{"id":"W4283155630","doi":"10.1145/3531146.3533231","title":"Data Cards: Purposeful and Transparent Dataset Documentation for Responsible AI","year":2022,"lang":"en","type":"article","venue":"2022 ACM Conference on Fairness, Accountability, and Transparency","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":183,"is_retracted":false,"has_abstract":true,"ca_institutions":"Google (Canada)","funders":"","keywords":"Documentation; Computer science; Information retrieval; Data science; Artificial intelligence; Programming language","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":["metaepi_narrow","sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00319801,0.0002587048,0.0003971932,0.0001152957,0.002499776,0.0006252712,0.001276363,0.0001421148,0.001107608],"category_scores_gemma":[0.0007344351,0.0002769102,0.00005817194,0.0002932191,0.0006750902,0.001562047,0.0002941811,0.0005127971,0.000003010521],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001307562,"about_ca_system_score_gemma":0.001003594,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01020072,"about_ca_topic_score_gemma":0.04500255,"domain_scores_codex":[0.9967067,0.0005069273,0.00044543,0.0009232921,0.00090339,0.000514219],"domain_scores_gemma":[0.9974937,0.0008916375,0.0001289964,0.001015613,0.0002429613,0.0002270491],"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":[0.003110604,0.002126134,0.05778648,0.001411939,0.000380887,0.00002513106,0.1711766,0.00006889227,0.0009928619,0.6627544,0.04176631,0.05839973],"study_design_scores_gemma":[0.003355489,0.001020944,0.02006749,0.00006230009,0.0003653372,0.000004023597,0.08045598,0.0004981803,0.00006647531,0.4409882,0.4518704,0.001245221],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6059167,0.003446637,0.002305879,0.183146,0.003865204,0.007275024,0.1834848,0.000421589,0.01013813],"genre_scores_gemma":[0.9905439,0.002914493,0.000183164,0.001041592,0.000144445,0.0002982079,0.004687295,0.00002221982,0.0001647071],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4101041,"threshold_uncertainty_score":0.9999683,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2238368987664745,"score_gpt":0.4491079445249052,"score_spread":0.2252710457584307,"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."}}