{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":43,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":43,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"0bc1e9e6864c","filters":{"venue":"Proceedings of the ACM on Management of Data"}},"results":[{"id":"W4380433238","doi":"10.1145/3588911","title":"A Universal Question-Answering Platform for Knowledge Graphs","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Topic Modeling","field":"Computer Science","cited_by":41,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"SPARQL; Computer science; Question answering; RDF; Information retrieval; Set (abstract data type); Named graph; Representation (politics); Query language; Artificial intelligence; Semantic Web; Programming language","authors":[{"name":"Reham Omar","is_ca":true},{"name":"Ishika Dhall","is_ca":true},{"name":"Panos Kalnis","is_ca":false},{"name":"E. M. E. Mansour","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09927151819617334,"gpt":0.3177039142849232,"spread":0.2184323960887498,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002968806,0.001260043,0.001002697,0.003021425,0.001101476,0.002726241,0.003433527,0.002045022,0.0174754],"category_scores_gemma":[0.01340951,0.000896132,0.002043104,0.002379297,0.001253013,0.008511955,0.006681971,0.002846384,0.008386753],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001705405,"about_ca_system_score_gemma":0.002082558,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00675333,"about_ca_topic_score_gemma":0.009189885,"domain_scores_codex":[0.997767,0.0005851346,0.0002432296,0.0007804951,0.0004831373,0.0001409752],"domain_scores_gemma":[0.9957061,0.001740068,0.0002330425,0.001535876,0.0005514247,0.0002334641],"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":[0.0008156921,0.0005498072,0.002437128,0.002274892,0.0002473286,0.0006966301,0.00201365,0.03124745,0.02512683,0.1413936,0.2107269,0.5824702],"study_design_scores_gemma":[0.0001705805,0.0001522786,0.00123337,0.0002692183,0.0001100895,0.0005038751,0.0004457574,0.4048423,0.02046156,0.2807549,0.2909272,0.000128828],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005405475,0.0006594458,0.8065154,0.001115167,0.000179412,0.0007826047,0.00944581,0.1695931,0.006303594],"genre_scores_gemma":[0.1155758,0.0008121913,0.8234657,0.001559789,0.0001396912,0.0009489192,0.0434387,0.006509243,0.007549932],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0174754,"threshold_uncertainty_score":0.05846101,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4380433250","doi":"10.1145/3588704","title":"How To Optimize My Blockchain? A Multi-Level Recommendation Approach","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Blockchain Technology Applications and Security","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Deutsche Forschungsgemeinschaft","keywords":"Blockchain; Computer science; Distributed computing; Data science; Computer security","authors":[{"name":"Jeeta Ann Chacko","is_ca":false},{"name":"Ruben Mayer","is_ca":false},{"name":"Hans‐Arno Jacobsen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1258081962844258,"gpt":0.3060546250598179,"spread":0.1802464287753921,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005083583,0.001046649,0.001041359,0.001509901,0.000948381,0.00329447,0.001819123,0.001248065,0.005077423],"category_scores_gemma":[0.01367918,0.0007324126,0.0007181841,0.001586552,0.0006721526,0.00617802,0.001215058,0.001759812,0.002162667],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001324573,"about_ca_system_score_gemma":0.001872767,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006333469,"about_ca_topic_score_gemma":0.01411885,"domain_scores_codex":[0.9957467,0.001903822,0.0003306338,0.0005648243,0.001162466,0.0002916517],"domain_scores_gemma":[0.9932632,0.002583824,0.0005235853,0.001957203,0.00146838,0.0002038603],"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":[0.0003276455,0.0003006782,0.0105804,0.000591523,0.0003199822,0.0002901067,0.0006340408,0.2399277,0.01158488,0.04179916,0.02221298,0.6714309],"study_design_scores_gemma":[0.00003494658,0.000106032,0.001039983,0.0001021826,0.00006653221,0.0001174137,0.0002790126,0.9446456,0.006652127,0.02863422,0.01827415,0.00004787011],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03829971,0.001449962,0.9388972,0.00432335,0.00007613708,0.0004000922,0.0004801721,0.004712659,0.01136072],"genre_scores_gemma":[0.3612569,0.001110819,0.6301237,0.0004892088,0.00006770757,0.0002085126,0.0009265032,0.0004645197,0.005352175],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006333469,"threshold_uncertainty_score":0.02688485,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4393212300","doi":"10.1145/3639311","title":"Fast Shapley Value Computation in Data Assemblage Tasks as Cooperative Simple Games","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Shapley value; Simple (philosophy); Computation; Assemblage (archaeology); Value (mathematics); Computer science; Mathematical economics; Mathematics; Game theory; Algorithm; Geography; Machine learning; Archaeology","authors":[{"name":"Xuan Luo","is_ca":true},{"name":"Jian Pei","is_ca":false},{"name":"Cheng Xu","is_ca":false},{"name":"Wenjie Zhang","is_ca":false},{"name":"Jianliang Xu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06813038036782831,"gpt":0.3467607599894598,"spread":0.2786303796216315,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004068304,0.001234872,0.002442324,0.001025583,0.001300864,0.003458945,0.002468433,0.001462496,0.006154745],"category_scores_gemma":[0.01412111,0.0008195044,0.001616426,0.001691439,0.002112315,0.007486124,0.003260073,0.002762007,0.0006498927],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002286338,"about_ca_system_score_gemma":0.002581684,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002089321,"about_ca_topic_score_gemma":0.002894734,"domain_scores_codex":[0.9970905,0.001230032,0.0001558505,0.0006239296,0.0005024816,0.0003972562],"domain_scores_gemma":[0.990493,0.007098612,0.0004899684,0.000984392,0.0004252312,0.0005088446],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0005371847,0.0002946581,0.001849684,0.000397461,0.0001610624,0.0002881309,0.000703899,0.5261145,0.004130696,0.3887685,0.003705767,0.07304844],"study_design_scores_gemma":[0.00005681674,0.00004371314,0.00009468983,0.00001185537,0.00001617808,0.00002934809,0.000076885,0.6923015,0.0009437691,0.3056066,0.000806148,0.00001257364],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09026507,0.0001183532,0.9046885,0.0003923664,0.00004372877,0.0002878548,0.0002122901,0.0003788177,0.003612996],"genre_scores_gemma":[0.6199874,0.0001466747,0.3751647,0.000185607,0.00005050128,0.0004040474,0.0004756186,0.0001805988,0.003404899],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006154745,"threshold_uncertainty_score":0.02151549,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4402426760","doi":"10.1145/3698820","title":"Memento Filter: A Fast, Dynamic, and Robust Range Filter","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Caching and Content Delivery","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Universitas Brawijaya","keywords":"Computer science; Range query (database); Filter (signal processing); Bloom filter; Key (lock); Range (aeronautics); Data mining; Information retrieval; Artificial intelligence; Algorithm; Computer vision; Search engine; Computer security; Web search query","authors":[{"name":"Navid Eslami","is_ca":false},{"name":"Niv Dayan","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05317012316491179,"gpt":0.2665536884634266,"spread":0.2133835652985148,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001831068,0.0009973404,0.001239916,0.00249449,0.001046283,0.002577498,0.003030185,0.001431913,0.005982],"category_scores_gemma":[0.01147281,0.0006940459,0.0006838584,0.002293058,0.0009327604,0.005998175,0.00279922,0.001057449,0.003661894],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001175643,"about_ca_system_score_gemma":0.001875698,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004785644,"about_ca_topic_score_gemma":0.005631734,"domain_scores_codex":[0.997474,0.0001871796,0.0002382139,0.0004365246,0.001411552,0.0002525423],"domain_scores_gemma":[0.994413,0.001678989,0.0005318301,0.001737738,0.00144581,0.0001927409],"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":[0.002314446,0.0002847209,0.01018517,0.0005392072,0.0001482248,0.0004609201,0.0005201029,0.01809736,0.05744437,0.02696996,0.07530449,0.807731],"study_design_scores_gemma":[0.0004266156,0.0009205572,0.004141358,0.0001881801,0.00014113,0.001656091,0.0004905959,0.6293814,0.1950389,0.04762153,0.119634,0.0003596746],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.05009426,0.003003741,0.8588315,0.0007250762,0.0003167331,0.0003955103,0.003554862,0.07214289,0.01093545],"genre_scores_gemma":[0.4174404,0.0008783995,0.5579243,0.001238479,0.0002155494,0.000620263,0.006552642,0.002553058,0.01257689],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005982,"threshold_uncertainty_score":0.02001184,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399174291","doi":"10.1145/3654963","title":"OTClean: Data Cleaning for Conditional Independence Violations using Optimal Transport","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"Universitas Brawijaya","keywords":"Computer science; Scalability; Mathematical optimization; Independence (probability theory); Mathematics","authors":[{"name":"Alireza Pirhadi","is_ca":true},{"name":"Mohammad Hossein Moslemi","is_ca":true},{"name":"Alexander Cloninger","is_ca":false},{"name":"Mostafa Milani","is_ca":true},{"name":"Babak Salimi","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1243438065986618,"gpt":0.3641923593716973,"spread":0.2398485527730355,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009386571,0.001605415,0.002095251,0.001930419,0.001767949,0.002945765,0.003806895,0.002607432,0.002827798],"category_scores_gemma":[0.0320976,0.001065605,0.002668152,0.002563363,0.003632285,0.005997322,0.007385392,0.005768319,0.0011272],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001953521,"about_ca_system_score_gemma":0.00621663,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005121819,"about_ca_topic_score_gemma":0.004157647,"domain_scores_codex":[0.9936463,0.002311984,0.0004228577,0.001250626,0.001959227,0.0004090894],"domain_scores_gemma":[0.9832876,0.008246134,0.001619218,0.004411487,0.002053997,0.0003816006],"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":[0.000361282,0.0001622498,0.002603657,0.0004787413,0.0002278738,0.0003282394,0.0005967455,0.6314083,0.01255909,0.09293812,0.009314306,0.2490214],"study_design_scores_gemma":[0.00002374032,0.00008232993,0.0002940556,0.00003725482,0.00002075147,0.0001242762,0.0001145595,0.9252272,0.00764299,0.06267842,0.003722261,0.00003199781],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00174141,0.0000663337,0.997221,0.0001358009,0.00002032716,0.00003859871,0.00005106481,0.000540071,0.0001854129],"genre_scores_gemma":[0.1593981,0.0002945233,0.8351722,0.0004357428,0.0001231696,0.0004202596,0.0009657904,0.000968841,0.002221215],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009386571,"threshold_uncertainty_score":0.04964149,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4389609557","doi":"10.1145/3626761","title":"DProvDB: