{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/scibert-pretrained-contextualized-embeddings","title":"SciBERT: A Pretrained Language Model for Scientific Text","arxiv_id":"1903.10676","date":"2019-03-26","proceeding":"IJCNLP 2019 11","authors":["Iz Beltagy","Kyle Lo","Arman Cohan"],"abstract":"Obtaining large-scale annotated data for NLP tasks in the scientific domain is challenging and expensive. We release SciBERT, a pretrained language model based on BERT (Devlin et al., 2018) to address the lack of high-quality, large-scale labeled scientific data. SciBERT leverages unsupervised pretraining on a large multi-domain corpus of scientific publications to improve performance on downstream scientific NLP tasks. We evaluate on a suite of tasks including sequence tagging, sentence classification and dependency parsing, with datasets from a variety of scientific domains. We demonstrate statistically significant improvements over BERT and achieve new state-of-the-art results on several of these tasks. The code and pretrained models are available at https://github.com/allenai/scibert/.","url_abs":"https://arxiv.org/abs/1903.10676v3","url_pdf":"https://arxiv.org/pdf/1903.10676v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"scibert-pretrained-contextualized-embeddings","repo_url":"https://github.com/allenai/scibert","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"scibert-pretrained-contextualized-embeddings","repo_url":"https://github.com/charles9n/bert-sklearn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"scibert-pretrained-contextualized-embeddings","repo_url":"https://github.com/georgetown-cset/ai-relevant-papers","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"scibert-pretrained-contextualized-embeddings","repo_url":"https://github.com/hoangcuongnguyen2001/scibert-for-technique-classification","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"scibert-pretrained-contextualized-embeddings","repo_url":"https://github.com/kuldeep7688/BioMedicalBertNer","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"scibert-pretrained-contextualized-embeddings","repo_url":"https://github.com/tetsu9923/scireviewgen","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"citation-intent-classification","task_name":"Citation Intent Classification"},{"task_slug":"dependency-parsing","task_name":"Dependency Parsing"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"medical-named-entity-recognition","task_name":"Medical Named Entity Recognition"},{"task_slug":"named-entity-recognition-ner","task_name":"Named Entity Recognition (NER)"},{"task_slug":"participant-intervention-comparison-outcome","task_name":"Participant Intervention Comparison Outcome Extraction"},{"task_slug":"relation-extraction","task_name":"Relation Extraction"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"sentence-classification","task_name":"Sentence Classification"},{"task_slug":"model","task_name":"model"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bert","method_name":"BERT"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-linear-decay","method_name":"Linear Warmup With Linear Decay"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"},{"method_slug":"wordpiece","method_name":"WordPiece"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/citation-intent-classification-on-scicite","task":"Citation Intent Classification","dataset":"SciCite","model":"SciBERT","rank_in_archive_order":6,"of":8,"metrics":{"Macro-F1":"86.32"},"uses_additional_data":true},{"leaderboard":"/sota/dependency-parsing-on-genia-las","task":"Dependency Parsing","dataset":"GENIA - LAS","model":"SciBERT (SciVocab)","rank_in_archive_order":2,"of":3,"metrics":{"F1":"91.41"},"uses_additional_data":true},{"leaderboard":"/sota/dependency-parsing-on-genia-las","task":"Dependency Parsing","dataset":"GENIA - LAS","model":"SciBERT (Base Vocab)","rank_in_archive_order":3,"of":3,"metrics":{"F1":"91.26"},"uses_additional_data":true},{"leaderboard":"/sota/dependency-parsing-on-genia-uas","task":"Dependency Parsing","dataset":"GENIA - UAS","model":"SciBERT (SciVocab)","rank_in_archive_order":2,"of":3,"metrics":{"F1":"92.46"},"uses_additional_data":true},{"leaderboard":"/sota/dependency-parsing-on-genia-uas","task":"Dependency Parsing","dataset":"GENIA - UAS","model":"SciBERT (Base Vocab)","rank_in_archive_order":3,"of":3,"metrics":{"F1":"92.32"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-bc5cdr","task":"Named Entity Recognition (NER)","dataset":"BC5CDR","model":"SciBERT (SciVocab)","rank_in_archive_order":11,"of":16,"metrics":{"F1":"88.94"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-bc5cdr","task":"Named