{"id":"W4403978473","doi":"10.3390/info15110687","title":"Mitigating Bias Due to Race and Gender in Machine Learning Predictions of Traffic Stop Outcomes","year":2024,"lang":"en","type":"article","venue":"Information","topic":"Policing Practices and Perceptions","field":"Social Sciences","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Nova Scotia Health Authority; St. Francis Xavier University","funders":"Nova Scotia Health Authority; Natural Sciences and Engineering Research Council of Canada; Compute Canada; Canada Foundation for Innovation; Nova Scotia Research Innovation Trust","keywords":"Race (biology); Law enforcement; Computer science; Machine learning; Artificial intelligence; Demographics; Selection bias; Enforcement; Omitted-variable bias; Feature selection; Point (geometry); Data science; Psychology; Statistics; Political science; Law; Mathematics; Sociology","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0005229643,0.00003329745,0.00005827296,0.000154409,0.0001578047,0.00008743807,0.00003096476,0.00003285611,0.00005071915],"category_scores_gemma":[0.0003049483,0.0000310903,0.00001623246,0.0002459009,0.00002491258,0.001052442,0.00001177515,0.0001074566,0.00002834213],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003489712,"about_ca_system_score_gemma":0.00004298104,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005562594,"about_ca_topic_score_gemma":0.003901325,"domain_scores_codex":[0.9995388,0.00005646045,0.0001659672,0.00003530313,0.0001189117,0.00008452295],"domain_scores_gemma":[0.9996851,0.0001864016,0.00004080131,0.00003002798,0.00002182483,0.00003584423],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.000002945621,0.000008308518,0.02059925,0.00006172738,0.00001023065,3.190567e-7,0.8869227,0.03415072,0.00002221757,0.003214406,0.0002303817,0.05477682],"study_design_scores_gemma":[0.0003214162,0.00007568674,0.5802049,0.0001730759,0.00003879853,0.000008904706,0.0774777,0.1956122,0.00001263116,0.0003028481,0.145544,0.0002277559],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9892293,0.00002275579,0.0007286728,0.001881257,0.0001213513,0.0001014929,0.00001078029,0.00006090022,0.007843479],"genre_scores_gemma":[0.9992778,0.00004882811,0.0003066848,0.0001022186,0.00002708965,0.00000652204,0.000008380539,0.000001905226,0.0002205274],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.809445,"threshold_uncertainty_score":0.8409016,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.067869844401849,"score_gpt":0.3712057789266373,"score_spread":0.3033359345247882,"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."}}