BELDA: Multilingual Biomedical Entity Linking Descriptions

BELDA provides precomputed mention and entity descriptions for biomedical entity linking. The dataset includes descriptions for the XL-BEL corpus across 10 languages and the NCBI corpus in English. It was authored by wenliang liang and last updated on June 1, 2026.

Use Cases

• Train biomedical entity linking models based on precomputed mention descriptions.
• Benchmark multilingual NLP systems based on the 10-language XL-BEL corpus.
• Fine-tune language models for biomedical text based on entity descriptions from the NCBI corpus.

Strengths

• Dataset size is 147.2 MB, indicating substantial textual content.
• Covers 10 languages for the XL-BEL component, supporting multilingual research.
• Released under a permissive CC-BY-4.0 license for broad reuse.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale training.

Provenance

Source

figshare

Collection Method

Precomputed from source corpora for the BELDA paper.

Freshness

Last updated 2026-06-01 05:38:09