Alab-NII/Awesome-SciLM

Pre-trained Language Model for Scientific Text

46

39 commits

updated Feb 22, 2024

See the code

README

Awesome PLM for Scientific Text (SciLM)

Resources for the paper: A Survey of Pre-trained Language Models for Processing Scientific Text

The Evolutionary tree of SciLMs

A tree file

Table of Contents

Existing SciLMs

Biomedical Domain

No.YearNameBase-modelObjective#ParametersCode
12019/01BioBERTBERTMLM, NSP110MGitHub
22019/02BERT-MIMICBERTMLM, NSP110M, 340MN/A
32019/04BioELMoELMoBi-LM93.6MGitHub
42019/04Clinical BERT (Emily)BERTMLM, NSP110MGitHub
52019/04ClinicalBERT (Kexin)BERTMLM, NSP110MGitHub
62019/06BlueBERTBERTMLM, NSP110M, 340MGitHub
72019/06G-BERTGNN + BERTSelf-Prediction, Dual-Prediction3MGitHub
82019/07BEHRTBERTMLM, NSPN/AGitHub
92019/08BioFLAIRFLAIRBi-LMN/AGitHub
102019/09EhrBERTBERTMLM, NSP110MGitHub
112019/12Clinical XLNetXLNetGeneralized Autoregressive Pretraining110MGitHub
122020/04GreenBioBERTBERTCBOW Word2Vec, Word Vector Space Alignment110MGitHub
132020/05BERT-XMLBERTMLM, NSPN/AN/A
142020/05Bio-ELECTRAELECTRAReplaced Token Prediction14MGitHub
152020/05Med-BERTBERTMLM, Prolonged LOS Prediction110MGitHub
162020/05ouBioBERTBERTMLM, NSP110MGitHub
172020/07PubMedBERTBERTMLM, NSP, Whole-Word Masking110MHuggingFace
182020/08MCBERTBERTMLM, NSP110M, 340MGitHub
192020/09BioALBERTALBERTMLM, SOP12M, 18MGitHub
202020/09BRLTMBERTMLMN/AGitHub
212020/10BioMegatronMegatronMLM, NSP345M, 800M, 1.2BGitHub
222020/10CharacterBERTBERT + Character-CNNMLM, NSP105MGitHub
232020/10ClinicalTransformerBERT - ALBERT - RoBERTa - ELECTRAMLM, NSP - MLM, SOP - MLM - Replaced Token Prediction110M - 12M - 125M - 110MGitHub
242020/10SapBERTBERTMulti-Similarity Loss110MGitHub
252020/10UmlsBERTBERTMLM110MGitHub
262020/11bert-for-radiologyBERTMLM, NSP110MGitHub
272020/11Bio-LMRoBERTaMLM125M, 355MGitHub
282020/11CODERPubMedBERT - mBERTContrastive Learning110M - 110MGitHub
292020/11exBERTBERTMLM, NSPN/AGitHub
302020/12BioMedBERTBERTMLM, NSP340MGitHub
312020/12LBERTBERTMLM, NSP110MGitHub
322021/04CovidBERTBioBERTMLM, NSP110MN/A
332021/04ELECTRAMedELECTRAReplaced Token PredictionN/AGitHub
342021/04KeBioLMPubMedBERTMLM, Entity Detection, Entity Linking110MGitHub
352021/04SINA-BERTBERTMLM110MN/A
362021/05ProteinBERTBERTCorrupted Token, Annotation Prediction16MGitHub
372021/05SciFiveT5Span Corruption Prediction220M, 770MGitHub
382021/06BioELECTRAELECTRAReplaced Token Prediction110MGitHub
392021/06EntityBERTBERTEntity-centric MLM110MN/A
402021/07MedGPTGPT-2 + GLU + RotaryEmbedLMN/AN/A
412021/08SMedBERTSMedBERTMasked Neighbor Modeling, Masked Mention Modeling, SOP, MLMN/AGitHub
422021/09Bio-cliRoBERTaMLM, Subword Masking or Whole Word Masking125MGitHub
432021/11UTH-BERTBERTMLM, NSP110MGitHub
442021/12ChestXRayBERTBERTMLM, NSP110MN/A
452021/12MedRoBERTa.nlRoBERTaMLM123MGitHub
462021/12PubMedELECTRAELECTRAReplaced Token Prediction110M, 335MHuggingFace
472022/01Clinical-BigBirdBigBirdMLM166MGitHub
482022/01Clinical-LongformerLongformerMLM149MGitHub
492022/03BioLinkBERTBERTMLM, Document Relation Prediction110M, 340MGitHub
502022/04BioBARTBARTText Infilling, Sentence Permutation140M, 400MGitHub
512022/05bsc-bio-ehr-esRoBERTaMLM125MGitHub
522022/05PathologyBERTBERTMLM, NSP110MHuggingFace
532022/06RadBERTRoBERTaMLM110MGitHub
542022/06ViHealthBERTBERTMLM, NSP, Capitalized Prediction110MGitHub
552022/07Clinical FlairFlairCharacter-level Bi-LMN/AGitHub
562022/08KM-BERTBERTMLM, NSP99MGitHub
572022/09BioGPTGPTAutoregressive Language Model347M, 1.5BGitHub
582022/10BioberturkBERTMLM, NSPN/AGitHub
592022/10DRAGONGreaseLMMLM, KG Link Prediction360MGitHub
602022/10UCSF-BERTBERTMLM, NSP135MN/A
612022/10ViPubmedT5ViT5Spans-masking learning220MGitHub
622022/12ALIBERTBERTMLM110MN/A
632022/12BioMedLMGPT2Autoregressive Language Model2.7BGitHub
642022/12BioReaderT5 & RETROMLM229.5MGitHub
652022/12clinicalT5T5Span-mask Denoising Objective220M, 770MN/A
662022/12GatortronBERTMLM8.9BGitHub
672022/12Med-PaLMFlan-PaLMInstruction Prompt Tuning540BOfficial Site
682023/01clinical-T5T5Fill-in-the-blank-style denoising objective220M, 770MPhysioNet
692023/01CPT-BigBirdBigBirdMLM166MN/A
702023/01CPT-LongformerLongformerMLM149MN/A
712023/02BioformerBioformerMLM, NSP43MGitHub
722023/02LightweightDistilBERTMLM, Knowledge Distillation65M, 25M, 18M, 15MGitHub
732023/03RAMMPubmedBERTMLM, Contrastive Learning, Image-Text MatchingN/AGitHub
742023/04DrBERTRoBERTaMLM110MGitHub
752023/04MOTORBLIPMLM, Contrastive Learning, Image-Text MatchingN/AGitHub
762023/05BiomedGPTBART backbone + BERT-encoder + GPT-decoderMLM33M, 93M, 182MGitHub
772023/05TurkRadBERTBERTMLM, NSP110MN/A
782023/06CamemBERT-bioBERTWhole Word MLM111MHuggingFace
792023/06ClinicalGPTT5Supervised Fine Tuning, Rank-based TrainingN/AN/A
802023/06EriBERTaRoBERTaMLM125MN/A
812023/06PharmBERTBERTMLM110MGitHub
822023/07BioNARTBERTNon-AutoRegressive Model110MGitHub
832023/07BIOptimusBERTMLM110MGitHub
842023/07KEBLMBERTMLM, Contrastive Learning, Ranking ObjectiveN/AN/A
852023/09CPLLMLlama2Autoregressive Language Model, Supervised Fine Tuning13BGitHub
862023/11MedCPTBERTContrastive Learning, Ranking Objective110MGitHub