Differentially Private Query Processing with Multi-Analyst Provenance","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Differential privacy; Computer science; Privilege (computing); Private information retrieval; Computer security; Database; Data mining","authors":[{"name":"Shufan Zhang","is_ca":true},{"name":"Xi He","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08080241617828675,"gpt":0.3053888409634237,"spread":0.224586424785137,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01572553,0.0009149397,0.001654726,0.001484508,0.001563521,0.004498536,0.004376586,0.002218149,0.002402892],"category_scores_gemma":[0.03274978,0.0008227591,0.001400884,0.002460568,0.002702323,0.00926779,0.009146815,0.00369499,0.0008822423],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002309648,"about_ca_system_score_gemma":0.004570264,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003670029,"about_ca_topic_score_gemma":0.004235386,"domain_scores_codex":[0.9872313,0.003905172,0.001071293,0.002112309,0.004831612,0.0008482149],"domain_scores_gemma":[0.9770702,0.008414316,0.001446342,0.009903898,0.002454264,0.0007108875],"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":[0.002408536,0.0005907496,0.007570982,0.0006984157,0.0003044351,0.0005729696,0.001612886,0.2273884,0.02933992,0.2948148,0.02850604,0.4061919],"study_design_scores_gemma":[0.0001682271,0.0002300862,0.0005085609,0.00004846725,0.00005097892,0.0003849877,0.0001938365,0.7626396,0.01561363,0.2056309,0.01446918,0.00006151495],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00715974,0.0003735075,0.9871787,0.0007026942,0.00006657071,0.0001995403,0.000392484,0.00296502,0.000961811],"genre_scores_gemma":[0.3252953,0.0004432142,0.6686343,0.0006067911,0.0001629997,0.0002993163,0.001391191,0.0005616874,0.002605266],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01572553,"threshold_uncertainty_score":0.08316559,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4389609606","doi":"10.1145/3626760","title":"Waffle: An Online Oblivious Datastore for Protecting Data Access Patterns","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Cryptography and Data Security","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Universitas Brawijaya","keywords":"Computer science; Overhead (engineering); Benchmark (surveying); Adversary; Flexibility (engineering); Data access; Bandwidth (computing); Database; Distributed computing; Computer network; Operating system; Computer security","authors":[{"name":"Sujaya Maiyya","is_ca":true},{"name":"Sharath Chandra Vemula","is_ca":false},{"name":"Divyakant Agrawal","is_ca":false},{"name":"Amr El Abbadi","is_ca":false},{"name":"Florian Kerschbaum","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3094247226058124,"gpt":0.395503019242454,"spread":0.08607829663664163,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002099393,0.001054326,0.0009131322,0.001242455,0.001413346,0.002104787,0.003939991,0.00135641,0.006629584],"category_scores_gemma":[0.008618623,0.0007704382,0.0008299383,0.001122073,0.003038401,0.009715975,0.007185211,0.002717954,0.002315535],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001329074,"about_ca_system_score_gemma":0.002121273,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001446023,"about_ca_topic_score_gemma":0.001856159,"domain_scores_codex":[0.9972838,0.0004650664,0.0002488547,0.000451465,0.001115141,0.0004356779],"domain_scores_gemma":[0.9868633,0.001826457,0.0006704905,0.009627723,0.0007228612,0.0002892649],"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":[0.003202374,0.0008259228,0.01131547,0.001405177,0.0006476273,0.001200116,0.001862883,0.1018181,0.1090292,0.1501264,0.08296312,0.5356038],"study_design_scores_gemma":[0.0003578041,0.0008276007,0.001882299,0.0002111083,0.0001800679,0.001491343,0.0003827055,0.6042026,0.1701308,0.1362415,0.08384022,0.0002519934],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1352917,0.00283763,0.7953869,0.001771905,0.000412783,0.0008176987,0.001843516,0.04365508,0.0179828],"genre_scores_gemma":[0.8468665,0.0007387288,0.133334,0.0007397102,0.0001400125,0.0005870848,0.001523947,0.001843134,0.01422692],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006629584,"threshold_uncertainty_score":0.02217817,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4381329253","doi":"10.1145/3589298","title":"Computing the Difference of Conjunctive Queries Efficiently","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Rewriting; Computer science; Query optimization; Heuristics; Operator (biology); Spatial query; Benchmark (surveying); Boolean conjunctive query; Set (abstract data type); Class (philosophy); Conjunctive query; SQL; Theoretical computer science; Algorithm; Sargable; Relational database; Search engine; Database; Information retrieval; Web search query","authors":[{"name":"Xiao Hu","is_ca":true},{"name":"Qichen Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06453656104710064,"gpt":0.3010111735680607,"spread":0.23647461252096,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003904338,0.001499114,0.00254915,0.001879831,0.001074226,0.004431486,0.004809613,0.001095627,0.006335379],"category_scores_gemma":[0.01875646,0.001285256,0.002298489,0.003207238,0.001956907,0.01254608,0.004223195,0.003002884,0.001743643],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002686167,"about_ca_system_score_gemma":0.002971568,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007076325,"about_ca_topic_score_gemma":0.00864323,"domain_scores_codex":[0.9888194,0.001160105,0.001044368,0.002558505,0.005366215,0.001051363],"domain_scores_gemma":[0.9839484,0.009363238,0.0008428504,0.00349149,0.002043106,0.0003109139],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.003098646,0.001108451,0.01603581,0.001701703,0.0008155484,0.001097064,0.001431762,0.1371795,0.0947395,0.1280542,0.02071015,0.5940276],"study_design_scores_gemma":[0.0002866439,0.0004769873,0.002261965,0.00004086325,0.0002381241,0.0006232569,0.0004665975,0.7633927,0.05571991,0.168871,0.007519176,0.0001027857],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1787084,0.0009691645,0.7968478,0.00113762,0.0002396522,0.0003942574,0.001575864,0.013154,0.006973244],"genre_scores_gemma":[0.4765908,0.0002503703,0.5156232,0.0005904895,0.0001587988,0.0001527021,0.00338564,0.001177794,0.002070216],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007076325,"threshold_uncertainty_score":0.02119392,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4381329270","doi":"10.1145/3589322","title":"Maestro: Automatic Generation of Comprehensive Benchmarks for Question Answering Over Knowledge Graphs","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"","keywords":"Computer science; Benchmark (surveying); Question answering; Knowledge base; Vocabulary; Natural language; Information retrieval; Set (abstract data type); Natural language understanding; Knowledge graph; Usability; Artificial intelligence; Natural language processing; Programming language; Human–computer interaction; Linguistics","authors":[{"name":"Abdelghny Orogat","is_ca":true},{"name":"Ahmed El-Roby","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1019414078019133,"gpt":0.3440120252621289,"spread":0.2420706174602156,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004043825,0.002868759,0.0009871292,0.005460505,0.000757851,0.002009164,0.003242787,0.001710442,0.009022462],"category_scores_gemma":[0.03305459,0.000874062,0.001540313,0.002840475,0.0007950977,0.003001204,0.00316305,0.00178496,0.004532638],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001360081,"about_ca_system_score_gemma":0.001837604,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004890708,"about_ca_topic_score_gemma":0.006124322,"domain_scores_codex":[0.9955574,0.0016311,0.0004574482,0.001028714,0.00107599,0.0002494612],"domain_scores_gemma":[0.98413,0.009549213,0.0007482918,0.001918249,0.003216448,0.0004378261],"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":[0.001331067,0.001064848,0.009405471,0.004960965,0.0004104489,0.001345342,0.0025553,0.05863292,0.04068023,0.02131938,0.3017087,0.5565853],"study_design_scores_gemma":[0.0007349434,0.0007182924,0.008176192,0.0004824784,0.0001674592,0.0007170358,0.001384482,0.6931043,0.06720524,0.03845834,0.1886228,0.0002284457],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08346705,0.002578621,0.4931529,0.001152231,0.0008159296,0.002927177,0.06082034,0.3354835,0.01960238],"genre_scores_gemma":[0.1786785,0.0007028808,0.5897623,0.0004868461,0.0001009741,0.003137813,0.2076795,0.01448213,0.004969027],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009022462,"threshold_uncertainty_score":0.0301832,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4407356147","doi":"10.1145/3709719","title":"Reliable Text-to-SQL with Adaptive Abstention","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Distributed and Parallel Computing Systems","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University; University of Toronto","funders":"","keywords":"Computer science; SQL; Programming language; Database","authors":[{"name":"Kaiwen Chen","is_ca":true},{"name":"Yueting Chen","is_ca":false},{"name":"Nick Koudas","is_ca":true},{"name":"Xiaohui Yu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03721397291859412,"gpt":0.2741668219245275,"spread":0.2369528490059334,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007771262,0.001265395,0.0009940852,0.0007792821,0.0005581075,0.002339134,0.004209791,0.001734918,0.003401929],"category_scores_gemma":[0.04146148,0.0007673644,0.0009996502,0.0006618293,0.001901956,0.00630832,0.00526101,0.002695811,0.001779571],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009060062,"about_ca_system_score_gemma":0.001545325,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002212956,"about_ca_topic_score_gemma":0.002502219,"domain_scores_codex":[0.9917423,0.003296509,0.0005446021,0.001386677,0.002748844,0.0002812516],"domain_scores_gemma":[0.9718204,0.01443824,0.001441899,0.009617512,0.002227117,0.0004548638],"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":[0.00268958,0.0008219838,0.01060277,0.0005317362,0.0001887054,0.0007908033,0.001968712,0.3637539,0.06737945,0.05882824,0.0196393,0.4728049],"study_design_scores_gemma":[0.00005226412,0.000112863,0.0002360229,0.00000934358,0.00001492883,0.00008006486,0.0000422573,0.9584796,0.01169343,0.02653138,0.002721254,0.00002650214],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03966335,0.0001518277,0.9179853,0.0004808357,0.00004185028,0.0001809628,0.0002729658,0.03987299,0.001349997],"genre_scores_gemma":[0.6025104,0.0001184928,0.3892972,0.0005299482,0.00007492406,0.00030052,0.000950656,0.00318349,0.003034322],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007771262,"threshold_uncertainty_score":0.04109889,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4380433188","doi":"10.1145/3588711","title":"dbET: Execution Time Distribution-based Plan Selection","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Huawei Technologies (Canada); University of Toronto; York University","funders":"","keywords":"Computer