{"id":"W4391018115","doi":"10.1145/3593579","title":"Record Fusion via Inference and Data Augmentation","year":2024,"lang":"en","type":"article","venue":"ACM / IMS Journal of Data Science","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Inference; Probabilistic logic; Sensor fusion; Data mining; Fusion; Data source; Artificial intelligence; Machine learning","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":[0.01136022,0.002045817,0.002861806,0.005840573,0.001401253,0.004045477,0.006431152,0.002550175,0.003121499],"category_scores_gemma":[0.04394877,0.001350543,0.00302881,0.008386554,0.002064916,0.01172472,0.008258821,0.005841135,0.002104602],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001753602,"about_ca_system_score_gemma":0.004067641,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006720821,"about_ca_topic_score_gemma":0.007839037,"domain_scores_codex":[0.9898854,0.002585058,0.0008387254,0.002932916,0.003308429,0.0004492989],"domain_scores_gemma":[0.9761038,0.01008696,0.00211629,0.0083159,0.003017836,0.0003592208],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000764413,0.0006248729,0.01398272,0.0006233287,0.0006596883,0.0003834111,0.001034321,0.1961879,0.007492799,0.03623721,0.01979479,0.7222146],"study_design_scores_gemma":[0.00004992895,0.00009711823,0.001325699,0.00008393019,0.0001459866,0.0002360265,0.0001464086,0.9185882,0.009345485,0.05918368,0.01073118,0.00006645304],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005955284,0.0004099309,0.9877331,0.0005256424,0.00009234047,0.0001325693,0.0008212654,0.003461639,0.0008682154],"genre_scores_gemma":[0.1695445,0.0004105394,0.8218092,0.0005150529,0.0002642959,0.0003005081,0.005018301,0.0002717017,0.001865879],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01136022,"threshold_uncertainty_score":0.06007928,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5120775299749007,"score_gpt":0.5555716160536497,"score_spread":0.04349408607874894,"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."}}