{"id":"W4286632124","doi":"10.1007/s10994-022-06191-y","title":"Improving fairness generalization through a sample-robust optimization method","year":2022,"lang":"en","type":"article","venue":"Machine Learning","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":16,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université du Québec à Montréal; École de Technologie Supérieure","funders":"Canada Research Chairs; Centre International de Mathématiques et Informatique de Toulouse; Agence Nationale de la Recherche","keywords":"Computer science; Robustness (evolution); Machine learning; Scalability; Artificial intelligence; Generalization; Fairness measure; Heuristic; Metric (unit); Set (abstract data type); Generalization error; Artificial neural network; Throughput; Mathematics","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":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002496083,0.0001023639,0.0001436974,0.00006496764,0.004227643,0.0002536459,0.000226326,0.00007992243,0.001260353],"category_scores_gemma":[0.002575912,0.0001170135,0.00006652403,0.0005274941,0.000062877,0.0004652363,0.0001667775,0.0005484383,0.000002216869],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002157527,"about_ca_system_score_gemma":0.0001739304,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.05063549,"about_ca_topic_score_gemma":0.001573094,"domain_scores_codex":[0.997317,0.001404917,0.0001853931,0.0002357818,0.0005396652,0.0003172166],"domain_scores_gemma":[0.9991289,0.0003729562,0.0001878856,0.00009690522,0.0001444523,0.0000689505],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000006904916,0.00002107552,0.002162129,0.00000685319,0.000009655364,0.000001742555,0.0315007,0.9111732,0.00006748178,0.0490833,0.00008320552,0.005883798],"study_design_scores_gemma":[0.0003926461,0.0001016899,0.000103809,0.000006503498,0.00003159801,0.000001615608,0.01392374,0.879012,0.00002240416,0.008235883,0.09783359,0.0003345301],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002526646,0.0001410682,0.9782072,0.005898884,0.0003701423,0.0001927166,0.00001120238,0.0002259648,0.01242617],"genre_scores_gemma":[0.7019581,0.0001654084,0.2905459,0.00243998,0.0007364126,0.00005937321,0.0002745184,0.00007187079,0.003748477],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6994314,"threshold_uncertainty_score":0.9996526,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04930835796936283,"score_gpt":0.3639637421336215,"score_spread":0.3146553841642587,"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."}}