science; Query plan; Plan (archaeology); Complement (music); Selection (genetic algorithm); Execution time; Overhead (engineering); Query optimization; Database; Sargable; Distributed computing; Programming language; Information retrieval; Web search query; Search engine; Artificial intelligence","authors":[{"name":"Yifan Li","is_ca":true},{"name":"Xiaohui Yu","is_ca":true},{"name":"Nick Koudas","is_ca":true},{"name":"Shu Lin","is_ca":true},{"name":"Calvin Sun","is_ca":true},{"name":"Chong Chen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04636754490894221,"gpt":0.2778059725457491,"spread":0.2314384276368069,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002191425,0.0009194135,0.0007175332,0.001338912,0.000364969,0.001218257,0.001518702,0.0005092492,0.00250589],"category_scores_gemma":[0.007232832,0.0003609835,0.0006459228,0.001309589,0.0005338353,0.001282688,0.001026577,0.000994251,0.000575365],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001070235,"about_ca_system_score_gemma":0.001936909,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003583151,"about_ca_topic_score_gemma":0.004992072,"domain_scores_codex":[0.9979256,0.000447716,0.0001201095,0.0003379941,0.001014279,0.0001542181],"domain_scores_gemma":[0.9966056,0.001815403,0.000288767,0.0005627452,0.0005788096,0.0001485873],"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":[0.0007843677,0.0003600623,0.01017004,0.0003150172,0.0001689978,0.0001762109,0.0001820584,0.3799504,0.02244128,0.02387184,0.01845036,0.5431294],"study_design_scores_gemma":[0.00005490303,0.00009687094,0.0009474748,0.000009824093,0.00002137434,0.00008734859,0.00003084069,0.9794613,0.007938948,0.007705579,0.003625116,0.00002045042],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02333282,0.000333321,0.9657782,0.0001843635,0.00004250998,0.0002542114,0.00079448,0.006588399,0.002691692],"genre_scores_gemma":[0.4175302,0.0002852889,0.5763616,0.0001757291,0.0000665745,0.0003549347,0.002413454,0.0009348546,0.001877265],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003583151,"threshold_uncertainty_score":0.01158953,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4381326963","doi":"10.1145/3589290","title":"OM3: An Ordered Multi-level Min-Max Representation for Interactive Progressive Visualization of Time Series","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Time Series Analysis and Forecasting","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China; National Science Foundation","keywords":"Computer science; Zoom; Visualization; Representation (politics); Ranging; Latency (audio); Series (stratigraphy); Data mining; Computer graphics (images)","authors":[{"name":"Yunhai Wang","is_ca":false},{"name":"Y. L. Wang","is_ca":false},{"name":"Xin Chen","is_ca":false},{"name":"Yue Zhao","is_ca":false},{"name":"Fan Zhang","is_ca":false},{"name":"Eugene Wu","is_ca":false},{"name":"Chi‐Wing Fu","is_ca":false},{"name":"Xiaohui Yu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1413222286035876,"gpt":0.3732153580399873,"spread":0.2318931294363998,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006623375,0.001212852,0.0007739375,0.001754242,0.0003527682,0.002098498,0.001355102,0.000602898,0.01122975],"category_scores_gemma":[0.002988804,0.0004406525,0.001091388,0.001655052,0.0002865764,0.002314164,0.00171095,0.001051947,0.001761985],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004726776,"about_ca_system_score_gemma":0.0007319992,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002750817,"about_ca_topic_score_gemma":0.003887535,"domain_scores_codex":[0.9996903,0.00006512357,0.00003524597,0.0000603471,0.000103693,0.0000452597],"domain_scores_gemma":[0.9993586,0.0002601797,0.00008371161,0.0001367713,0.0001087515,0.00005189639],"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":[0.001551027,0.0002292115,0.003586286,0.001048772,0.0001977533,0.0004979431,0.001158976,0.1192138,0.06841015,0.05356545,0.06439162,0.686149],"study_design_scores_gemma":[0.00009692129,0.0001283124,0.001388544,0.00007594651,0.00004735347,0.0002095549,0.0001725948,0.8962713,0.0243526,0.02922616,0.04794028,0.00009042401],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.009280868,0.0002284947,0.971929,0.0001884757,0.00006920367,0.00007132625,0.002085598,0.01445648,0.001690497],"genre_scores_gemma":[0.1313417,0.0004536077,0.8590795,0.000145208,0.00007433854,0.0003629488,0.004351815,0.002159966,0.002030966],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01122975,"threshold_uncertainty_score":0.03756726,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4393183687","doi":"10.1145/3639279","title":"DTT: An Example-Driven Tabular Transformer for Joinability by Leveraging Large Language Models","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Topic Modeling","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Transformer; Computer science; Engineering; Electrical engineering; Voltage","authors":[{"name":"Arash Nobari","is_ca":true},{"name":"Davood Rafiei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1034483500304466,"gpt":0.3135535429041299,"spread":0.2101051928736833,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001350108,0.00101697,0.0007322612,0.001430152,0.0005445267,0.001773986,0.002506195,0.0009190601,0.01073176],"category_scores_gemma":[0.006887335,0.0005499555,0.001753532,0.001671582,0.000907535,0.006842368,0.002464308,0.00242467,0.005227981],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001117516,"about_ca_system_score_gemma":0.002265301,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006429025,"about_ca_topic_score_gemma":0.01094178,"domain_scores_codex":[0.9990103,0.000188743,0.00008800172,0.0002983102,0.0003332073,0.00008137789],"domain_scores_gemma":[0.9977513,0.0008313867,0.0001592505,0.0008287902,0.0003334803,0.00009578584],"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":[0.0005623537,0.0004250442,0.004315146,0.0004880039,0.0001202508,0.0003965413,0.0005003671,0.1052744,0.01323588,0.05652421,0.06684957,0.7513082],"study_design_scores_gemma":[0.00005408286,0.00008074307,0.0002798055,0.00003499546,0.0000325621,0.0001611911,0.00009108228,0.9183519,0.01312005,0.04855638,0.01920556,0.00003163437],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01483866,0.0002956834,0.9390623,0.0003892163,0.0001514832,0.0001771698,0.003214503,0.0384702,0.003400696],"genre_scores_gemma":[0.2596592,0.0004729463,0.7078319,0.0006330098,0.0001268195,0.0003470774,0.01802761,0.003564385,0.009337069],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01073176,"threshold_uncertainty_score":0.03590131,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"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,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"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","authors":[{"name":"Binyang Dai","is_ca":false},{"name":"Xiao Hu","is_ca":true},{"name":"Ke Yi","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.12639640691388,"gpt":0.35644566552144,"spread":0.23004925860756,"validation_status":"score_only:v0-immature-baseline"},"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"},"labels":[],"label_agreement":null},{"id":"W4396892532","doi":"10.1145/3651599","title":"Fast Matrix Multiplication for Query Processing","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Distributed and Parallel Computing Systems","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Matrix multiplication; Multiplication (music); Query optimization; Matrix (chemical analysis); Arithmetic; Information retrieval; Mathematics; Physics; Chemistry; Combinatorics","authors":[{"name":"Xiao Hu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0612702140294622,"gpt":0.3339377198722386,"spread":0.2726675058427764,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004352161,0.002444409,0.001840726,0.001907639,0.002124946,0.004287455,0.002957052,0.001491289,0.01304338],"category_scores_gemma":[0.01709468,0.0009869848,0.001451163,0.004016181,0.001966051,0.01215508,0.004060097,0.003769915,0.00637396],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002221589,"about_ca_system_score_gemma":0.002078311,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003184962,"about_ca_topic_score_gemma":0.002792108,"domain_scores_codex":[0.9915015,0.0019331,0.000548203,0.00130859,0.003737409,0.0009713545],"domain_scores_gemma":[0.9858628,0.007652321,0.0005144928,0.003369038,0.002337236,0.0002641127],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001836926,0.0003311536,0.00139671,0.0009424587,0.0001603966,0.0002718156,0.0006539543,0.06656437,0.03258683,0.3281526,0.04247215,0.5246307],"study_design_scores_gemma":[0.0002099478,0.0004593997,0.0003547284,0.0001158539,0.00009552384,0.0005269857,0.0001912516,0.5941912,0.03524163,0.3184463,0.05006021,0.0001069459],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007047295,0.001233545,0.9819518,0.0005297004,0.0002662956,0.0001383472,0.0001826943,0.003597392,0.005052878],"genre_scores_gemma":[0.1788082,0.001390302,0.8097851,0.0005218305,0.0006168006,0.0004367912,0.0008355433,0.00115972,0.006445722],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01304338,"threshold_uncertainty_score":0.04363441,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4404130423","doi":"10.1145/3695835","title":"Computing A Well-Representative Summary of Conjunctive Query Results","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Conjunctive query; Boolean conjunctive query; Computer science; Query optimization; Query expansion; Information retrieval; Theoretical computer science; Web search query; Sargable; Data mining; Search engine; Relational database","authors":[{"name":"Pankaj K. Agarwal","is_ca":false},{"name":"Aryan Esmailpour","is_ca":false},{"name":"Xiao Hu","is_ca":true},{"name":"Stavros Sintos","is_ca":false},{"name":"Jun Yang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05242105394825337,"gpt":0.3208429987592012,"spread":0.2684219448109479,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002503074,0.001533826,0.00285996,0.002158416,0.0008257327,0.003653101,0.002329829,0.001562423,0.002003384],"category_scores_gemma":[0.01737966,0.0007958171,0.001160312,0.003511764,0.0008313067,0.006377035,0.001842354,0.001380209,0.001163051],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001243963,"about_ca_system_score_gemma":0.001654913,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002199498,"about_ca_topic_score_gemma":0.003327274,"domain_scores_codex":[0.9965177,0.0004423555,0.0005561039,0.0009240954,0.001264963,0.0002948486],"domain_scores_gemma":[0.9915677,0.003514435,0.001140783,0.001797422,0.001612746,0.0003669614],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.002234249,0.0005151244,0.01360348,0.001604894,0.0005805501,0.0008289337,0.001732122,0.3488257,0.06622405,0.02509601,0.01531151,0.5234433],"study_design_scores_gemma":[0.00005956554,0.0003294642,0.002284706,0.00003722071,0.0001352305,0.0003863691,0.000454584,0.9437246,0.02398044,0.02468061,0.003879738,0.00004740794],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1202329,0.001285593,0.8692543,0.0008227641,0.00009340795,0.0002964107,0.002009414,0.004583778,0.001421393],"genre_scores_gemma":[0.4021399,0.0005813177,0.5896192,0.0002075053,0.0001687549,0.0002363161,0.005477588,0.0003584222,0.00121098],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003653101,"threshold_uncertainty_score":0.01323766,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411403538","doi":"10.1145/3725325","title":"MIRAGE-ANNS: Mixed Approach Graph-based Indexing for Approximate Nearest Neighbor Search","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Image and Video Retrieval Techniques","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Search engine indexing; Computer science; Graph; Nearest neighbor search; k-nearest neighbors algorithm; Data mining; Search engine; Context (archaeology); Search algorithm; Theoretical computer science; Machine learning; Artificial intelligence; Algorithm; Information retrieval","authors":[{"name":"Sairaj Voruganti","is_ca":true},{"name":"M. TAMER ÖZSU","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08133916376663859,"gpt":0.3436731363883481,"spread":0.2623339726217095,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009713725,0.001177634,0.001958652,0.003383806,0.001074827,0.001842184,0.002951786,0.001380574,0.005559051],"category_scores_gemma":[0.007208297,0.0005282782,0.00106915,0.004963316,0.000690302,0.004272624,0.003054677,0.001312413,0.003408729],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001166225,"about_ca_system_score_gemma":0.001904601,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01162437,"about_ca_topic_score_gemma":0.02488246,"domain_scores_codex":[0.9980995,0.0003727742,0.000170405,0.0003224144,0.0009103546,0.0001246376],"domain_scores_gemma":[0.9975682,0.0006891334,0.0001596941,0.0009910166,0.0004862581,0.0001056583],"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":[0.0007046199,0.0005144769,0.002821071,0.0005419258,0.0002130772,0.0001948666,0.0002712469,0.08542538,0.01225173,0.02590708,0.0562634,0.8148911],"study_design_scores_gemma":[0.00009499804,0.0002647561,0.0005505135,0.00003105924,0.00004486342,0.0002800091,0.0001059149,0.9449919,0.006976833,0.02442574,0.02217705,0.00005634715],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02991495,0.003280116,0.9317886,0.0004552763,0.0003724673,0.0005306472,0.002818977,0.0219753,0.008863769],"genre_scores_gemma":[0.141462,0.0009515744,0.8417385,0.0003277295,0.0001727426,0.0003659627,0.008109569,0.0008215054,0.00605058],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01162437,"threshold_uncertainty_score":0.02311343,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399156413","doi":"10.1145/3654985","title":"Wii: Dynamic Budget Reallocation In Index Tuning","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Workload; Spurious relationship; Index (typography); Set (abstract data type); Process (computing); Mathematical optimization; Budget constraint; Resource allocation; Constraint (computer-aided design); Operations research; Operating system; Mathematics; Economics","authors":[{"name":"Xiaoying Wang","is_ca":true},{"name":"Wentao Wu","is_ca":false},{"name":"Chi Wang","is_ca":false},{"name":"Vivek Narasayya","is_ca":false},{"name":"Surajit Chaudhuri","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0325549985898155,"gpt":0.3031739514041989,"spread":0.2706189528143834,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003833732,0.002055915,0.001597966,0.001744279,0.001343785,0.003018651,0.00578077,0.001749914,0.006150592],"category_scores_gemma":[0.01394574,0.001213375,0.0007202094,0.002130751,0.001431858,0.005488981,0.004830284,0.002770213,0.002509863],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001764449,"about_ca_system_score_gemma":0.003443395,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0041381,"about_ca_topic_score_gemma":0.005006796,"domain_scores_codex":[0.9960472,0.000862374,0.0002886525,0.0007937863,0.001295572,0.0007124543],"domain_scores_gemma":[0.9949787,0.002000418,0.0005429355,0.001805924,0.0003574446,0.0003145262],"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":[0.001366968,0.000721878,0.005859275,0.000658042,0.0001825604,0.0002672613,0.0007492848,0.2060806,0.04762916,0.03774072,0.04254052,0.6562037],"study_design_scores_gemma":[0.0001629955,0.0001813528,0.0009731894,0.00006143509,0.00005368632,0.0001660452,0.0001122084,0.9469015,0.01715155,0.02087984,0.01327244,0.00008381068],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03057851,0.001696323,0.9256335,0.0006408496,0.0002107261,0.000445145,0.0004621736,0.03269375,0.007639083],"genre_scores_gemma":[0.3796559,0.0006171966,0.6092455,0.0005121783,0.0001631668,0.0007158569,0.001158898,0.003240855,0.004690495],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006150592,"threshold_uncertainty_score":0.02057582,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4389609976","doi":"10.1145/3626739","title":"NOCAP: Near-Optimal Correlation-Aware Partitioning Joins","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Caching and Content Delivery","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Universitas Brawijaya","keywords":"Joins; Computer science; Join (topology); Skew; Exploit; Hash function; Key (lock); Relation (database); Parallel computing; Hash join; Theoretical computer science; Algorithm; Data mining; Mathematics","authors":[{"name":"Zichen Zhu","is_ca":false},{"name":"Xiao Hu","is_ca":true},{"name":"Manos Athanassoulis","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07309228208787909,"gpt":0.2823106997284716,"spread":0.2092184176405925,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001297433,0.0008176316,0.0008789172,0.0007492944,0.001015685,0.001233833,0.001985692,0.0005827884,0.001714394],"category_scores_gemma":[0.005264302,0.0004687443,0.0004466605,0.001355937,0.0007132058,0.002176806,0.002157996,0.0007667255,0.0006394842],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001074167,"about_ca_system_score_gemma":0.002014847,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002048284,"about_ca_topic_score_gemma":0.004050997,"domain_scores_codex":[0.9981171,0.0003646137,0.000121041,0.0003030495,0.0008806339,0.0002136843],"domain_scores_gemma":[0.9972907,0.0009722587,0.0002731678,0.0008186423,0.0005108867,0.0001342897],"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":[0.001223096,0.0002778305,0.004221012,0.0003509229,0.0001018919,0.00023581,0.0004241857,0.4498422,0.03765874,0.05193076,0.02080492,0.4329285],"study_design_scores_gemma":[0.00004510286,0.0001162957,0.0003421079,0.00001473443,0.00001833166,0.0002084517,0.00008381171,0.9606369,0.0102449,0.02313298,0.005138298,0.00001812027],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06633905,0.001081845,0.9232761,0.0002432608,0.0001212318,0.0001434646,0.0002317254,0.002447382,0.006116036],"genre_scores_gemma":[0.4477042,0.0003614108,0.5474527,0.000198511,0.00007007952,0.0001633662,0.0008482967,0.0003722477,0.002829134],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002048284,"threshold_uncertainty_score":0.007793665,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399156410","doi":"10.1145/3654934","title":"Data Acquisition for Improving Model Confidence","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; York University","funders":"","keywords":"Computer science; Data acquisition; Machine learning; Context (archaeology); Knowledge acquisition; Range (aeronautics); Process (computing); Artificial intelligence; Data mining; Data quality; Data science; Engineering","authors":[{"name":"Yifan Li","is_ca":true},{"name":"Xiaohui Yu","is_ca":true},{"name":"Nick Koudas","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1126036958320424,"gpt":0.3468254042837654,"spread":0.234221708451723,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01375825,0.002195951,0.002893644,0.003292915,0.001276069,0.003068318,0.004153091,0.002334895,0.003508053],"category_scores_gemma":[0.09475756,0.001248984,0.00176109,0.002821975,0.001729842,0.008389064,0.006667011,0.005406511,0.001921993],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0011821,"about_ca_system_score_gemma":0.00405518,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003463033,"about_ca_topic_score_gemma":0.005274174,"domain_scores_codex":[0.9912096,0.002772268,0.0009594766,0.001963483,0.002614927,0.0004802168],"domain_scores_gemma":[0.9360967,0.03444773,0.004057187,0.01517976,0.009325502,0.0008930379],"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":[0.00161165,0.001265085,0.02763243,0.0008276705,0.0004181537,0.0003846941,0.001237165,0.1730256,0.0189517,0.02035347,0.01468798,0.7396045],"study_design_scores_gemma":[0.0001074617,0.0003580607,0.002721831,0.00009817799,0.00007984424,0.0002477498,0.0002827177,0.9482003,0.01522121,0.0268307,0.005794269,0.00005767327],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04688839,0.0008294734,0.9459043,0.0007758488,0.00007480355,0.0002772324,0.0006579277,0.003196,0.001396068],"genre_scores_gemma":[0.4180884,0.0004560025,0.5730137,0.0008035888,0.0001673211,0.0006793083,0.004735751,0.0007837301,0.001272218],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01375825,"threshold_uncertainty_score":0.07276142,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4381328483","doi":"10.1145/3589320","title":"Mitigating Filter Bubbles Under a Competitive Diffusion Model","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Opinion Dynamics and Social Influence","field":"Physics and Astronomy","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Heuristic; Computer science; Filter (signal processing); Viewpoints; Mathematical optimization; Maximization; Algorithm; Artificial intelligence; Mathematics","authors":[{"name":"Prithu Banerjee","is_ca":true},{"name":"Wei Chen","is_ca":false},{"name":"Laks V. S. Lakshmanan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06327816623333686,"gpt":0.3162685902262851,"spread":0.2529904239929482,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003613269,0.00133164,0.002064356,0.001111268,0.0009830291,0.002379366,0.003093045,0.003667214,0.00300664],"category_scores_gemma":[0.01378578,0.0005933263,0.001040002,0.001348522,0.00131587,0.004441551,0.00156883,0.002406441,0.0005481372],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002887631,"about_ca_system_score_gemma":0.001915446,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01263497,"about_ca_topic_score_gemma":0.009763266,"domain_scores_codex":[0.9984488,0.0006045018,0.00006425171,0.0003378087,0.0002757447,0.000268711],"domain_scores_gemma":[0.9917474,0.006379397,0.0006748287,0.0003495383,0.0005263512,0.000322529],"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":[0.000190497,0.0001660049,0.001956099,0.0001953721,0.00006113563,0.0001723949,0.000205656,0.8893878,0.001465291,0.06381093,0.006845329,0.03554351],"study_design_scores_gemma":[0.00001692779,0.0000153624,0.00007090015,0.000003911781,0.000006964724,0.00001183338,0.00001258587,0.9901388,0.0001419323,0.009149848,0.0004272622,0.000003801287],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0806342,0.001369611,0.9067876,0.003166252,0.0001043819,0.0002425546,0.0004934085,0.0006552261,0.006546792],"genre_scores_gemma":[0.8140827,0.001048494,0.1760568,0.0006751773,0.0002781812,0.0004011712,0.0007124179,0.0001611807,0.006583833],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01263497,"threshold_uncertainty_score":0.02512282,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4389611627","doi":"10.1145/3626750","title":"Rethink