{"id":"W4414687920","doi":"10.1109/ccwc62904.2025.10903878","title":"An LLM-Based Framework for Synthetic Data Generation","year":2025,"lang":"en","type":"article","venue":"","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ontario Tech University","funders":"","keywords":"Key (lock); Synthetic data; Differential privacy; Scarcity; Information privacy; Identification (biology)","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":[],"consensus_categories":[],"category_scores_codex":[0.008014677,0.0007549914,0.0006356215,0.00145677,0.0007805285,0.002793589,0.003367598,0.00142232,0.005955935],"category_scores_gemma":[0.02359746,0.000615052,0.001614093,0.001329645,0.001772544,0.002848041,0.005385766,0.002694706,0.001729833],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001448291,"about_ca_system_score_gemma":0.002312401,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003105323,"about_ca_topic_score_gemma":0.004619014,"domain_scores_codex":[0.9959292,0.001693429,0.0002727507,0.0005755706,0.00134837,0.0001807523],"domain_scores_gemma":[0.9927434,0.003845908,0.0003718286,0.001887284,0.0008716978,0.0002798902],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003433205,0.0002070555,0.0025256,0.0003301731,0.0001123011,0.0005455097,0.0005974881,0.4451474,0.009074369,0.382232,0.01908958,0.1397953],"study_design_scores_gemma":[0.0000294549,0.00003188013,0.00008995167,0.00002287781,0.000006430399,0.00009854718,0.00003685619,0.9005049,0.002501834,0.08443444,0.01222444,0.00001841761],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001082006,0.00005222067,0.9955049,0.000206723,0.00002987782,0.00009652242,0.0003423985,0.001799394,0.0008860285],"genre_scores_gemma":[0.08820868,0.0001443809,0.9059401,0.0003260134,0.00005624089,0.0005660286,0.002380304,0.0006958832,0.001682351],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008014677,"threshold_uncertainty_score":0.04238617,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06413203143772486,"score_gpt":0.3591671951627064,"score_spread":0.2950351637249815,"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."}}