Papers › Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing

Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing

31 Jul 2020arXiv:2007.15779archive 2025-07-28

Yu Gu, Robert Tinn, Hao Cheng, Michael Lucas, Naoto Usuyama, Xiaodong Liu, Tristan Naumann, Jianfeng Gao, Hoifung Poon

Pretraining large neural language models, such as BERT, has led to impressive gains on many natural language processing (NLP) tasks. However, most pretraining efforts focus on general domain corpora, such as newswire and Web. A prevailing assumption is that even domain-specific pretraining can benefit by starting from general-domain language models. In this paper, we challenge this assumption by showing that for domains with abundant unlabeled text, such as biomedicine, pretraining language models from scratch results in substantial gains over continual pretraining of general-domain language models. To facilitate this investigation, we compile a comprehensive biomedical NLP benchmark from publicly-available datasets. Our experiments show that domain-specific pretraining serves as a solid foundation for a wide range of biomedical NLP tasks, leading to new state-of-the-art results across the board. Further, in conducting a thorough evaluation of modeling choices, both for pretraining and task-specific fine-tuning, we discover that some common practices are unnecessary with BERT models, such as using complex tagging schemes in named entity recognition (NER). To help accelerate research in biomedical NLP, we have released our state-of-the-art pretrained and task-specific models for the community, and created a leaderboard featuring our BLURB benchmark (short for Biomedical Language Understanding & Reasoning Benchmark) at https://aka.ms/BLURB.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

rohanshad/cmr_transformer mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Continual PretrainingDocument ClassificationDrug–drug Interaction ExtractionLanguage ModelingLanguage ModellingNERNamed Entity RecognitionNamed Entity Recognition (NER)PICOParticipant Intervention Comparison Outcome ExtractionQuestion AnsweringRelation ExtractionSentence SimilarityText Classificationnamed-entity-recognition

Datasets

Introduced by this paper, per the archive.

BLURB

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Document Classification HOC PubMedBERT uncased Micro F1 82.32 #5 of 5 Archive leaderboard report
Drug–drug Interaction Extraction DDI extraction 2013 corpus PubMedBERT F1 0.8236 #3 of 10 Archive leaderboard report
Drug–drug Interaction Extraction DDI extraction 2013 corpus PubMedBERT Micro F1 82.36 #3 of 10 Archive leaderboard report
Named Entity Recognition (NER) BC2GM PubMedBERT uncased F1 84.52 #10 of 13 Archive leaderboard report
Named Entity Recognition (NER) JNLPBA PubMedBERT uncased F1 79.1 #9 of 17 Archive leaderboard report
Named Entity Recognition (NER) NCBI-disease PubMedBERT uncased F1 87.82 #16 of 26 Archive leaderboard report
PICO EBM PICO PubMedBERT uncased Macro F1 word level 73.38 #3 of 3 Archive leaderboard report
Participant Intervention Comparison Outcome Extraction EBM-NLP PubMedBERT uncased F1 73.38 #2 of 5 Archive leaderboard report
Question Answering BLURB PubMedBERT (uncased; abstracts) Accuracy 71.7 #4 of 4 Archive leaderboard report
Question Answering BioASQ PubMedBERT uncased Accuracy 87.56 #5 of 7 Archive leaderboard report
Question Answering PubMedQA PubMedBERT uncased Accuracy 55.84 #28 of 30 Archive leaderboard report
Relation Extraction ChemProt PubMedBERT uncased Micro F1 77.24 #12 of 13 Archive leaderboard report
Relation Extraction DDI PubMedBERT uncased Micro F1 82.36 #3 of 3 Archive leaderboard report
Relation Extraction GAD PubMedBERT uncased Micro F1 82.34 #3 of 3 Archive leaderboard report
Text Classification BLURB PubMedBERT (uncased; abstracts) F1 82.32 #3 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AdamAttentionAttention DropoutBERTDense ConnectionsDropoutLayer NormalizationLinear LayerLinear Warmup With Linear DecayMulti-Head AttentionResidual ConnectionSoftmaxWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections