{"id":"W2492590231","doi":"10.14778/2983200.2983203","title":"Distributed data deduplication","year":2016,"lang":"en","type":"article","venue":"Proceedings of the VLDB Endowment","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":80,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Data deduplication; Tuple; Computer science; Blocking (statistics); Block (permutation group theory); Relation (database); Backup; Locality; Process (computing); Theoretical computer science; Data mining; Database; 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":[0.002900639,0.0009624009,0.001522964,0.0009954371,0.001569028,0.002168587,0.002882895,0.0007436742,0.00288138],"category_scores_gemma":[0.008180099,0.0004749566,0.000636077,0.002344567,0.0008355245,0.003063466,0.003867195,0.001069586,0.001192661],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005335631,"about_ca_system_score_gemma":0.001859531,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007185501,"about_ca_topic_score_gemma":0.001153588,"domain_scores_codex":[0.9977374,0.0004630889,0.0002614409,0.0006499902,0.0006765156,0.0002115748],"domain_scores_gemma":[0.9898888,0.001886859,0.0005392479,0.005753434,0.001631465,0.0003000933],"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":[0.001907728,0.0007390953,0.01053621,0.001538366,0.0003914431,0.001142978,0.001013716,0.1671132,0.06034495,0.0474415,0.07294006,0.6348907],"study_design_scores_gemma":[0.000382287,0.000619042,0.003919673,0.0001325178,0.0001595308,0.00259024,0.001079541,0.7292244,0.1134366,0.07963128,0.06867234,0.000152516],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.05992539,0.001859596,0.9254689,0.0007775997,0.0004755423,0.0007247435,0.00170175,0.003398426,0.00566808],"genre_scores_gemma":[0.6663778,0.001041001,0.3197482,0.0004116982,0.0002313806,0.0006985886,0.003307602,0.0003751323,0.007808567],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.002900639,"threshold_uncertainty_score":0.01534027,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2701566814294362,"score_gpt":0.4095847037039765,"score_spread":0.1394280222745403,"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."}}