{"id":"W4396883491","doi":"10.1145/3665252.3665266","title":"Technical Perspective: Synthetic Data Needs a Reproducibility Benchmark","year":2024,"lang":"en","type":"article","venue":"ACM SIGMOD Record","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Benchmark (surveying); Perspective (graphical); Reproducibility; Data mining; Data science; Database; Artificial intelligence; Statistics; Mathematics","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":["metaresearch","scholarly_communication","open_science","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.02619974,0.0002010179,0.0003126135,0.0006347472,0.0002334209,0.001572976,0.007071964,0.00007990532,0.001771189],"category_scores_gemma":[0.06861946,0.000143177,0.0001436447,0.003349209,0.0002468397,0.0007824716,0.007319225,0.0003361301,0.00159456],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001344499,"about_ca_system_score_gemma":0.0001481231,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002851028,"about_ca_topic_score_gemma":0.0001135288,"domain_scores_codex":[0.992431,0.0003109316,0.0007699945,0.004559235,0.001523517,0.0004053491],"domain_scores_gemma":[0.9643829,0.004099552,0.0001280736,0.03087793,0.0003232709,0.0001882631],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001541384,0.00007628812,0.0006581307,0.000008819616,0.0000161159,0.00002669464,0.0002264695,0.000007613108,0.000139815,0.003120524,0.4692768,0.5264273],"study_design_scores_gemma":[0.0001194759,0.00009639403,0.00333063,0.0000909988,0.00005857412,0.00004767972,0.002127489,0.06341286,0.00006242527,0.1908135,0.7394693,0.0003707499],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2805924,0.00853009,0.2811883,0.09557384,0.04660543,0.003059756,0.001045012,0.004376065,0.2790291],"genre_scores_gemma":[0.9724417,0.00002305956,0.01762886,0.0001553593,0.0004794348,0.00001349098,0.00004664752,0.00002023447,0.009191191],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6918494,"threshold_uncertainty_score":0.9994635,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2209871303754449,"score_gpt":0.4407284741139029,"score_spread":0.219741343738458,"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."}}