Entity Recognition (NER)","dataset":"BC5CDR","model":"SciBERT (Base Vocab)","rank_in_archive_order":13,"of":16,"metrics":{"F1":"88.11"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-jnlpba","task":"Named Entity Recognition (NER)","dataset":"JNLPBA","model":"SciBERT (Base Vocab)","rank_in_archive_order":17,"of":17,"metrics":{"F1":"75.77"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-ncbi-disease","task":"Named Entity Recognition (NER)","dataset":"NCBI-disease","model":"SciBERT (Base Vocab)","rank_in_archive_order":21,"of":26,"metrics":{"F1":"86.88"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-ncbi-disease","task":"Named Entity Recognition (NER)","dataset":"NCBI-disease","model":"SciBERT (SciVocab)","rank_in_archive_order":23,"of":26,"metrics":{"F1":"86.45"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-scierc","task":"Named Entity Recognition (NER)","dataset":"SciERC","model":"SciBERT (SciVocab)","rank_in_archive_order":5,"of":7,"metrics":{"F1":"67.57"},"uses_additional_data":true},{"leaderboard":"/sota/named-entity-recognition-ner-on-scierc","task":"Named Entity Recognition (NER)","dataset":"SciERC","model":"SciBERT (Base Vocab)","rank_in_archive_order":6,"of":7,"metrics":{"F1":"65.24"},"uses_additional_data":true},{"leaderboard":"/sota/participant-intervention-comparison-outcome","task":"Participant Intervention Comparison Outcome Extraction","dataset":"EBM-NLP","model":"SciBERT (SciVocab)","rank_in_archive_order":3,"of":5,"metrics":{"F1":"71.18"},"uses_additional_data":true},{"leaderboard":"/sota/participant-intervention-comparison-outcome","task":"Participant Intervention Comparison Outcome Extraction","dataset":"EBM-NLP","model":"SciBERT (Base Vocab)","rank_in_archive_order":4,"of":5,"metrics":{"F1":"70.82"},"uses_additional_data":true},{"leaderboard":"/sota/relation-extraction-on-chemprot","task":"Relation Extraction","dataset":"ChemProt","model":"SciBert (Finetune)","rank_in_archive_order":1,"of":13,"metrics":{"F1":"83.64"},"uses_additional_data":false},{"leaderboard":"/sota/relation-extraction-on-chemprot","task":"Relation Extraction","dataset":"ChemProt","model":"SciBERT (Base Vocab)","rank_in_archive_order":10,"of":13,"metrics":{"F1":"73.7"},"uses_additional_data":true},{"leaderboard":"/sota/relation-extraction-on-jnlpba","task":"Relation Extraction","dataset":"JNLPBA","model":"SciBERT (SciVocab)","rank_in_archive_order":1,"of":1,"metrics":{"F1":"76.09"},"uses_additional_data":false},{"leaderboard":"/sota/relation-extraction-on-scierc","task":"Relation Extraction","dataset":"SciERC","model":"SciBERT (SciVocab)","rank_in_archive_order":1,"of":3,"metrics":{"F1":"74.64"},"uses_additional_data":true},{"leaderboard":"/sota/relation-extraction-on-scierc","task":"Relation Extraction","dataset":"SciERC","model":"SciBERT (Base Vocab)","rank_in_archive_order":2,"of":3,"metrics":{"F1":"74.42"},"uses_additional_data":true},{"leaderboard":"/sota/sentence-classification-on-acl-arc","task":"Sentence Classification","dataset":"ACL-ARC","model":"SciBERT","rank_in_archive_order":2,"of":4,"metrics":{"F1":"70.98"},"uses_additional_data":false},{"leaderboard":"/sota/sentence-classification-on-paper-field","task":"Sentence Classification","dataset":"Paper Field","model":"SciBERT (SciVocab)","rank_in_archive_order":1,"of":2,"metrics":{"F1":"65.71"},"uses_additional_data":true},{"leaderboard":"/sota/sentence-classification-on-paper-field","task":"Sentence Classification","dataset":"Paper Field","model":"SciBERT (Base Vocab)","rank_in_archive_order":2,"of":2,"metrics":{"F1":"64.02"},"uses_additional_data":true},{"leaderboard":"/sota/sentence-classification-on-pubmed-20k-rct","task":"Sentence Classification","dataset":"PubMed 20k RCT","model":"SciBERT (Base Vocab)","rank_in_archive_order":2,"of":2,"metrics":{"F1":"86.81"},"uses_additional_data":true},{"leaderboard":"/sota/sentence-classification-on-scicite","task":"Sentence Classification","dataset":"SciCite","model":"SciBERT","rank_in_archive_order":1,"of":5,"metrics":{"F1":"84.9"},"uses_additional_data":false},{"leaderboard":"/sota/sentence-classification-on-sciencecite","task":"Sentence Classification","dataset":"ScienceCite","model":"SciBERT (SciVocab)","rank_in_archive_order":1,"of":2,"metrics":{"F1":"84.99"},"uses_additional_data":true},{"leaderboard":"/sota/sentence-classification-on-sciencecite","task":"Sentence Classification","dataset":"ScienceCite","model":"SciBERT (Base Vocab)","rank_in_archive_order":2,"of":2,"metrics":{"F1":"84.43"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1903.10676","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}