ELECTRAMed

Reported on 4 benchmarks across 3 tasks · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing3 results

Named Entity Recognition (NER)onBC5CDR
F1· uses extra data· 2021-04-19
90.03
best: 91.9 (BINDER)
SOTA
ELECTRAMed: a new pre-trained language representation model for biomedical NLP arXiv:2104.09585
Relation ExtractiononChemProt
F1· uses extra data· 2021-04-19
72.94
best: 83.64 (SciBert (Finetune))
ELECTRAMed: a new pre-trained language representation model for biomedical NLP arXiv:2104.09585
Named Entity Recognition (NER)onNCBI-disease
F1· uses extra data· 2021-04-19
87.54
best: 89.71 (BioBERT)
ELECTRAMed: a new pre-trained language representation model for biomedical NLP arXiv:2104.09585

Medical1 result

Information ExtractiononDDI extraction 2013 corpus
Micro F1· uses extra data· 2021-04-19
79.13
best: 84.08 (DESC+MOL+SciBERT)
ELECTRAMed: a new pre-trained language representation model for biomedical NLP arXiv:2104.09585