metadata

language:
  - es
tags:
  - biomedical
  - spanish
license: apache-2.0
metrics:
  - ppl
widget:
  - text: El único antecedente personal a reseñar era la <mask> arterial.
  - text: >-
      Las radiologías óseas de cuerpo entero no detectan alteraciones <mask>, ni
      alteraciones vertebrales.
  - text: >-
      En el <mask> toraco-abdómino-pélvico no se encontraron hallazgos
      patológicos de interés.

Biomedical language model for Spanish

BibTeX citation

If you use any of these resources (datasets or models) in your work, please cite our latest paper:

@misc{carrino2021biomedical,
      title={Biomedical and Clinical Language Models for Spanish: On the Benefits of Domain-Specific Pretraining in a Mid-Resource Scenario}, 
      author={Casimiro Pio Carrino and Jordi Armengol-Estapé and Asier Gutiérrez-Fandiño and Joan Llop-Palao and Marc Pàmies and Aitor Gonzalez-Agirre and Marta Villegas},
      year={2021},
      eprint={2109.03570},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

Model and tokenization

This model is a RoBERTa-based model trained on a biomedical corpus collected from several sources (see next section).

Training corpora and preprocessing

The training corpus is composed of several biomedical corpora in Spanish, collected from publicly available corpora and crawlers:

Name	No. tokens	Description
Medical crawler	745,705,946	Crawler of more than 3,000 URLs belonging to Spanish health domain
Scielo	60,007,289	Collection of biomedical literature in Spanish crawled from the Scielo repository in 2019
BARR2_background	24,516,442	Biomedical Abbreviation Recognition and Resolution (BARR2) containing Spanish clinical case study sections from a variety of clinical disciplines
Wikipedia_life_sciences	13,890,501	Wikipedia articles beloging to the Life Sciences category crawled on 04/01/2021
Patents	13,463,387	Google Patent in Medical Domain for Spain (Spanish). The accepted codes (Medical Domain) for Json files of patents are: "A61B", "A61C","A61F", "A61H", "A61K", "A61L","A61M", "A61B", "A61P"
EMEA	5,377,448	Spanish-side documents extracted from the a parallel corpus made out of PDF documents from the European Medicines Agency.
mespen_Medline	4,166,077	Spanish-side documents extracted from a collection of Spanish-English parallel corpora consistiing of biomedical scientific literature. The collection of parallel resources are aggregated from the IBECS, SciELO, Pubmed and MedlinePlus sources.
PubMed	1,858,966	Collection of biomedical literature in Spanish crawled from the PubMed repository in 2019

To obtain a high-quality training corpus, a cleaning pipeline with the following operations has been applied:

data parsing in different formats
- sentence splitting
- language detection
- filtering of ill-formed sentences
- deduplication of repetitive contents
- keep the original document boundaries

Finally, the corpora are concatenated and further global deduplication among the corpora have been applied. The result is a medium-size biomedical corpus for Spanish composed of about 860M tokens.

Evaluation and results

The model has been evaluated on the Named Entity Recognition (NER) using the following datasets:

PharmaCoNER: is a track on chemical and drug mention recognition from Spanish medical texts (for more info see: https://temu.bsc.es/pharmaconer/).
CANTEMIST: is a shared task specifically focusing on named entity recognition of tumor morphology, in Spanish (for more info see: https://zenodo.org/record/3978041#.YTt5qH2xXbQ).
ICTUSnet: consists of 1,006 hospital discharge reports of patients admitted for stroke from 18 different Spanish hospitals. It contains more than 79,000 annotations for 51 different kinds of variables.

The evaluation results are compared against the mBERT and BETO models:

F1 - Precision - Recall	roberta-base-biomedical-es	mBERT	BETO
PharmaCoNER	89.48 - 87.85 - 91.18	87.46 - 86.50 - 88.46	88.18 - 87.12 - 89.28
CANTEMIST	83.87 - 81.70 - 86.17	82.61 - 81.12 - 84.15	82.42 - 80.91 - 84.00
ICTUSnet	88.12 - 85.56 - 90.83	86.75 - 83.53 - 90.23	85.95 - 83.10 - 89.02

Intended uses & limitations

The model is ready-to-use only for masked language modelling to perform the Fill Mask task (try the inference API or read the next section)

However, the is intended to be fine-tuned on non-generative downstream tasks such as Question Answering, Text Classification or Named Entity Recognition.

How to use

from transformers import AutoTokenizer, AutoModelForMaskedLM

tokenizer = AutoTokenizer.from_pretrained("BSC-TeMU/roberta-base-biomedical-es")

model = AutoModelForMaskedLM.from_pretrained("BSC-TeMU/roberta-base-biomedical-es")

from transformers import pipeline

unmasker = pipeline('fill-mask', model="BSC-TeMU/roberta-base-biomedical-es")

unmasker("El único antecedente personal a reseñar era la <mask> arterial.")

# Output
[
  {
    "sequence": " El único antecedente personal a reseñar era la hipertensión arterial.",
    "score": 0.9855039715766907,
    "token": 3529,
    "token_str": " hipertensión"
  },
  {
    "sequence": " El único antecedente personal a reseñar era la diabetes arterial.",
    "score": 0.0039140828885138035,
    "token": 1945,
    "token_str": " diabetes"
  },
  {
    "sequence": " El único antecedente personal a reseñar era la hipotensión arterial.",
    "score": 0.002484665485098958,
    "token": 11483,
    "token_str": " hipotensión"
  },
  {
    "sequence": " El único antecedente personal a reseñar era la Hipertensión arterial.",
    "score": 0.0023484621196985245,
    "token": 12238,
    "token_str": " Hipertensión"
  },
  {
    "sequence": " El único antecedente personal a reseñar era la presión arterial.",
    "score": 0.0008009297889657319,
    "token": 2267,
    "token_str": " presión"
  }
]