Chemical Domain

No.YearNameBase-modelObjective#ParametersCode
12020/03NukeBERTBERTMLM, NSP110MGitHub
22020/10ChemBERTaRoBERTaMLM125MGitHub
32021/05NukeLMSciBERT, RoBERTaMLM125M, 355M, 110MGitHub
42021/06ChemBERTRoBERTaMLM110MGitHub
52021/09MatSciBERTBERTMLM110MGitHub
62021/10MatBERTBERTMLM110MGitHub
72022/05BatteryBERTBERT, SciBERTMLM110MGitHub
82022/05ChemGPTGPTAutoregressive Language Model1BGitHub
92022/08MaterialsBERT (Shetty)PubMedBERTMLM, NSP, Whole-Word Masking110MGitHub
102022/08ProcessBERTBERTMLM, NSP110MN/A
112022/09ChemBERTa-2RoBERTaMLM, Multi-task Regression125MGitHub
122022/09MaterialBERT (Yoshitake)BERTMLM, NSP110MMDR
132023/01MolGenBARTSeq2Seq MLM, Autoregressive Language Modeling460M, 7BGithub
142023/08GIT-MolGIT-FormerXmodal-Text Matching, Xmodal-Text Contrastive Learning700MN/A
152023/10MolCAGalacticaMolecule Captioning, Molecule-Text Contrastive Learning1.4BGitHub

