mbruton/galician_srl

Dataset Card for GalicianSRL

Dataset Summary

The GalicianSRL dataset is a Galician-language dataset of tokenized sentences and the semantic role for each token within a sentence. Semantic roles are limited to verbal roots, argument 0, argument 1, and argument 2. This dataset was created to support the task of semantic role labeling in the Galician language, as no publically available datasets existed as of the date of publication to the contributor's knowledge.

Languages

The text in the dataset is in Galician.

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 GalicianSRL dataset looks as follows:

{'tokens': ['O', 'Pleno', 'poderá', ',', 'con', 'todo', ',', 'avocar', 'en', 'calquera', 'momento', 'o', 'debate', 'e', 'votación', 'de', 'calquera', 'proxecto', 'ou', 'proposición', 'de', 'lei', 'que', 'xa', 'fora', 'obxecto', 'de', 'esta', 'delegación', '.'],
 'tags': [0, 1, 4, 0, 0, 0, 0, 17, 0, 0, 16, 0, 15, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 'ids': 504}

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 training and test set. The final structure and split sizes are as follow:

DatasetDict({
    train: Dataset({
        features: ['tokens', 'tags', 'ids'],
        num_rows: 1005
    })
    test: Dataset({
        features: ['tokens', 'tags', 'ids'],
        num_rows: 252
    })
})

Dataset Creation

Curation Rationale

GalicianSRL was built to provide a dataset for semantic role labeling in Galician and expand NLP resources available for the Galician language.

Source Data

Initial Data Collection and Normalization

Data was collected from both the CTG UD annotated corpus and the TreeGal UD annotated corpus, and combined to collect the requsite information for this task. For more information, please refer to the publication listed in the citation.

Considerations for Using the Data

Limitations

The purpose of this dataset is to help develop a working semantic role labeling system for Galician, as SRL systems have been shown to improve a variety of NLP tasks. It should be noted however that Galician is considered a low-resource language at this time, and as such the dataset has an extrememly limited scope. This dataset would benefit from manual validation of a native speaker of Galician, the inclusion of additional sentences, and an extention of arguments past arg0, arg1, and arg2.

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.",
}