.

Supported Tasks and Leaderboards

[More Information Needed]

Languages

The language supported is Portuguese.

Dataset Structure

Data Instances

An example from the ASSIN dataset looks as follows:

{
  "entailment_judgment": 0,
  "hypothesis": "André Gomes entra em campo quatro meses depois de uma lesão na perna esquerda o ter afastado dos relvados.",
  "premise": "Relembre-se que o atleta estava afastado dos relvados desde maio, altura em que contraiu uma lesão na perna esquerda.",
  "relatedness_score": 3.5,
  "sentence_pair_id": 1
}

Data Fields

  • sentence_pair_id: a int64 feature.
  • premise: a string feature.
  • hypothesis: a string feature.
  • relatedness_score: a float32 feature.
  • entailment_judgment: a classification label, with possible values including NONE, ENTAILMENT, PARAPHRASE.

Data Splits

The data is split into train, validation and test set. The split sizes are as follow:

Train Val Test
full 5000 1000 4000
ptbr 2500 500 2000
ptpt 2500 500 2000

Dataset Creation

Curation Rationale

[More Information Needed]

Source Data

Initial Data Collection and Normalization

[More Information Needed]

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

[More Information Needed]

Personal and Sensitive Information

[More Information Needed]

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

[More Information Needed]

Licensing Information

[More Information Needed]

Citation Information

@inproceedings{fonseca2016assin,
  title={ASSIN: Avaliacao de similaridade semantica e inferencia textual},
  author={Fonseca, E and Santos, L and Criscuolo, Marcelo and Aluisio, S},
  booktitle={Computational Processing of the Portuguese Language-12th International Conference, Tomar, Portugal},
  pages={13--15},
  year={2016}
}

Contributions

Thanks to @jonatasgrosman for adding this dataset.

Models trained or fine-tuned on assin

None yet