{"id":"W4226354324","doi":"10.3390/jrfm15040165","title":"Time to Assess Bias in Machine Learning Models for Credit Decisions","year":2022,"lang":"en","type":"article","venue":"Journal of risk and financial management","topic":"Law, AI, and Intellectual Property","field":"Computer Science","cited_by":38,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Underwriting; Loan; Analytics; Compliance (psychology); Actuarial science; Credit score; Computer science; Finance; Business; Artificial intelligence; Data science; Psychology","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":[0.001208008,0.00008097773,0.0001892035,0.0002820075,0.0002847089,0.00007867882,0.000471275,0.00001689541,0.00004221183],"category_scores_gemma":[0.0002812463,0.00006424607,0.00007462651,0.00033429,0.00001457598,0.0002652625,0.0005178619,0.0002364475,0.000006333312],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006105458,"about_ca_system_score_gemma":0.00003286732,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001365573,"about_ca_topic_score_gemma":0.000006197919,"domain_scores_codex":[0.9989662,0.00009291286,0.0003209896,0.0001610292,0.0002860785,0.0001727885],"domain_scores_gemma":[0.9994159,0.0001941006,0.0001399734,0.0001149196,0.00006262921,0.00007247495],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002729454,0.0002242456,0.0002344567,0.00001280208,0.00001896915,0.00008801107,0.002625569,0.1233827,0.00002353011,0.01140694,0.03109474,0.830615],"study_design_scores_gemma":[0.0008388976,0.00105553,0.0008436583,0.0000354573,0.00001798144,0.00002039826,0.00009963477,0.4705389,0.00001395023,0.02310469,0.5032805,0.0001504704],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01781595,0.0003205332,0.9797434,0.0003245574,0.0003208231,0.0001939196,0.00000605613,0.0000098244,0.001264996],"genre_scores_gemma":[0.9364023,0.001081812,0.05817746,0.0007906902,0.0002277826,0.00003597869,0.000001697559,0.00001591249,0.003266336],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9215659,"threshold_uncertainty_score":0.2619879,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06786324974592264,"score_gpt":0.2570016938658317,"score_spread":0.1891384441199091,"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."}}