{"id":"W2891947597","doi":"10.48550/arxiv.1711.06664","title":"Predict Responsibly: Improving Fairness and Accuracy by Learning to Defer","year":2017,"lang":"en","type":"article","venue":"arXiv (Cornell University)","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":57,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Decision maker; Process (computing); Decision process; Machine learning; Artificial intelligence; Simple (philosophy); Work (physics); Risk analysis (engineering); Management science; Engineering","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":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.0009550039,0.0001006174,0.0001308506,0.00006675789,0.003265032,0.0005177092,0.0004947627,0.00016439,0.00004422263],"category_scores_gemma":[0.003786454,0.0001168619,0.00004680731,0.0001221191,0.000410882,0.001090633,0.0002535865,0.0002997859,0.00003375385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009056446,"about_ca_system_score_gemma":0.0001840162,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006794811,"about_ca_topic_score_gemma":0.002239452,"domain_scores_codex":[0.9989562,0.0001912102,0.00007502818,0.0003276399,0.0001054591,0.000344483],"domain_scores_gemma":[0.9987749,0.0003320145,0.0001243829,0.0002664912,0.0001744143,0.0003277806],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004991193,0.0001516166,0.2418927,0.00005911239,0.0001420961,0.0002752528,0.06260695,0.0009024353,0.005009074,0.659819,0.006472166,0.02217046],"study_design_scores_gemma":[0.005352428,0.001224088,0.2807681,0.0003623287,0.0004357938,0.000004017913,0.1244242,0.008528354,0.001369871,0.151105,0.4228742,0.003551505],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9678242,0.00002194093,0.0009879305,0.002287035,0.0001349045,0.0001487815,0.00000550336,0.000076574,0.02851313],"genre_scores_gemma":[0.9890499,0.0001699336,0.0000350118,0.0001989448,0.00009614638,2.469946e-7,0.000001033193,0.00001051598,0.01043827],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.508714,"threshold_uncertainty_score":0.999819,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08803717292815989,"score_gpt":0.2645044095299627,"score_spread":0.1764672366018029,"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."}}