UMLSmapper

UMLSMapper performs semantic annotation of biomedical literature to enable cross-lingual biomedical knowledge mining from under-resourced languages such as Spanish.


Key Features:

  • Cross-Lingual Annotation: Supports semantic annotation of biomedical texts across languages with emphasis on Spanish-language documents.
  • Linguistic Analysis: Employs linguistic analysis to annotate Spanish texts directly using information retrieval techniques and concept disambiguation.
  • Machine Translation-Based Annotation: Utilizes Spanish-English machine translation to annotate English documents and transfer annotations back to the original Spanish text.
  • Combined Approach: Integrates linguistic analysis and machine translation-based annotation into a hybrid system that demonstrates improved performance over individual methods.
  • Information Retrieval: Applies information retrieval techniques as part of the annotation process.
  • Concept Disambiguation: Uses concept disambiguation to resolve ambiguous biomedical concepts during annotation.
  • Annotation Transfer: Transfers annotations produced on translated English documents back to the source Spanish texts.

Scientific Applications:

  • Biomedical Knowledge Mining: Enables mining of biomedical knowledge from Spanish-language literature through standardized semantic annotations.
  • Access to Non-English Biomedical Knowledge: Facilitates comprehensive access to biomedical knowledge contained in under-resourced Spanish texts for research.
  • Multinational Research Support: Supports multinational studies by reducing language barriers in biomedical text annotation and information extraction.

Methodology:

Annotates Spanish texts via linguistic analysis that uses information retrieval techniques and concept disambiguation; employs Spanish-English machine translation to annotate English documents and transfer annotations back to Spanish; evaluates and compares linguistic analysis, machine translation-based annotation, and their combination, demonstrating that the combined system enhances performance.

Topics

Details

Added:
1/14/2020
Last Updated:
1/16/2021

Operations

Publications

Perez N, Accuosto P, Bravo À, Cuadros M, Martínez-Garcia E, Saggion H, Rigau G. Cross-lingual semantic annotation of biomedical literature: experiments in Spanish and English. Bioinformatics. 2019;36(6):1872-1880. doi:10.1093/bioinformatics/btz853. PMID:31730202.

PMID: 31730202
Funding: - Maria de Maeztu Units of Excellence Programme: MDM-2015-0502 - CROSSTEXT: MINECO/FEDER, TIN2015-72646-EXP - DeepReading: RTI2018-096846-B-C21 MCIU/AEI/FEDER