Multi-domain

No.YearNameBase-modelObjective#ParametersCode
12019/03SciBERT (CS + Bio)BERTMLM, NSP110MGitHub
22019/11S2ORC-SciBERTBERTMLM, NSP110MGitHub
32020/04SPECTERBERTTriple-loss110MGitHub
42021/03OAG-BERTBERTMLM110MGitHub
52022/05ScholarBERTBERTMLM770MHuggingFace
62022/06SciDEBERTaDeBERTaMLMN/AGitHub
72022/09CSL-T5T5Fill-in-the-blank-style denoising objective220MGitHub
82022/10AcademicRoBERTaRoBERTaMLM125MGitHub
92022/11GalacticaGPTAutoregressive Language Model125M, 1.3B, 6.7B, 30B, 120BGitHub
102022/11VarMAERoBERTaMLM110MN/A
112022/12SciBARTBARTMLM124M, 386MGithub
122023/05PattonGNN + BERTNetwork-contextualized MLM, Masked Node PredictionN/AGitHub

Other Scientific Domains

Sorted by Domain-name

No.YearNameBase-modelObjective#ParametersCodeDomain
12022/04SecureBERTRoBERTaMLM125MGitHub Cybersecurity
22022/12CySecBERTBERTMLM, NSP110MN/ACybersecurity
32021/05MathBERT (Peng)BERTMLM, Masked Substructure Prediction, Context Correspondence Prediction110MN/AMath
42021/06MathBERT (Shen)RoBERTaMLM110MGitHub Math
52021/10ClimateBertDistilROBERTAMLM66MGitHub Climate
62020/02SciGPT2GPT2LM124MGitHub CS
72023/06K2LLaMACosine Loss7BGitHub Geoscience
82023/03ManuBERTBERTMLM110M, 126MHuggingFaceManufaturing
92023/01ProtSTBERTMasked Protein Modeling, Contrastive Learning, Multi-modal Masked PredictionN/AGitHub Protein
102023/01SciEdBERTBERTMLM110MN/AScience Education
112022/06SsciBERTBERTMLM, NSP110MGitHub Social Science
bio-lms
chem-lms
geoscience-lm
language-models
large-language-models
math-lms
scientific-text
scilms

Alab-NII/Awesome-SciLM

Pre-trained Language Model for Scientific Text

46

39 commits

updated Feb 22, 2024

See the code

README

Awesome PLM for Scientific Text (SciLM)

Resources for the paper: A Survey of Pre-trained Language Models for Processing Scientific Text

