{"id":"W3209074625","doi":"10.1080/08839514.2021.1988443","title":"Semi-Supervised Self-Training of Hate and Offensive Speech from Social Media","year":2021,"lang":"en","type":"article","venue":"Applied Artificial Intelligence","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Regina","funders":"","keywords":"Computer science; Classifier (UML); Artificial intelligence; Offensive; Leverage (statistics); Social media; Machine learning; Support vector machine; Arabic; Training set; Supervised learning; Labeled data; Semi-supervised learning; Natural language processing; Speech recognition; Pattern recognition (psychology); 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002226777,0.0001618902,0.0002774584,0.00006720195,0.0001868518,0.0001370861,0.0003396211,0.0001240936,0.00005866884],"category_scores_gemma":[0.00009377539,0.00018147,0.00006336372,0.0004503735,0.0001028532,0.0001255243,0.0002184591,0.0001961081,0.00008037332],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002257389,"about_ca_system_score_gemma":0.0001149068,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000714146,"about_ca_topic_score_gemma":0.0001291226,"domain_scores_codex":[0.9984133,0.00005376225,0.0003911274,0.0005243862,0.0003154686,0.0003019312],"domain_scores_gemma":[0.9989859,0.0003113735,0.0001203577,0.0003024475,0.0001798812,0.0001000834],"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":[0.00001936009,0.00006429067,0.00001774424,0.00001298321,0.00004409049,0.00004970299,0.01739565,0.00002334637,0.1699411,0.106811,0.00004725755,0.7055734],"study_design_scores_gemma":[0.00004920841,0.00002049843,0.0002559152,0.00001787247,0.00001814913,0.00001104203,0.002873933,0.008124658,0.8789272,0.1091408,0.0003345964,0.000226075],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.77006,0.0001381805,0.2251474,0.0004654296,0.0006776357,0.0001747443,0.00001354366,0.0002489204,0.003074188],"genre_scores_gemma":[0.9604223,0.00004313631,0.03905191,0.0001919426,0.0002480061,0.000009908195,0.000009741794,0.00001237678,0.00001067473],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7089862,"threshold_uncertainty_score":0.7400132,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04537766661798438,"score_gpt":0.2479983857249821,"score_spread":0.2026207191069977,"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."}}