Query Optimization in HTAP Databases","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Computer science; Online analytical processing; Online transaction processing; Database; Schedule; Isolation (microbiology); Parallel computing; Database transaction; Transaction processing; Data warehouse; Operating system","authors":[{"name":"Haoze Song","is_ca":false},{"name":"Wenchao Zhou","is_ca":false},{"name":"Feifei Li","is_ca":false},{"name":"X. Peng","is_ca":false},{"name":"Heming Cui","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09719269435293139,"gpt":0.3220787802633234,"spread":0.224886085910392,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001903186,0.001018267,0.0005647843,0.0004420497,0.0006665493,0.002014514,0.00188619,0.0006810611,0.0008921293],"category_scores_gemma":[0.00313869,0.0005935167,0.0007143142,0.0009996183,0.001184802,0.002845692,0.001481822,0.001609954,0.0002895877],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00109433,"about_ca_system_score_gemma":0.001637753,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007183782,"about_ca_topic_score_gemma":0.009549847,"domain_scores_codex":[0.9977642,0.0005690016,0.0001711701,0.0003724082,0.0008748528,0.0002483763],"domain_scores_gemma":[0.9982565,0.0007570561,0.0001485207,0.000515947,0.000241733,0.00008024688],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001603151,0.0005429073,0.01616252,0.0004601394,0.0003444317,0.0005720763,0.001305523,0.4459093,0.1500756,0.02453426,0.01126261,0.3472274],"study_design_scores_gemma":[0.00008435783,0.0002820017,0.0026703,0.0000189825,0.00008517593,0.0001797192,0.0003435395,0.9275171,0.0505226,0.01153078,0.006723115,0.00004241042],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4720565,0.003075907,0.4992596,0.001585273,0.0001517183,0.000296648,0.0006037639,0.01379587,0.009174848],"genre_scores_gemma":[0.7739594,0.0004374286,0.2218969,0.0003682889,0.00005826372,0.0000875105,0.0004923144,0.0008221208,0.001877792],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007183782,"threshold_uncertainty_score":0.01428396,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4396893056","doi":"10.1145/3651603","title":"On the Feasibility of Forgetting in Data Streams","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":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Universitas Brawijaya","keywords":"STREAMS; Forgetting; Computer science; Psychology; Cognitive psychology; Computer network","authors":[{"name":"A. Pavan","is_ca":false},{"name":"Sourav Chakraborty","is_ca":false},{"name":"N. V. Vinodchandran","is_ca":false},{"name":"Kuldeep S. Meel","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1701356361458202,"gpt":0.3650236530530431,"spread":0.1948880169072229,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01898933,0.001338923,0.002251856,0.001694084,0.002628179,0.005843553,0.004297584,0.00426761,0.006249095],"category_scores_gemma":[0.1990873,0.001453617,0.002192589,0.002321017,0.006900932,0.01920074,0.006686232,0.007770666,0.0007319573],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002334645,"about_ca_system_score_gemma":0.003275792,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00530368,"about_ca_topic_score_gemma":0.003296832,"domain_scores_codex":[0.9867564,0.005640956,0.0008933102,0.003025953,0.002524634,0.001158737],"domain_scores_gemma":[0.6630304,0.3007315,0.009230653,0.0161083,0.007624387,0.00327485],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001528887,0.0001826485,0.007595639,0.0004861897,0.0001243824,0.0008690245,0.001235961,0.2366104,0.001820066,0.694072,0.004920815,0.05055392],"study_design_scores_gemma":[0.00007572677,0.00008957333,0.0004546927,0.00008619647,0.00002845474,0.0002484314,0.0001598501,0.5434396,0.001024853,0.4525622,0.001787,0.00004359136],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09116527,0.00188664,0.883114,0.009870627,0.0003288548,0.0001971635,0.0007745716,0.0007067366,0.01195608],"genre_scores_gemma":[0.8604854,0.001693638,0.1286929,0.00138366,0.0008354569,0.0002903069,0.0009309074,0.000328189,0.005359473],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01898933,"threshold_uncertainty_score":0.1004264,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4396892922","doi":"10.1145/3651613","title":"A faster FPRAS for #NFA","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Machine Learning and Algorithms","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Universitas Brawijaya","keywords":"Computer science","authors":[{"name":"Kuldeep S. Meel","is_ca":true},{"name":"Sourav Chakraborty","is_ca":false},{"name":"Umang Mathur","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05590839665298635,"gpt":0.3236092414076727,"spread":0.2677008447546864,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00475499,0.001817694,0.002363025,0.00114747,0.001495342,0.003813324,0.004470988,0.003228385,0.008695833],"category_scores_gemma":[0.02341166,0.001183536,0.004366961,0.001541751,0.002535764,0.01068981,0.004179268,0.005557718,0.003308073],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004072524,"about_ca_system_score_gemma":0.005823125,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005625187,"about_ca_topic_score_gemma":0.005871987,"domain_scores_codex":[0.9906036,0.002252907,0.0006271108,0.002693163,0.002598427,0.001224782],"domain_scores_gemma":[0.9754719,0.01080831,0.001232892,0.01064884,0.00128836,0.0005498411],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.003399998,0.0009794749,0.005673829,0.001445047,0.0004606624,0.0003804668,0.001115238,0.1399591,0.05690054,0.26307,0.02673006,0.4998856],"study_design_scores_gemma":[0.0003419158,0.0003755117,0.000743587,0.0001450022,0.0002295195,0.0004878261,0.0002050307,0.7627742,0.02536229,0.1881727,0.02104154,0.0001209571],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04686186,0.000915658,0.9237599,0.002667821,0.0002629349,0.0003441986,0.0007330354,0.01836523,0.006089339],"genre_scores_gemma":[0.2540558,0.0002965858,0.7367905,0.00117706,0.0001579661,0.0005562076,0.00117891,0.001000773,0.004786175],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008695833,"threshold_uncertainty_score":0.02954835,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141691","doi":"10.1145/3725235","title":"Fast Matrix Multiplication meets the Submodular Width","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Complexity and Algorithms in Graphs","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Universitas Brawijaya","keywords":"Matrix multiplication; Submodular set function; Multiplication (music); Matrix (chemical analysis); Computer science; Mathematics; Arithmetic; Combinatorics; Physics; Materials science","authors":[{"name":"Mahmoud Abo Khamis","is_ca":false},{"name":"Dan Suciu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0438812084694511,"gpt":0.3106361800602223,"spread":0.2667549715907712,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0055987,0.002120182,0.003192565,0.001435435,0.001556068,0.007350797,0.003315025,0.002849023,0.006124865],"category_scores_gemma":[0.03382564,0.001275419,0.002933536,0.003613947,0.003633638,0.02727495,0.004628514,0.006633448,0.00177301],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00273529,"about_ca_system_score_gemma":0.003355478,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002435267,"about_ca_topic_score_gemma":0.001981932,"domain_scores_codex":[0.9912349,0.001926647,0.000701551,0.002892561,0.002035923,0.001208354],"domain_scores_gemma":[0.9617113,0.02605462,0.002088789,0.007482172,0.00175634,0.0009066683],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.002000887,0.0005816286,0.003798922,0.001423446,0.0002735129,0.0003237961,0.00106007,0.1167731,0.02232661,0.5965684,0.02249635,0.2323732],"study_design_scores_gemma":[0.00009803737,0.0002011851,0.0004680363,0.0000508993,0.00005919961,0.0003810377,0.0001640898,0.2046033,0.006303048,0.7832288,0.004398472,0.0000438426],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07979991,0.002171641,0.8963785,0.007230886,0.0002226268,0.0003413378,0.00158452,0.002007359,0.01026324],"genre_scores_gemma":[0.6055515,0.001624069,0.3818351,0.002188528,0.001055987,0.000709019,0.001420135,0.0006739312,0.004941721],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007350797,"threshold_uncertainty_score":0.02960908,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4417070120","doi":"10.1145/3769829","title":"ST-Raptor: LLM-Powered Semi-Structured Table Question Answering","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"National Key Research and Development Program of China; National Natural Science Foundation of China","keywords":"Table (database); Correctness; Tree (set theory); Set (abstract data type); Benchmark (surveying); Pipeline (software); Question answering","authors":[{"name":"Boyu Niu","is_ca":false},{"name":"Xuanhe Zhou","is_ca":false},{"name":"Wei Zhou","is_ca":false},{"name":"J. Wang","is_ca":true},{"name":"Guoliang Li","is_ca":false},{"name":"Xinyi Zhang","is_ca":false},{"name":"Fan Wu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1373690653317707,"gpt":0.4113977177010211,"spread":0.2740286523692504,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007122713,0.00170836,0.001209398,0.002025845,0.0009965374,0.003233228,0.004999233,0.002571185,0.01818994],"category_scores_gemma":[0.03604563,0.001182408,0.002598759,0.001270633,0.002242547,0.008227798,0.007029453,0.003278474,0.01059901],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001343792,"about_ca_system_score_gemma":0.004356821,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005198085,"about_ca_topic_score_gemma":0.006700838,"domain_scores_codex":[0.9913845,0.003823354,0.0005900752,0.001429417,0.002350216,0.0004224913],"domain_scores_gemma":[0.9803193,0.0116582,0.0006564366,0.004245865,0.002703304,0.0004169844],"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":[0.001315583,0.0005915498,0.004868027,0.002285352,0.0002006015,0.001036256,0.005028586,0.04882607,0.04919832,0.1477364,0.1173388,0.6215744],"study_design_scores_gemma":[0.0001434509,0.0002185468,0.0006576998,0.0001336518,0.00005046182,0.0004660384,0.0004850698,0.8496262,0.03120858,0.06915355,0.04773639,0.0001204691],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005129872,0.0001312527,0.9337869,0.0004794906,0.00008264359,0.0005232813,0.001699521,0.05521103,0.002955927],"genre_scores_gemma":[0.09696149,0.0001442294,0.8896387,0.000618017,0.00007187905,0.0007174174,0.005927822,0.002885153,0.003035236],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01818994,"threshold_uncertainty_score":0.0608514,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411403532","doi":"10.1145/3725397","title":"Computing Inconsistency Measures Under Differential Privacy","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Universitas Brawijaya","keywords":"Differential privacy; Computer science; Tuple; Graph; Data mining; Data quality; Information sensitivity; Theoretical computer science; Mathematics; Computer security","authors":[{"name":"Shubhankar