The Evolutionary tree of SciLMs

A tree file

Table of Contents

Existing SciLMs

Biomedical Domain

No.YearNameBase-modelObjective#ParametersCode
12019/01BioBERTBERTMLM, NSP110MGitHub
22019/02BERT-MIMICBERTMLM, NSP110M, 340MN/A
32019/04BioELMoELMoBi-LM93.6MGitHub
42019/04Clinical BERT (Emily)BERTMLM, NSP110MGitHub
52019/04ClinicalBERT (Kexin)BERTMLM, NSP110MGitHub
62019/06BlueBERTBERTMLM, NSP110M, 340MGitHub
72019/06G-BERTGNN + BERTSelf-Prediction, Dual-Prediction3MGitHub
82019/07BEHRTBERTMLM, NSPN/AGitHub
92019/08BioFLAIRFLAIRBi-LMN/AGitHub
102019/09EhrBERTBERTMLM, NSP110MGitHub
112019/12Clinical XLNetXLNetGeneralized Autoregressive Pretraining110MGitHub
122020/04GreenBioBERTBERTCBOW Word2Vec, Word Vector Space Alignment110MGitHub
132020/05BERT-XMLBERTMLM, NSPN/AN/A
142020/05Bio-ELECTRAELECTRAReplaced Token Prediction14MGitHub
152020/05Med-BERTBERTMLM, Prolonged LOS Prediction110MGitHub
162020/05ouBioBERTBERTMLM, NSP110MGitHub
172020/07PubMedBERTBERTMLM, NSP, Whole-Word Masking110MHuggingFace
182020/08MCBERTBERTMLM, NSP110M, 340MGitHub
192020/09BioALBERTALBERTMLM, SOP12M, 18MGitHub
202020/09BRLTMBERTMLMN/AGitHub
212020/10BioMegatronMegatronMLM, NSP345M, 800M, 1.2BGitHub
222020/10CharacterBERTBERT + Character-CNNMLM, NSP105MGitHub
232020/10ClinicalTransformerBERT - ALBERT - RoBERTa - ELECTRAMLM, NSP - MLM, SOP - MLM - Replaced Token Prediction110M - 12M - 125M - 110MGitHub
242020/10SapBERTBERTMulti-Similarity Loss110MGitHub
252020/10UmlsBERTBERTMLM110MGitHub
262020/11bert-for-radiologyBERTMLM, NSP110MGitHub
272020/11Bio-LMRoBERTaMLM125M, 355MGitHub
282020/11CODERPubMedBERT - mBERTContrastive Learning110M - 110MGitHub
292020/11exBERTBERTMLM, NSPN/AGitHub
302020/12BioMedBERTBERTMLM, NSP340MGitHub
312020/12LBERTBERTMLM, NSP110MGitHub
322021/04CovidBERTBioBERTMLM, NSP110MN/A
332021/04ELECTRAMedELECTRAReplaced Token PredictionN/AGitHub
342021/04KeBioLMPubMedBERTMLM, Entity Detection, Entity Linking110MGitHub
352021/04SINA-BERTBERTMLM110MN/A
362021/05ProteinBERTBERTCorrupted Token, Annotation Prediction16MGitHub
372021/05SciFiveT5Span Corruption Prediction220M, 770MGitHub
382021/06BioELECTRAELECTRAReplaced Token Prediction110MGitHub
392021/06EntityBERTBERTEntity-centric MLM110MN/A
402021/07MedGPTGPT-2 + GLU + RotaryEmbedLMN/AN/A
412021/08SMedBERTSMedBERTMasked Neighbor Modeling, Masked Mention Modeling, SOP, MLMN/AGitHub
422021/09Bio-cliRoBERTaMLM, Subword Masking or Whole Word Masking125MGitHub
432021/11UTH-BERTBERTMLM, NSP110MGitHub
442021/12ChestXRayBERTBERTMLM, NSP110MN/A
452021/12MedRoBERTa.nlRoBERTaMLM123MGitHub
462021/12PubMedELECTRAELECTRAReplaced Token Prediction110M, 335MHuggingFace
472022/01Clinical-BigBirdBigBirdMLM166MGitHub
482022/01Clinical-LongformerLongformerMLM149MGitHub
492022/03BioLinkBERTBERTMLM, Document Relation Prediction110M, 340MGitHub
502022/04BioBARTBARTText Infilling, Sentence Permutation140M, 400MGitHub
512022/05bsc-bio-ehr-esRoBERTaMLM125MGitHub
522022/05PathologyBERTBERTMLM, NSP110MHuggingFace
532022/06RadBERTRoBERTaMLM110MGitHub
542022/06ViHealthBERTBERTMLM, NSP, Capitalized Prediction110MGitHub
552022/07Clinical FlairFlairCharacter-level Bi-LMN/AGitHub
562022/08KM-BERTBERTMLM, NSP99MGitHub
572022/09BioGPTGPTAutoregressive Language Model347M, 1.5BGitHub
582022/10BioberturkBERTMLM, NSPN/AGitHub
592022/10DRAGONGreaseLMMLM, KG Link Prediction360MGitHub
602022/10UCSF-BERTBERTMLM, NSP135MN/A
612022/10ViPubmedT5ViT5Spans-masking learning220MGitHub
622022/12ALIBERTBERTMLM110MN/A
632022/12BioMedLMGPT2Autoregressive Language Model2.7BGitHub
642022/12BioReaderT5 & RETROMLM229.5MGitHub
652022/12clinicalT5T5Span-mask Denoising Objective220M, 770MN/A
662022/12GatortronBERTMLM8.9BGitHub
672022/12Med-PaLMFlan-PaLMInstruction Prompt Tuning540BOfficial Site
682023/01clinical-T5T5Fill-in-the-blank-style denoising objective220M, 770MPhysioNet
692023/01CPT-BigBirdBigBirdMLM166MN/A
702023/01CPT-LongformerLongformerMLM149MN/A
712023/02BioformerBioformerMLM, NSP43MGitHub
722023/02LightweightDistilBERTMLM, Knowledge Distillation65M, 25M, 18M, 15MGitHub
732023/03RAMMPubmedBERTMLM, Contrastive Learning, Image-Text MatchingN/AGitHub
742023/04DrBERTRoBERTaMLM110MGitHub
752023/04MOTORBLIPMLM, Contrastive Learning, Image-Text MatchingN/AGitHub
762023/05BiomedGPTBART backbone + BERT-encoder + GPT-decoderMLM33M, 93M, 182MGitHub
772023/05TurkRadBERTBERTMLM, NSP110MN/A
782023/06CamemBERT-bioBERTWhole Word MLM111MHuggingFace
792023/06ClinicalGPTT5Supervised Fine Tuning, Rank-based TrainingN/AN/A
802023/06EriBERTaRoBERTaMLM125MN/A
812023/06PharmBERTBERTMLM110MGitHub
822023/07BioNARTBERTNon-AutoRegressive Model110MGitHub
832023/07BIOptimusBERTMLM110MGitHub
842023/07KEBLMBERTMLM, Contrastive Learning, Ranking ObjectiveN/AN/A
852023/09CPLLMLlama2Autoregressive Language Model, Supervised Fine Tuning13BGitHub
862023/11MedCPTBERTContrastive Learning, Ranking Objective110MGitHub

