{"id":"W4399174293","doi":"10.1145/3654921","title":"Reservoir Sampling over Joins","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Joins; Tuple; Computer science; Join (topology); Graph; Theoretical computer science; Sampling (signal processing); Operator (biology); Data mining; Mathematics; Discrete 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003015956,0.0007769198,0.001783506,0.0009654054,0.00126574,0.002381568,0.002643452,0.001085576,0.002230403],"category_scores_gemma":[0.01335884,0.000574609,0.001051603,0.002165377,0.001378831,0.00556631,0.00282901,0.001683432,0.0005774895],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001062091,"about_ca_system_score_gemma":0.002047708,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004195617,"about_ca_topic_score_gemma":0.004849725,"domain_scores_codex":[0.9973459,0.0005581336,0.0001816709,0.0007090566,0.0009259118,0.000279302],"domain_scores_gemma":[0.9932824,0.003448451,0.0004660331,0.001637653,0.0008532012,0.0003121359],"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":[0.00104348,0.0003130459,0.009125411,0.0004155293,0.0002032825,0.000449977,0.000558938,0.5258257,0.02396658,0.1249498,0.01220498,0.3009433],"study_design_scores_gemma":[0.00002675141,0.00007723278,0.0002910621,0.00001032443,0.00001969095,0.0001071244,0.00005669325,0.9583165,0.00559247,0.03265283,0.002834606,0.00001466632],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0223351,0.0003836729,0.9739688,0.0002340238,0.00006270807,0.0001062528,0.0002479372,0.001349232,0.001312327],"genre_scores_gemma":[0.468046,0.0005295398,0.525272,0.0003548824,0.0002109682,0.0002923068,0.001545711,0.0004016349,0.003347029],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004195617,"threshold_uncertainty_score":0.01595008,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.12639640691388,"score_gpt":0.35644566552144,"score_spread":0.23004925860756,"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."}}