{"id":"W3155638757","doi":"","title":"Zero-shot Cross-lingual Content Filtering: Offensive Language and Hate Speech Detection","year":2021,"lang":"en","type":"article","venue":"Queen Mary Research Online (Queen Mary University of London)","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Engineering and Physical Sciences Research Council; European Commission; Javna Agencija za Raziskovalno Dejavnost RS; Institute for Catastrophic Loss Reduction","keywords":"Offensive; Computer science; Natural language processing; Speech recognition; Zero (linguistics); Classifier (UML); Artificial intelligence; Voice activity detection; Task (project management); Linguistics; Speech processing; Mathematics; Engineering","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0008514117,0.0002313733,0.0003661218,0.0003990194,0.0004946289,0.0001870044,0.0007358438,0.000192596,0.0001367844],"category_scores_gemma":[0.000252213,0.0002694452,0.0001460612,0.0007596883,0.0003651534,0.0007181349,0.001299378,0.0007842279,0.00004478715],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001926905,"about_ca_system_score_gemma":0.0003216537,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005668798,"about_ca_topic_score_gemma":0.0006543925,"domain_scores_codex":[0.9969246,0.0005821335,0.0002758651,0.000780943,0.0008085052,0.0006279908],"domain_scores_gemma":[0.9975312,0.0003104444,0.0001296751,0.0007173033,0.001000626,0.000310782],"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":[0.001739634,0.001025787,0.01211962,0.0007224719,0.000519397,0.01538896,0.003852296,0.0001777358,0.3436852,0.001151075,0.003501653,0.6161162],"study_design_scores_gemma":[0.008166324,0.003235172,0.1972465,0.0007101768,0.0001080534,0.001070226,0.009117852,0.01158414,0.6921455,0.002351287,0.07232395,0.001940775],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9850677,0.0002467442,0.01042026,0.00167523,0.0002214756,0.0003467689,0.00007055426,0.0001631849,0.001788139],"genre_scores_gemma":[0.93372,0.001204999,0.02937557,0.0002102513,0.000202703,0.00000143777,0.000158759,0.00003758093,0.03508869],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6141754,"threshold_uncertainty_score":0.9999758,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05353052957588103,"score_gpt":0.3054008333783639,"score_spread":0.2518703038024829,"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."}}