{"id":"W3099008231","doi":"10.18653/v1/2020.findings-emnlp.129","title":"exBERT: Extending Pre-trained Models with Domain-specific Vocabulary Under Constrained Training Resources","year":2020,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":75,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Air Force Research Laboratory; MediaTek","keywords":"Computer science; Vocabulary; Benchmark (surveying); Artificial intelligence; Embedding; Domain (mathematical analysis); Context (archaeology); Machine learning; Computation; Training (meteorology); Training set; Natural language processing; Language model; Algorithm; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003081963,0.0002986382,0.0003434236,0.0001023517,0.0002158043,0.0003244567,0.001027506,0.00009280977,0.00007531247],"category_scores_gemma":[0.00000964571,0.0002421264,0.00008656112,0.0004738691,0.0001254206,0.000807122,0.0002436926,0.0002659293,0.00001586736],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004373954,"about_ca_system_score_gemma":0.0001115222,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001482071,"about_ca_topic_score_gemma":0.00000776046,"domain_scores_codex":[0.9974967,0.00008974205,0.0004038213,0.0009070759,0.0005285351,0.0005741619],"domain_scores_gemma":[0.9987118,0.0001811912,0.0001093636,0.0006023334,0.00005984489,0.0003354306],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000547196,0.00003245838,0.0000388298,0.00002644627,0.00006418759,0.0001323377,0.0357643,0.06198706,0.003751951,0.8663506,0.0001794552,0.03161764],"study_design_scores_gemma":[0.001541165,0.000225407,0.0001856523,0.0000856127,0.00001076362,0.0001329862,0.003273304,0.9596447,0.0004485802,0.0264769,0.007291042,0.0006839284],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05884343,0.0002040901,0.918448,0.005813148,0.0001032991,0.0003154206,0.000001576133,0.0006982194,0.01557287],"genre_scores_gemma":[0.6447484,0.000007867744,0.3535695,0.001378482,0.0001446608,0.00001828043,0.000001645912,0.00002130662,0.0001097547],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8976576,"threshold_uncertainty_score":0.9873629,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05795474199009502,"score_gpt":0.2343886152408255,"score_spread":0.1764338732507305,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}