{"id":"W2802105481","doi":"10.18653/v1/s18-2005","title":"Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systems","year":2018,"lang":"en","type":"article","venue":"","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":395,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Computer science; Task (project management); Benchmark (surveying); Race (biology); Sentiment analysis; Artificial intelligence; Affect (linguistics); Gender bias; Natural language processing; Machine learning; SemEval; Psychology; Social psychology; Sociology","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":[0.003104413,0.0004597492,0.0003009906,0.001473259,0.00116625,0.0008257236,0.0004903189,0.0005855319,0.001300243],"category_scores_gemma":[0.0082988,0.0002482088,0.0003645114,0.001120655,0.0006687326,0.0009364818,0.001174982,0.0005857146,0.0009654298],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008889002,"about_ca_system_score_gemma":0.0006695578,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006126354,"about_ca_topic_score_gemma":0.01314726,"domain_scores_codex":[0.996986,0.0009833695,0.0003203874,0.0005311794,0.0009184782,0.0002605829],"domain_scores_gemma":[0.9932388,0.002964801,0.0004198598,0.0008414267,0.002293182,0.000241855],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002475457,0.001129,0.4125614,0.001901127,0.000524634,0.000887074,0.009107465,0.006435086,0.08470576,0.004793693,0.1205988,0.3548805],"study_design_scores_gemma":[0.0002731148,0.001061708,0.7783202,0.0001917142,0.0002344447,0.0008284273,0.003640534,0.05190961,0.05671678,0.002464807,0.1042034,0.0001552915],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9801595,0.0006979308,0.002853278,0.0004080311,0.0001859877,0.0001339065,0.007131304,0.0004620111,0.007968077],"genre_scores_gemma":[0.961028,0.0002757828,0.006244482,0.000259029,0.00009599561,0.0001881253,0.02820292,0.0001563802,0.003549331],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.006126354,"threshold_uncertainty_score":0.01641792,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06497271295094052,"score_gpt":0.290410277792946,"score_spread":0.2254375648420055,"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."}}