Mohapatra","is_ca":true},{"name":"Amir Gilad","is_ca":false},{"name":"Xi He","is_ca":true},{"name":"Benny Kimelfeld","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.075174269373825,"gpt":0.3138096824857975,"spread":0.2386354131119726,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01124608,0.001089466,0.001727223,0.003810137,0.0009724767,0.003498511,0.002840875,0.001780709,0.001235764],"category_scores_gemma":[0.1053837,0.0008085389,0.0014077,0.004564944,0.002836776,0.009065102,0.004607816,0.003442042,0.0002197802],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002927241,"about_ca_system_score_gemma":0.001776477,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001880871,"about_ca_topic_score_gemma":0.001379983,"domain_scores_codex":[0.9868668,0.004698911,0.001196025,0.002966286,0.003534422,0.0007375528],"domain_scores_gemma":[0.897612,0.07356503,0.009071289,0.0135626,0.004920806,0.001268373],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0006920431,0.0002252235,0.03729807,0.0004678714,0.0004514067,0.000451241,0.0007690486,0.6322364,0.005047184,0.1748155,0.003571696,0.1439743],"study_design_scores_gemma":[0.00003128777,0.000105796,0.002642564,0.00003870021,0.00004109241,0.000234414,0.0001592976,0.7395459,0.002653165,0.2535807,0.0009357052,0.00003127989],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1134857,0.0007033829,0.8818488,0.001015355,0.00004290956,0.00009738134,0.0007903306,0.0007338881,0.001282153],"genre_scores_gemma":[0.7753016,0.0004168809,0.2214714,0.0003280731,0.0001247116,0.0002136083,0.001464542,0.0001779287,0.0005011652],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01124608,"threshold_uncertainty_score":0.05947572,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4388620459","doi":"10.1145/3617336","title":"OptiQL: Robust Optimistic Locking for Memory-Optimized Indexes","year":2023,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Lock (firearm); Robustness (evolution); Multi-core processor; Parallel computing; Byte; Distributed computing; Mutual exclusion; Operating system; Computer network","authors":[{"name":"Ge Shi","is_ca":true},{"name":"Ziyi Yan","is_ca":true},{"name":"Tianzheng Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1022769341983313,"gpt":0.3075132284456033,"spread":0.205236294247272,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001867386,0.0005058966,0.0006107509,0.0007382544,0.0008424926,0.001924992,0.003716351,0.0005040687,0.003830154],"category_scores_gemma":[0.005626799,0.0005522447,0.0003916731,0.0008589427,0.001067556,0.002625042,0.002798971,0.001194805,0.00105289],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001055843,"about_ca_system_score_gemma":0.002360287,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002725313,"about_ca_topic_score_gemma":0.0039171,"domain_scores_codex":[0.9979016,0.0002373851,0.0001871433,0.0001983934,0.001140888,0.000334645],"domain_scores_gemma":[0.9968676,0.0008091504,0.0004283568,0.001151971,0.0004989945,0.000243934],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.003256638,0.0004863706,0.009904305,0.000789372,0.0002146327,0.0005273384,0.0007909556,0.1418533,0.1637477,0.08332118,0.03174678,0.5633614],"study_design_scores_gemma":[0.0005341651,0.0006915484,0.001310721,0.00005871931,0.00009649257,0.0005164099,0.000164531,0.7604949,0.156446,0.029659,0.04984042,0.0001870604],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04499241,0.001222134,0.924493,0.0002732244,0.0002353384,0.0002714456,0.0003868841,0.02268636,0.005439189],"genre_scores_gemma":[0.6840305,0.0004903018,0.3048957,0.0004318643,0.000135075,0.0003091201,0.0007108535,0.001559621,0.007437028],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003830154,"threshold_uncertainty_score":0.01281315,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4414427902","doi":"10.1145/3749156","title":"A Comprehensive Benchmark on Spectral GNNs: The Impact on Efficiency, Memory, and Effectiveness","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Nanyang Technological University; Ministry of Education - Singapore","keywords":"Benchmarking; Graph; Computation; Benchmark (surveying); Power graph analysis","authors":[{"name":"Ningyi Liao","is_ca":false},{"name":"Haoyu Liu","is_ca":false},{"name":"Zulun Zhu","is_ca":false},{"name":"Siqiang Luo","is_ca":false},{"name":"Laks V. S. Lakshmanan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02864368781006588,"gpt":0.3161488555937306,"spread":0.2875051677836648,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002263535,0.001547973,0.0009944419,0.0013298,0.0007921295,0.001511725,0.001873747,0.001695682,0.003391187],"category_scores_gemma":[0.01357295,0.000331977,0.0008192121,0.001260603,0.001075999,0.003761234,0.001339257,0.001510874,0.001077448],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001849986,"about_ca_system_score_gemma":0.00183975,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009189421,"about_ca_topic_score_gemma":0.01157457,"domain_scores_codex":[0.9986529,0.0003347236,0.0001000502,0.000329168,0.0003675241,0.0002156267],"domain_scores_gemma":[0.9950089,0.002661868,0.0002033629,0.001040265,0.0008963341,0.000189281],"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":[0.0008076993,0.0004219628,0.004741617,0.001107908,0.0002222813,0.0001753167,0.0001589423,0.729959,0.00834077,0.02648868,0.01426778,0.213308],"study_design_scores_gemma":[0.00006093071,0.0002560268,0.0007742235,0.00008439938,0.00005349481,0.0001410167,0.000142197,0.9645317,0.01076563,0.01872789,0.004434336,0.00002818169],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4522813,0.007020213,0.4893217,0.002442744,0.0009083007,0.00044166,0.00280787,0.01393755,0.03083869],"genre_scores_gemma":[0.7965266,0.001797562,0.1930272,0.0006968562,0.00008587243,0.0002322962,0.00374859,0.0008970127,0.002988137],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009189421,"threshold_uncertainty_score":0.01827186,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141435","doi":"10.1145/3725240","title":"Optimal Bounds for Private Minimum Spanning Trees via Input Perturbation","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Spanning tree; Minimum spanning tree; Minimum degree spanning tree; Combinatorics; Perturbation (astronomy); Mathematics; Physics","authors":[{"name":"Rasmus Pagh","is_ca":false},{"name":"Lukas Retschmeier","is_ca":false},{"name":"Hao Wu","is_ca":true},{"name":"H. H. Zhang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05061472259112793,"gpt":0.3086069096880938,"spread":0.2579921870969659,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00877942,0.002809689,0.003388687,0.001306024,0.001947948,0.005432856,0.005606494,0.003941247,0.006327416],"category_scores_gemma":[0.07445807,0.001386039,0.002058415,0.003160866,0.004206153,0.01854259,0.008033188,0.008867905,0.002103406],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005218908,"about_ca_system_score_gemma":0.002851008,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008971163,"about_ca_topic_score_gemma":0.001030073,"domain_scores_codex":[0.9903246,0.003533841,0.000449838,0.001932826,0.002587065,0.001171862],"domain_scores_gemma":[0.9349446,0.04905697,0.002441622,0.01020983,0.002028832,0.00131823],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.002328437,0.0004298385,0.002196353,0.0008706733,0.0002218336,0.0004472334,0.0008015051,0.6075369,0.0100692,0.2877742,0.01207034,0.07525355],"study_design_scores_gemma":[0.00007106192,0.0001110209,0.0002038615,0.00007048128,0.00005086856,0.0001868592,0.00008207118,0.7460705,0.003320727,0.2478764,0.001927912,0.0000282198],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0568811,0.002292663,0.915876,0.006377148,0.0002698445,0.0003100783,0.001256326,0.002454257,0.01428257],"genre_scores_gemma":[0.8011921,0.00162464,0.1866693,0.001586872,0.0004605689,0.000588624,0.001349859,0.00100573,0.005522352],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00877942,"threshold_uncertainty_score":0.04643059,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141774","doi":"10.1145/3725254","title":"Towards Update-Dependent Analysis of Query Maintenance","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Universitas Brawijaya","keywords":"Computer science; Query optimization; Information retrieval","authors":[{"name":"Qichen Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02913723348163801,"gpt":0.2919665382818373,"spread":0.2628293048001993,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005349948,0.001424246,0.00154552,0.002267504,0.001535538,0.004319074,0.005327314,0.001923601,0.004758934],"category_scores_gemma":[0.04531915,0.001286705,0.001663091,0.002471725,0.003311713,0.01619741,0.004588827,0.006128846,0.0006670789],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004477948,"about_ca_system_score_gemma":0.002279012,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002503661,"about_ca_topic_score_gemma":0.001546341,"domain_scores_codex":[0.9911203,0.001916179,0.0004077524,0.001551095,0.003848989,0.001155749],"domain_scores_gemma":[0.9401087,0.0459373,0.003401339,0.006764219,0.002862202,0.0009263075],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001585209,0.0005647189,0.007568233,0.0008807058,0.0001926669,0.0004111821,0.001461774,0.4350949,0.02336047,0.4115904,0.01168732,0.1056024],"study_design_scores_gemma":[0.00003927478,0.00008705042,0.0006790821,0.000022096,0.00004690688,0.0001104114,0.00008859996,0.8484834,0.003087997,0.1450814,0.002251905,0.00002187601],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.122953,0.002085907,0.8560124,0.003981926,0.0001578681,0.000219538,0.0008090761,0.001439383,0.01234091],"genre_scores_gemma":[0.8487242,0.001278881,0.1407419,0.001129284,0.0008809345,0.0003904511,0.001136612,0.0007428442,0.004974794],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005349948,"threshold_uncertainty_score":0.03248996,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399174990","doi":"10.1145/3654984","title":"Unstructured Data Fusion for Schema and Data Extraction","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Information retrieval; Schema (genetic algorithms); Component (thermodynamics); Data mining; Table (database)","authors":[{"name":"Kaiwen Chen","is_ca":true},{"name":"Nick Koudas","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1355387895140462,"gpt":0.3611646660040275,"spread":0.2256258764899813,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004862013,0.001593387,0.001557241,0.006114514,0.001077072,0.002766668,0.002541584,0.001382983,0.00415175],"category_scores_gemma":[0.01821226,0.0008305815,0.00274288,0.00765215,0.001160615,0.007479253,0.004695519,0.002757436,0.003603161],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001554175,"about_ca_system_score_gemma":0.002681629,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00352582,"about_ca_topic_score_gemma":0.005729317,"domain_scores_codex":[0.9950709,0.001003211,0.0005477504,0.00112372,0.002064773,0.0001896531],"domain_scores_gemma":[0.9894502,0.003347657,0.0006868435,0.004246322,0.002065433,0.0002035431],"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":[0.0004036581,0.0002785662,0.005788533,0.001537492,0.0003448768,0.0008471529,0.00110832,0.04544663,0.02907532,0.05368097,0.0351611,0.8263273],"study_design_scores_gemma":[0.00008141572,0.0002292187,0.00245652,0.000295901,0.0001905113,0.001447118,0.0009068243,0.6274142,0.08410067,0.1404037,0.1423143,0.0001596038],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005283117,0.001013606,0.9790235,0.0004291187,0.00009096947,0.0002791884,0.003824016,0.008567123,0.001489474],"genre_scores_gemma":[0.04773195,0.0006501893,0.9341906,0.0002889132,0.00004926557,0.0002496573,0.01518842,0.0004975746,0.001153388],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006114514,"threshold_uncertainty_score":0.02571309,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141791","doi":"10.1145/3725253","title":"Towards