{"id":"W2980073313","doi":"10.1162/coli_a_00363","title":"Scalable Micro-planned Generation of Discourse from Structured Data","year":2019,"lang":"en","type":"article","venue":"Zenodo (CERN European Organization for Nuclear Research)","topic":"Topic Modeling","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Natural language processing; Interpretability; Scalability; Artificial intelligence; Natural language generation; Sentence; Pipeline (software); Paragraph; Fluency; Natural language understanding; Data manipulation language; Natural language; Information retrieval; Programming language; Database; World Wide Web","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":[0.001696789,0.0009946523,0.0007113981,0.001597794,0.0006496345,0.001596066,0.001587084,0.0007630737,0.007550546],"category_scores_gemma":[0.008035829,0.0005589837,0.00123636,0.001222279,0.000643528,0.002357733,0.001969509,0.001125186,0.004113657],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007689247,"about_ca_system_score_gemma":0.001877761,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002695042,"about_ca_topic_score_gemma":0.004360904,"domain_scores_codex":[0.9987833,0.0003809242,0.00009299593,0.0004314519,0.0002597541,0.00005153469],"domain_scores_gemma":[0.9964249,0.002167141,0.0001971625,0.0005731996,0.0005316508,0.0001059235],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006137777,0.0002780102,0.002990936,0.001544853,0.0001537692,0.001014442,0.002863346,0.04618977,0.05499563,0.05493936,0.0722007,0.7622154],"study_design_scores_gemma":[0.0001285445,0.000133202,0.0008206692,0.0001036706,0.00007517619,0.0003626421,0.0006722469,0.7980615,0.06775742,0.05471589,0.0770875,0.00008150739],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01004811,0.0002184777,0.9574152,0.0003635624,0.0001088448,0.0002812419,0.003777748,0.02540688,0.002379897],"genre_scores_gemma":[0.0986937,0.0002069791,0.8821574,0.0001383264,0.00007518547,0.0004002144,0.01343284,0.001855715,0.003039528],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007550546,"threshold_uncertainty_score":0.02525908,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06693568368884965,"score_gpt":0.266282299033941,"score_spread":0.1993466153450913,"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."}}