{"id":"W3158220826","doi":"10.1177/20597991211010416","title":"Big data and democratic speech: Predicting deliberative quality using machine learning techniques","year":2021,"lang":"en","type":"article","venue":"Methodological Innovations","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Ottawa","funders":"","keywords":"Newspaper; Quality (philosophy); Deliberative democracy; Computer science; Legislation; Democracy; Data science; Artificial intelligence; Machine learning; Political science; Law; Politics; Epistemology","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004050226,0.0001403154,0.0002502039,0.0001152774,0.0005329793,0.0002247453,0.000410948,0.0001225998,0.00001361057],"category_scores_gemma":[0.006968223,0.0001210062,0.00002563328,0.00151519,0.00007730081,0.0004246507,0.0009464601,0.0004919732,0.00000204018],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003092507,"about_ca_system_score_gemma":0.00008744124,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001039846,"about_ca_topic_score_gemma":0.00003592133,"domain_scores_codex":[0.996848,0.00166382,0.0004492867,0.0005987742,0.0002102867,0.0002297712],"domain_scores_gemma":[0.9978697,0.0009578858,0.0001815493,0.000603732,0.0003288091,0.00005828915],"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":[0.000008215475,0.00009199478,0.008680311,0.00003015073,0.00005518339,0.0000646316,0.0003158712,0.00006218404,0.3327806,0.0195471,0.00004364099,0.6383201],"study_design_scores_gemma":[0.0004585331,0.000259728,0.01218725,0.0001063421,0.00005463051,0.0007466393,0.0004067054,0.4077194,0.4814721,0.08949899,0.006284329,0.000805328],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1511118,0.0001218608,0.8467237,0.0009376603,0.0001737414,0.0001033947,0.000009589052,0.0003457724,0.0004725563],"genre_scores_gemma":[0.06419937,0.0000294468,0.9349084,0.0004610681,0.0001359068,0.000007798407,0.00005000068,0.000007300765,0.0002006594],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.6375148,"threshold_uncertainty_score":0.8342118,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5102523518130553,"score_gpt":0.4385821259810788,"score_spread":0.07167022583197652,"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."}}