{"id":"W4410282617","doi":"10.18280/ijsse.150310","title":"A Hybrid Semantic Enrichment Approach for Multi-Label Toxic Speech Detection","year":2025,"lang":"en","type":"article","venue":"International Journal of Safety and Security Engineering","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Computer science; Natural language processing; Speech recognition; Artificial intelligence","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003846833,0.00009624616,0.0001373049,0.0002688605,0.0000595429,0.0001101801,0.0003281902,0.00003994169,9.58117e-7],"category_scores_gemma":[0.00008769098,0.00009348631,0.00007269609,0.0001073956,0.000009212293,0.0003242896,0.00007321921,0.000173688,4.879785e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009932664,"about_ca_system_score_gemma":0.00003122201,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007015547,"about_ca_topic_score_gemma":0.000001668072,"domain_scores_codex":[0.9992058,0.000013233,0.0003075316,0.0001391257,0.0002088523,0.000125413],"domain_scores_gemma":[0.9994897,0.00005289962,0.0001055618,0.00007870916,0.0002209925,0.00005210286],"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":[0.0005839021,0.0007492109,0.0001610157,0.0003537137,0.001105917,0.0001441686,0.001516821,0.02019718,0.06088953,0.03581505,0.0001410486,0.8783424],"study_design_scores_gemma":[0.001182992,0.0000839116,0.0002563788,0.00007523815,0.00001441823,0.0003284095,0.00002208945,0.9697686,0.02490757,0.0006736518,0.002586109,0.0001005863],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03380816,0.0002903731,0.9642014,0.0002768857,0.001210154,0.0001016407,0.000002946752,0.00003367857,0.00007474702],"genre_scores_gemma":[0.8855551,0.0001831825,0.1139967,0.00006612926,0.000137818,0.000003903624,0.000001593993,0.000004646196,0.00005094325],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9495715,"threshold_uncertainty_score":0.3812261,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01027643679605816,"score_gpt":0.2443037246224196,"score_spread":0.2340272878263615,"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."}}