{"id":"W4390305878","doi":"10.48550/arxiv.2312.15058","title":"The State of Documentation Practices of Third-party Machine Learning Models and Datasets","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Documentation; Computer science; Coding (social sciences); Transparency (behavior); Data science; Data mining; Information retrieval; Computer security; Programming language; Statistics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.224768,0.0006916631,0.0009740304,0.01404105,0.003902428,0.01690059,0.006090766,0.002958184,0.00454022],"category_scores_gemma":[0.572901,0.002124399,0.00110237,0.01421084,0.009673934,0.01777921,0.009798322,0.00574289,0.00407753],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007587036,"about_ca_system_score_gemma":0.0164205,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009822396,"about_ca_topic_score_gemma":0.004211716,"domain_scores_codex":[0.7516437,0.117098,0.03407838,0.01766196,0.07611649,0.003401397],"domain_scores_gemma":[0.1504477,0.4139572,0.0600227,0.271838,0.09926933,0.004465119],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.0006233972,0.0005258457,0.08065837,0.002289886,0.0002613937,0.0002993754,0.02451462,0.004674229,0.007235799,0.1486578,0.04672192,0.6835374],"study_design_scores_gemma":[0.0002489047,0.0006367143,0.08471648,0.0115645,0.0002854915,0.002249337,0.03128976,0.03001368,0.04363447,0.1920688,0.6024271,0.0008648131],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3761341,0.01192239,0.4380874,0.06202013,0.0008980995,0.001108951,0.009745332,0.0113225,0.08876109],"genre_scores_gemma":[0.7378297,0.00537557,0.2248146,0.005404464,0.0004857382,0.001192787,0.009563188,0.004670697,0.0106633],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7752321,"threshold_uncertainty_score":0.9559991,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2417275582009056,"score_gpt":0.3267408284696807,"score_spread":0.08501327026877512,"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."}}