Santiago Martínez Novoa

Santiago Martínez Novoa

Machine Learning Researcher | Multilingual NLP & Trustworthy LLM Evaluation

CV Resume

MeSHClass-ES and AnatEM-ES: Open Resources for Spanish Biomedical NLP

Published in BioNLP 2026 (ACL Workshop), San Diego, California, pp. 617–629, 2026

Abstract

Despite Spanish being one of the most widely spoken languages in the world, biomedical NLP resources and systematic evaluations remain limited relative to English. We address this gap by constructing and releasing two Spanish biomedical corpora: (1) MeSHClass-ES, a 29,063 abstract bilingual corpus translated from PubMed with Opus-MT, and (2) AnatEM-ES, the AnatEM anatomical entity corpus translated with a chunk-level LLM-based pipeline that jointly preserves BIO annotations across 13,849 entity mentions. Both corpora achieve a mean COMET score of 0.73 despite using different translation systems. We benchmark nine encoder models spanning general-domain Spanish, domain-specific, and multilingual architectures for both tasks. RigoBERTa-2.0 leads both tasks (micro-F1 classification 0.69, tied with SciBETO-large; NER F1 0.66). Both domain pretraining and model capacity drive performance, with the gap slightly more pronounced for NER (4-point spread) than classification (3-point spread). XLM-RoBERTa-large emerges as a competitive multilingual baseline. A parallel evaluation of four open-weight decoders (7-9B) reveals a task-dependent encoder-decoder gap: QLoRA-adapted Gemma-2-9B reaches 88% of the best encoder on classification (micro-F1 .61 vs .69), but for NER every decoder configuration we tested stays at or below 40% of the best encoder F1. We release both corpora on the HuggingFace Hub, translation pipelines, and evaluation code on GitHub.

Slides

Slide 1
1 / 28
Slide 2
2 / 28
Slide 3
3 / 28
Slide 4
4 / 28
Slide 5
5 / 28
Slide 6
6 / 28
Slide 7
7 / 28
Slide 8
8 / 28
Slide 9
9 / 28
Slide 10
10 / 28
Slide 11
11 / 28
Slide 12
12 / 28
Slide 13
13 / 28
Slide 14
14 / 28
Slide 15
15 / 28
Slide 16
16 / 28
Slide 17
17 / 28
Slide 18
18 / 28
Slide 19
19 / 28
Slide 20
20 / 28
Slide 21
21 / 28
Slide 22
22 / 28
Slide 23
23 / 28
Slide 24
24 / 28
Slide 25
25 / 28
Slide 26
26 / 28
Slide 27
27 / 28
Slide 28
28 / 28

Scroll sideways to browse. Open as PDF.

Details

Author: Santiago Martínez Novoa, Lina Gómez Mesa, Juan Prieto, Rubén Manrique Year: 2026 Publication Type: Paper presented at BioNLP 2026 (Biomedical Natural Language Processing Workshop, ACL), San Diego, California, pp. 617-629
Presented: Oral presentation (virtual), July 4, 2026

Citation

Martínez Novoa, S., Gómez Mesa, L., Prieto, J., & Manrique, R. (2026). MeSHClass-ES and AnatEM-ES: Open Resources for Spanish Biomedical NLP. In BioNLP 2026, pp. 617-629. Association for Computational Linguistics. https://doi.org/10.18653/v1/2026.bionlp-1.49

See Paper Slides Download BibTeX