mbruton/spanish_srl

Dataset Card for SpanishSRL

Dataset Summary

The SpanishSRL dataset is a Spanish-language dataset of tokenized sentences and the semantic role for each token within a sentence. Standard semantic roles for Spanish are identified as well as verbal root; standard roles include "arg0|[agt, cau, exp, src]", "arg1|[ext, loc, pat, tem]", "arg2[atr, ben, efi, exp, ext, ins, loc]", "arg3[ben, ein, fin, ori]", "arg4[des, efi]", and "argM[adv, atr, cau, ext, fin, ins, loc, mnr, tmp]".

Languages

The text in the dataset is in Spanish.

Dataset Structure

Data Instances

A typical data point comprises a tokenized sentence, tags for each token, and a sentence id number. An example from the SpanishSRL dataset looks as follows:

{'tokens': ['Ante', 'unas', 'mil', 'personas', ',', 'entre', 'ellas', 'la', 'ministra', 'de', 'Ciencia_y_Tecnología', ',', 'Anna_Birulés', ',', 'el', 'alcalde', 'de', 'Barcelona', ',', 'Joan_Clos', ',', 'la', 'Delegada', 'del', 'Gobierno', ',', 'Julia_García_Valdecasas', ',', 'y', 'una', 'nutrida', 'representación', 'del', 'gobierno', 'catalán', ',', 'Pujol', 'dio', 'un', 'toque', 'de', 'alerta', 'sobre', 'el', 'aumento', 'de', 'los', 'accidentes', 'laborales', '.'],
 'tags': [34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 37, 0, 0, 0, 0, 28, 0, 0, 0, 0, 0, 0, 0],
 'ids': 66}

Tags are assigned an id number according to the index of its label as listed in:

>>> dataset['train'].features['tags'].feature.names

Data Fields

• tokens: a list of strings
• tags: a list of integers
• ids: a sentence id, as an integer

Data Splits

The data is split into a development, training, and test set. The final structure and split sizes are as follow:

DatasetDict({
    dev: Dataset({
        features: ['tokens', 'tags', 'ids'],
        num_rows: 1654
    })
    test: Dataset({
        features: ['tokens', 'tags', 'ids'],
        num_rows: 1724
    })
    train: Dataset({
        features: ['tokens', 'tags', 'ids'],
        num_rows: 14328
    })
})

Dataset Creation

Curation Rationale

SpanishSRL was built to test the verbal indexing method as introduced in the publication listed in the citation against an established baseline.

Source Data

Initial Data Collection and Normalization

Data was collected from the 2009 CoNLL Shared Task. For more information, please refer to the publication listed in the citation.

Additional Information

Dataset Curators

The dataset was created by Micaella Bruton, as part of her Master's thesis.

Citation Information

@inproceedings{bruton-beloucif-2023-bertie,
    title = "{BERT}ie Bott{'}s Every Flavor Labels: A Tasty Introduction to Semantic Role Labeling for {G}alician",
    author = "Bruton, Micaella  and
      Beloucif, Meriem",
    editor = "Bouamor, Houda  and
      Pino, Juan  and
      Bali, Kalika",
    booktitle = "Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing",
    month = dec,
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.emnlp-main.671",
    doi = "10.18653/v1/2023.emnlp-main.671",
    pages = "10892--10902",
    abstract = "In this paper, we leverage existing corpora, WordNet, and dependency parsing to build the first Galician dataset for training semantic role labeling systems in an effort to expand available NLP resources. Additionally, we introduce verb indexing, a new pre-processing method, which helps increase the performance when semantically parsing highly-complex sentences. We use transfer-learning to test both the resource and the verb indexing method. Our results show that the effects of verb indexing were amplified in scenarios where the model was both pre-trained and fine-tuned on datasets utilizing the method, but improvements are also noticeable when only used during fine-tuning. The best-performing Galician SRL model achieved an f1 score of 0.74, introducing a baseline for future Galician SRL systems. We also tested our method on Spanish where we achieved an f1 score of 0.83, outperforming the baseline set by the 2009 CoNLL Shared Task by 0.025 showing the merits of our verb indexing method for pre-processing.",
}