{"id":"W2952264314","doi":"10.48550/arxiv.0906.4019","title":"On the prevalence and scientific impact of duplicate publications in different scientific fields (1980-2007)","year":2009,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Metadata; Library science; Field (mathematics); Scientific literature; Scientific field; Data science; Information retrieval; Geography; Computer science; World Wide Web; Mathematics; Biology; 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":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.003636736,0.000233369,0.0002938776,0.0003084463,0.0008917399,0.0004286118,0.001221427,0.0008354036,0.000433589],"category_scores_gemma":[0.001657126,0.0001510602,0.0001785906,0.0006094601,0.001503703,0.0001644982,0.0003275402,0.002328161,0.00005064354],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001936084,"about_ca_system_score_gemma":0.0005750085,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001725333,"about_ca_topic_score_gemma":0.0008508682,"domain_scores_codex":[0.9974047,0.000432199,0.0004136616,0.000708862,0.0005943686,0.0004462267],"domain_scores_gemma":[0.9978358,0.00068126,0.0003075095,0.0007955702,0.0002409963,0.0001388627],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00002636999,0.0003571056,0.847349,0.0001519634,0.00005143862,0.000001475603,0.02224694,0.0001078509,0.0004068098,0.100915,0.02687173,0.001514368],"study_design_scores_gemma":[0.00009794953,0.00004197388,0.9543198,0.0003276522,0.0000316312,5.689673e-7,0.0002759192,0.0002300883,0.0001832097,0.04033259,0.003919149,0.0002394171],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9832956,0.0005124648,0.00005953035,0.009394536,0.0007769538,0.0006589025,0.00007683506,0.00004247219,0.005182708],"genre_scores_gemma":[0.9911066,0.000310468,0.00001381596,0.0001070175,0.000151818,0.00004605049,0.00002604788,0.000007580513,0.008230651],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1069709,"threshold_uncertainty_score":0.9999735,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07229518415060894,"score_gpt":0.3446323455733842,"score_spread":0.2723371614227753,"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."}}