{"id":"W3208821253","doi":"10.2200/s01123ed1v01y202108hlt053","title":"Pretrained Transformers for Text Ranking: BERT and Beyond","year":2021,"lang":"en","type":"article","venue":"Synthesis lectures on human language technologies","topic":"Topic Modeling","field":"Computer Science","cited_by":124,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Canada First Research Excellence Fund; Compute Canada","keywords":"Transformer; Computer science; Artificial intelligence; Natural language processing; Encoder; Ranking (information retrieval); Question answering; Information retrieval; Artificial neural network; Engineering","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0001885195,0.0002060703,0.000256541,0.0001988283,0.0002437618,0.0001553103,0.0005724548,0.0002065641,0.0000208065],"category_scores_gemma":[0.0006763894,0.0001738103,0.00009652957,0.0002052966,0.0001136468,0.0001314255,0.0001192607,0.0001988165,0.000002096408],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003584091,"about_ca_system_score_gemma":0.00003222991,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007015778,"about_ca_topic_score_gemma":0.0000572777,"domain_scores_codex":[0.9986782,0.00003388632,0.000191635,0.0005643268,0.0001870777,0.0003448566],"domain_scores_gemma":[0.9989268,0.0004144186,0.00005357841,0.00053715,0.00003968117,0.00002833674],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00001110915,0.00004346453,0.0000329175,0.00008529462,0.00006447042,0.00004525747,0.002746945,0.00002673637,0.05890688,0.07022757,0.0002686774,0.8675407],"study_design_scores_gemma":[0.0005238819,0.0001632003,0.0002085803,0.00009847185,0.00003822495,0.00004336768,0.003040518,0.003535236,0.9461379,0.04391542,0.001762768,0.0005324521],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3145986,0.005394342,0.6559633,0.01130738,0.0002018338,0.0007794431,0.00002221744,0.003860497,0.007872435],"genre_scores_gemma":[0.9650732,0.00003126814,0.03427654,0.0003055146,0.0000246799,0.0001133053,0.000002487438,0.0000177038,0.0001552532],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.887231,"threshold_uncertainty_score":0.7087777,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02076188500124745,"score_gpt":0.2760993430802039,"score_spread":0.2553374580789565,"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."}}