{"id":"W4287120901","doi":"10.1162/tacl_a_00492","title":"Generate, Annotate, and Learn: NLP with Synthetic Text","year":2022,"lang":"en","type":"article","venue":"Transactions of the Association for Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":37,"is_retracted":false,"has_abstract":true,"ca_institutions":"Google (Canada)","funders":"Defense Advanced Research Projects Agency","keywords":"Computer science; Artificial intelligence; Transformer; Natural language processing; Classifier (UML); Machine learning; Labeled data; Task (project management); Distillation; Language model","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002707636,0.0008602803,0.0004802741,0.0005363958,0.0005635999,0.001047708,0.001452629,0.001258601,0.004044543],"category_scores_gemma":[0.0154307,0.0003060373,0.0005009374,0.0005995124,0.001289586,0.00317167,0.001912209,0.001857437,0.00167907],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008224361,"about_ca_system_score_gemma":0.0006645097,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002102799,"about_ca_topic_score_gemma":0.003657861,"domain_scores_codex":[0.9982247,0.001002507,0.00005676873,0.0003669994,0.0002757785,0.00007326284],"domain_scores_gemma":[0.9900479,0.007523117,0.0002412275,0.00138134,0.0006321153,0.0001742586],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001051926,0.0006229797,0.003493272,0.0006655763,0.000131244,0.0005417097,0.0008018955,0.5767578,0.01292091,0.0633667,0.0351302,0.3045157],"study_design_scores_gemma":[0.00005055728,0.00005155817,0.0001548217,0.00001680769,0.000006946076,0.00004028418,0.00006872984,0.9605917,0.009361397,0.02536711,0.004277063,0.00001299745],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1763613,0.0007228936,0.7923977,0.002251224,0.0004031529,0.0002820592,0.003608165,0.01317655,0.01079697],"genre_scores_gemma":[0.6470042,0.0001651419,0.3397616,0.0004123927,0.00009916467,0.0003240173,0.00590084,0.0007953396,0.005537218],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004044543,"threshold_uncertainty_score":0.01431948,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01282966128692613,"score_gpt":0.2278401156433895,"score_spread":0.2150104543564634,"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."}}