Practical FPRAS for #NFA: Exploiting the Power of Dependence","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Power (physics); Physics; Thermodynamics","authors":[{"name":"Kuldeep S. Meel","is_ca":true},{"name":"Alexis de Colnet","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3512875317024356,"gpt":0.4853916397535502,"spread":0.1341041080511146,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007619141,0.00180405,0.002123521,0.001946411,0.001543376,0.004060732,0.004912734,0.002449607,0.005085383],"category_scores_gemma":[0.03482883,0.00141468,0.003339996,0.002231611,0.003848786,0.01228066,0.004895981,0.005551391,0.00243956],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004026576,"about_ca_system_score_gemma":0.006692867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007882884,"about_ca_topic_score_gemma":0.01059101,"domain_scores_codex":[0.9893157,0.00340184,0.0008179074,0.002321109,0.00302411,0.001119297],"domain_scores_gemma":[0.9381166,0.03765275,0.003170654,0.01650433,0.003685525,0.0008700839],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001174947,0.0005070941,0.006478435,0.000898745,0.0002158488,0.0002841472,0.000770773,0.1362394,0.02052883,0.1559456,0.0176535,0.6593027],"study_design_scores_gemma":[0.00009375368,0.0001519788,0.000382783,0.0001059531,0.00007005306,0.0002331257,0.0001195729,0.8247061,0.01115805,0.155501,0.00741437,0.00006310589],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02657093,0.0004364779,0.9576004,0.001382251,0.00007257985,0.0002405763,0.0002503225,0.01003598,0.003410481],"genre_scores_gemma":[0.2345517,0.0003020953,0.7591935,0.0007360288,0.0001244426,0.0003332699,0.0006163393,0.0008630748,0.003279631],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007882884,"threshold_uncertainty_score":0.04029435,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4396892769","doi":"10.1145/3651144","title":"On Reporting Durable Patterns in Temporal Proximity Graphs","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Graph Theory Research","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science","authors":[{"name":"Pankaj K. Agarwal","is_ca":false},{"name":"Xiao Hu","is_ca":true},{"name":"Stavros Sintos","is_ca":false},{"name":"Jun Yang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08898262796644435,"gpt":0.3509226316586121,"spread":0.2619400036921677,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00434885,0.001260465,0.001666299,0.004998861,0.001511091,0.003356389,0.00251887,0.002103411,0.002496754],"category_scores_gemma":[0.07849787,0.0007972623,0.0007825799,0.01040198,0.001452022,0.01159207,0.004709005,0.001853199,0.001312314],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001575254,"about_ca_system_score_gemma":0.001721842,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008415269,"about_ca_topic_score_gemma":0.008173718,"domain_scores_codex":[0.9901949,0.001750855,0.00126123,0.002067976,0.004112924,0.0006122034],"domain_scores_gemma":[0.9394614,0.02979005,0.008557294,0.01502904,0.005701464,0.00146072],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.002112324,0.0004479681,0.06291169,0.001382423,0.0003623506,0.001407275,0.002773007,0.2936864,0.01984587,0.07045066,0.03765471,0.5069654],"study_design_scores_gemma":[0.0001239958,0.0003400637,0.00765728,0.0001130317,0.0001226616,0.002037141,0.001314301,0.801558,0.01030641,0.1574155,0.01890158,0.0001100644],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2144545,0.003842941,0.7467605,0.00332354,0.0004151045,0.0005476992,0.01020989,0.0128782,0.007567597],"genre_scores_gemma":[0.7403728,0.00127317,0.244774,0.0006585868,0.0002726724,0.0003044567,0.008913123,0.0005866873,0.002844581],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008415269,"threshold_uncertainty_score":0.02299917,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141667","doi":"10.1145/3725228","title":"An Improved Fully Dynamic Algorithm for Counting 4-Cycles in General Graphs Using Fast Matrix Multiplication","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Universitas Brawijaya","keywords":"Matrix multiplication; Multiplication (music); Algorithm; Computer science; Matrix (chemical analysis); Multiplication algorithm; Arithmetic; Mathematics; Parallel computing; Combinatorics; Physics; Materials science","authors":[{"name":"Sepehr Assadi","is_ca":true},{"name":"Vihan Shah","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02497041821931234,"gpt":0.3219790420322582,"spread":0.2970086238129459,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001041592,0.00165595,0.001616771,0.002666318,0.001278059,0.002537347,0.003604642,0.001327597,0.01203036],"category_scores_gemma":[0.006146848,0.0008749006,0.001291326,0.003829972,0.0009519055,0.006178799,0.003448999,0.001739551,0.004252287],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001900787,"about_ca_system_score_gemma":0.002633743,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009977807,"about_ca_topic_score_gemma":0.01361348,"domain_scores_codex":[0.9978501,0.0003181821,0.0001513171,0.0006304374,0.0007290445,0.0003209251],"domain_scores_gemma":[0.9964744,0.001018215,0.0002986877,0.001339835,0.0006638947,0.0002049714],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0007400023,0.0003360304,0.002027279,0.0004003229,0.0001325759,0.0002690744,0.0005853364,0.1395438,0.02373864,0.08544931,0.02562875,0.7211488],"study_design_scores_gemma":[0.0001362776,0.0001463081,0.000377422,0.00002866726,0.00002931542,0.0001791326,0.00009432121,0.8876814,0.005403137,0.0967899,0.009089132,0.00004502717],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01929984,0.0003318963,0.9684415,0.0003921062,0.0001310426,0.0002229926,0.0005214214,0.005271208,0.005387903],"genre_scores_gemma":[0.1358585,0.0001517815,0.8552765,0.0002378359,0.0001097614,0.0003911819,0.001612988,0.0006244942,0.005736893],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01203036,"threshold_uncertainty_score":0.04024559,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141648","doi":"10.1145/3725241","title":"Output-Optimal Algorithms for Join-Aggregate Queries","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Join (topology); Aggregate (composite); Computer science; Algorithm; Theoretical computer science; Mathematics; Combinatorics","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.06301675797609008,"gpt":0.3193127812564128,"spread":0.2562960232803227,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003603997,0.001711749,0.001780588,0.001483787,0.001346087,0.005278497,0.003008788,0.001668148,0.009325603],"category_scores_gemma":[0.01365406,0.0008835681,0.001980558,0.003356733,0.001483186,0.01081165,0.005173522,0.002788544,0.003959262],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002497193,"about_ca_system_score_gemma":0.002292391,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001804854,"about_ca_topic_score_gemma":0.002181964,"domain_scores_codex":[0.991699,0.001227083,0.0009802621,0.001777707,0.003043319,0.001272578],"domain_scores_gemma":[0.9901032,0.004178574,0.0006068093,0.003560973,0.001230023,0.000320424],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.004086814,0.0009916375,0.004244715,0.001530201,0.000262033,0.0002267126,0.001603623,0.1192017,0.04411084,0.2754199,0.03317313,0.5151488],"study_design_scores_gemma":[0.0003238949,0.000363624,0.000888912,0.00008822438,0.0001343487,0.0003425855,0.0004074856,0.5841764,0.03246497,0.3637544,0.01697115,0.00008406988],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04153471,0.001066912,0.9352366,0.0005878003,0.0001148361,0.0002255035,0.001093003,0.009313324,0.0108273],"genre_scores_gemma":[0.3614436,0.0007742166,0.6211678,0.000658564,0.0003264076,0.0005413918,0.004926794,0.00230897,0.007852244],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009325603,"threshold_uncertainty_score":0.03119719,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411141319","doi":"10.1145/3725250","title":"Smallest Synthetic Witnesses for Conjunctive Queries","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Universitas Brawijaya","keywords":"Conjunctive query; Computer science; Information retrieval; Relational database","authors":[{"name":"Aryan Esmailpour","is_ca":false},{"name":"Boris Glavic","is_ca":false},{"name":"Stavros Sintos","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2317777557871395,"gpt":0.4214874451748175,"spread":0.1897096893876779,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004020431,0.0008595621,0.001367024,0.0008886742,0.001081945,0.003187925,0.002167635,0.001437684,0.007455106],"category_scores_gemma":[0.02483471,0.0007734019,0.001584535,0.001491578,0.001520767,0.01071285,0.004040366,0.002093174,0.0008187562],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001215059,"about_ca_system_score_gemma":0.001468727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009988025,"about_ca_topic_score_gemma":0.001493088,"domain_scores_codex":[0.9950676,0.0009177409,0.0006559367,0.001470382,0.001471563,0.0004166955],"domain_scores_gemma":[0.981984,0.01209021,0.001313778,0.00304089,0.001016252,0.0005548978],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.004327036,0.0008808787,0.01579568,0.003210411,0.000586007,0.001174752,0.003803917,0.1259838,0.06883511,0.356698,0.02492024,0.3937841],"study_design_scores_gemma":[0.0004019359,0.0004683236,0.002008405,0.000151919,0.0001996635,0.0009720423,0.001399521,0.5079336,0.04656733,0.4245609,0.01524769,0.00008871698],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2338613,0.001030061,0.7444987,0.003600675,0.0001313919,0.0008121167,0.0048771,0.003692457,0.007496172],"genre_scores_gemma":[0.590995,0.0003445956,0.3972062,0.0005603964,0.000147233,0.0004115448,0.00641402,0.0005747877,0.00334617],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007455106,"threshold_uncertainty_score":0.02493984,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4396892510","doi":"10.1145/3651598","title":"Topology-aware Parallel Joins","year":2024,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Interconnection Networks and Systems","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Joins; Join (topology); Skew; Cartesian product; Network topology; Computer science; Topology (electrical circuits); Extension topology; Product topology; Computation; Logical topology; Theoretical computer science; Algorithm; Mathematics; Discrete mathematics; General topology; Combinatorics; Computer network; Topological space","authors":[{"name":"Xiao