Chemical Domain

No.YearNameBase-modelObjective#ParametersCode
12020/03NukeBERTBERTMLM, NSP110MGitHub
22020/10ChemBERTaRoBERTaMLM125MGitHub
32021/05NukeLMSciBERT, RoBERTaMLM125M, 355M, 110MGitHub
42021/06ChemBERTRoBERTaMLM110MGitHub
52021/09MatSciBERTBERTMLM110MGitHub
62021/10MatBERTBERTMLM110MGitHub
72022/05BatteryBERTBERT, SciBERTMLM110MGitHub
82022/05ChemGPTGPTAutoregressive Language Model1BGitHub
92022/08MaterialsBERT (Shetty)PubMedBERTMLM, NSP, Whole-Word Masking110MGitHub
102022/08ProcessBERTBERTMLM, NSP110MN/A
112022/09ChemBERTa-2RoBERTaMLM, Multi-task Regression125MGitHub
122022/09MaterialBERT (Yoshitake)BERTMLM, NSP110MMDR
132023/01MolGenBARTSeq2Seq MLM, Autoregressive Language Modeling460M, 7BGithub
142023/08GIT-MolGIT-FormerXmodal-Text Matching, Xmodal-Text Contrastive Learning700MN/A
152023/10MolCAGalacticaMolecule Captioning, Molecule-Text Contrastive Learning1.4BGitHub

Multi-domain

No.YearNameBase-modelObjective#ParametersCode
12019/03SciBERT (CS + Bio)BERTMLM, NSP110MGitHub
22019/11S2ORC-SciBERTBERTMLM, NSP110MGitHub
32020/04SPECTERBERTTriple-loss110MGitHub
42021/03OAG-BERTBERTMLM110MGitHub
52022/05ScholarBERTBERTMLM770MHuggingFace
62022/06SciDEBERTaDeBERTaMLMN/AGitHub
72022/09CSL-T5T5Fill-in-the-blank-style denoising objective220MGitHub
82022/10AcademicRoBERTaRoBERTaMLM125MGitHub
92022/11GalacticaGPTAutoregressive Language Model125M, 1.3B, 6.7B, 30B, 120BGitHub
102022/11VarMAERoBERTaMLM110MN/A
112022/12SciBARTBARTMLM124M, 386MGithub
122023/05PattonGNN + BERTNetwork-contextualized MLM, Masked Node PredictionN/AGitHub

Other Scientific Domains

Sorted by Domain-name

No.YearNameBase-modelObjective#ParametersCodeDomain
12022/04SecureBERTRoBERTaMLM125MGitHub Cybersecurity
22022/12CySecBERTBERTMLM, NSP110MN/ACybersecurity
32021/05MathBERT (Peng)BERTMLM, Masked Substructure Prediction, Context Correspondence Prediction110MN/AMath
42021/06MathBERT (Shen)RoBERTaMLM110MGitHub Math
52021/10ClimateBertDistilROBERTAMLM66MGitHub Climate
62020/02SciGPT2GPT2LM124MGitHub CS
72023/06K2LLaMACosine Loss7BGitHub Geoscience
82023/03ManuBERTBERTMLM110M, 126MHuggingFaceManufaturing
92023/01ProtSTBERTMasked Protein Modeling, Contrastive Learning, Multi-modal Masked PredictionN/AGitHub Protein
102023/01SciEdBERTBERTMLM110MN/AScience Education
112022/06SsciBERTBERTMLM, NSP110MGitHub Social Science
bio-lms
chem-lms
geoscience-lm
language-models
large-language-models
math-lms
scientific-text
scilms