Hu","is_ca":true},{"name":"Paraschos Koutris","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06779863890756006,"gpt":0.303460955020995,"spread":0.2356623161134349,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0030725,0.0007192231,0.001382537,0.0006399719,0.001287841,0.002654974,0.003239954,0.001177652,0.002612579],"category_scores_gemma":[0.01119118,0.0009376251,0.0008179936,0.001279894,0.001586985,0.005046075,0.002535108,0.001836185,0.0006369869],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001559299,"about_ca_system_score_gemma":0.001582704,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00131852,"about_ca_topic_score_gemma":0.001278704,"domain_scores_codex":[0.9964007,0.001010255,0.0001372397,0.0007577371,0.001339269,0.0003547986],"domain_scores_gemma":[0.9927077,0.003934541,0.0008038918,0.001364078,0.0009020229,0.000287856],"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":[0.0001997548,0.00009281591,0.0005317449,0.00009463796,0.00004859893,0.0000597483,0.00009269454,0.8795198,0.003452198,0.08833679,0.001438821,0.02613235],"study_design_scores_gemma":[0.00001787096,0.00004804106,0.00004509521,0.000003546296,0.000009980741,0.00004107434,0.00001754746,0.9630043,0.001150741,0.03477133,0.0008854434,0.000005023095],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03208281,0.0002810652,0.9633032,0.0002587583,0.00005443965,0.00006727056,0.00005832039,0.0003498937,0.003544164],"genre_scores_gemma":[0.750513,0.0005854366,0.2428366,0.000191401,0.000173328,0.0002037205,0.000249633,0.0002241834,0.005022705],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003239954,"threshold_uncertainty_score":0.01624912,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W7109087591","doi":"10.1145/3769841","title":"Visualization-Oriented Progressive Time Series Transformation","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Transformation (genetics); Computation; Visualization; Data visualization; Multivariate statistics; Data transformation; Time series; Data manipulation language","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.0252930631897247,"gpt":0.3198411704681106,"spread":0.2945481072783859,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000878109,0.001030419,0.0008060341,0.001086287,0.0003698542,0.00175155,0.001623163,0.0005662288,0.005041774],"category_scores_gemma":[0.007139997,0.0004475253,0.0008872088,0.001396234,0.0006402779,0.002165023,0.002586669,0.001413252,0.00197088],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003863313,"about_ca_system_score_gemma":0.0008041564,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002227839,"about_ca_topic_score_gemma":0.002556802,"domain_scores_codex":[0.9991565,0.0001513692,0.00006222341,0.0002177626,0.0003548849,0.00005724297],"domain_scores_gemma":[0.9974407,0.0008889734,0.0001673503,0.0008273431,0.0005402891,0.0001353641],"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":[0.001026503,0.0002321682,0.004763535,0.0006184254,0.0002035605,0.0005756604,0.001157022,0.2009265,0.1715511,0.05201286,0.03719935,0.5297334],"study_design_scores_gemma":[0.00006871174,0.00009562179,0.0008909188,0.00001720164,0.00002535601,0.0002233713,0.0001174775,0.8986111,0.05466192,0.02917497,0.01606489,0.00004849501],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009393284,0.0001560956,0.974853,0.0001376246,0.00005781912,0.00005072405,0.0004150684,0.01390908,0.001027254],"genre_scores_gemma":[0.3550189,0.0005482093,0.6348733,0.0001734588,0.0001367502,0.0002651629,0.002632742,0.002976996,0.003374528],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005041774,"threshold_uncertainty_score":0.01686645,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411403412","doi":"10.1145/3725404","title":"Fast Maximum Common Subgraph Search: A Redundancy-Reduced Backtracking Approach","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Backtracking; Computer science; Benchmark (surveying); Redundancy (engineering); Graph; Computation; Theoretical computer science; Algorithm","authors":[{"name":"Kaiqiang Yu","is_ca":false},{"name":"Kaixin Wang","is_ca":false},{"name":"Cheng Long","is_ca":false},{"name":"Laks V. S. Lakshmanan","is_ca":true},{"name":"Reynold Cheng","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05965004820938316,"gpt":0.2988278835663659,"spread":0.2391778353569828,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002120654,0.001570724,0.00195736,0.003547509,0.001379081,0.001670575,0.004199226,0.002168079,0.00325601],"category_scores_gemma":[0.01018635,0.001004205,0.0021187,0.004058698,0.001664508,0.003910285,0.003335437,0.001960062,0.001299781],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001590196,"about_ca_system_score_gemma":0.00387508,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007519753,"about_ca_topic_score_gemma":0.01458142,"domain_scores_codex":[0.9971667,0.0007294914,0.0001625584,0.0006636289,0.0009005046,0.0003771628],"domain_scores_gemma":[0.9939954,0.002968369,0.0005198902,0.001660992,0.0006148114,0.0002405213],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0009612046,0.0008866323,0.004590924,0.0008047934,0.0003628701,0.0007638453,0.001029309,0.4489129,0.03341317,0.04924227,0.01760561,0.4414265],"study_design_scores_gemma":[0.0001126976,0.0001465086,0.0004902415,0.00004744813,0.00008621269,0.0002889663,0.0001496593,0.9520606,0.006815512,0.03647703,0.003293525,0.00003151609],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02847587,0.0005215927,0.9637913,0.0004461709,0.00004773442,0.0003351156,0.0003308574,0.003654317,0.00239711],"genre_scores_gemma":[0.1592596,0.0002411288,0.8357441,0.0002480579,0.00005513367,0.000270445,0.001520397,0.0005061312,0.002155001],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007519753,"threshold_uncertainty_score":0.01495194,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411403262","doi":"10.1145/3725319","title":"Low-Latency Transaction Scheduling via Userspace Interrupts: Why Wait or Yield When You Can Preempt?","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Universitas Brawijaya","keywords":"Computer science; Preemption; Context switch; Scheduling (production processes); Database transaction; Latency (audio); Operating system; Parallel computing; Embedded system; Distributed computing; Database","authors":[{"name":"Kaisong Huang","is_ca":true},{"name":"Zhuoyue Zhao","is_ca":false},{"name":"Dong Xie","is_ca":false},{"name":"Tianzheng Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05149839268555786,"gpt":0.2853484248596599,"spread":0.2338500321741021,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003160573,0.0004478312,0.0004877634,0.0003399819,0.0007214118,0.002782672,0.002260931,0.0007344066,0.00115912],"category_scores_gemma":[0.009897531,0.0004427385,0.000267209,0.0005724852,0.0008800514,0.003343276,0.001074112,0.001824372,0.0006112628],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007631886,"about_ca_system_score_gemma":0.001581981,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002805056,"about_ca_topic_score_gemma":0.004081802,"domain_scores_codex":[0.9977167,0.0003672025,0.0001465061,0.0003071519,0.001124831,0.0003376394],"domain_scores_gemma":[0.9944144,0.001759774,0.0007267229,0.001672153,0.00089389,0.0005330122],"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":[0.003451281,0.000825295,0.04657723,0.0008608284,0.000242588,0.0008984035,0.002456892,0.02664104,0.1228898,0.0583507,0.0246965,0.7121093],"study_design_scores_gemma":[0.0006376706,0.002488514,0.01984961,0.0004905466,0.0005121358,0.003247761,0.002234732,0.4783119,0.2509798,0.07648045,0.1644333,0.0003335829],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4052021,0.0164054,0.5485639,0.00541003,0.001255371,0.0002046324,0.0002120507,0.01037788,0.01236854],"genre_scores_gemma":[0.9242613,0.001811578,0.06920459,0.0009070029,0.0002160001,0.00004700834,0.0001131737,0.0004905087,0.0029489],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003160573,"threshold_uncertainty_score":0.01671487,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W7109234313","doi":"10.1145/3769790","title":"Enjima: A Resource-Adaptive Stream Processing System","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Latency (audio); Stream processing; Dataflow; Pipeline (software); Workload; Throughput; Scheduling (production processes); Memory management; Response time","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.04016599984991381,"gpt":0.2851652727165403,"spread":0.2449992728666265,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007664327,0.0007740131,0.0007297269,0.0006554783,0.0004353111,0.00132545,0.002299268,0.0004597279,0.005636983],"category_scores_gemma":[0.002095441,0.0004721901,0.0003545192,0.0007648136,0.0003196421,0.002074215,0.001943996,0.00127307,0.002023542],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007095376,"about_ca_system_score_gemma":0.001419092,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002021376,"about_ca_topic_score_gemma":0.002358289,"domain_scores_codex":[0.9994066,0.00006310135,0.00007282425,0.0001633148,0.0002078174,0.00008643699],"domain_scores_gemma":[0.9992535,0.0001265878,0.00004956985,0.0002221747,0.0002260452,0.000122119],"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":[0.005967497,0.001358652,0.0134969,0.0007785579,0.0003980988,0.00125661,0.0005784752,0.04085836,0.166778,0.01906561,0.1789621,0.5705011],"study_design_scores_gemma":[0.00120173,0.001188355,0.007695702,0.0001010354,0.0003000828,0.001013743,0.0001854237,0.75293,0.07964373,0.01050672,0.14499,0.0002434837],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1590155,0.002968136,0.6372416,0.0008712966,0.001036442,0.001361755,0.004887958,0.1549608,0.03765649],"genre_scores_gemma":[0.5582297,0.00171595,0.3940147,0.001843796,0.0002913846,0.000789406,0.01337608,0.002122561,0.02761638],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005636983,"threshold_uncertainty_score":0.0188576,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4407355838","doi":"10.1145/3709681","title":"Dialogue Benchmark Generation from Knowledge Graphs with Cost-Effective Retrieval-Augmented LLMs","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on Management of Data","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Benchmark (surveying); Computer science; Artificial intelligence; Information retrieval; Geography","authors":[{"name":"Reham Omar","is_ca":true},{"name":"Omij Mangukiya","is_ca":true},{"name":"E. M. E. Mansour","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05128988014828679,"gpt":0.2954122021952202,"spread":0.2441223220469334,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003410154,0.002092182,0.001239639,0.002265146,0.0007089255,0.001739802,0.00282885,0.001160755,0.00489074],"category_scores_gemma":[0.02211322,0.0006527437,0.0008963281,0.001167516,0.0007579128,0.002754302,0.00327131,0.001364743,0.002398876],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001149572,"about_ca_system_score_gemma":0.001469553,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004539392,"about_ca_topic_score_gemma":0.006106475,"domain_scores_codex":[0.9959283,0.002090082,0.0002469329,0.0007815512,0.000730618,0.0002224558],"domain_scores_gemma":[0.9888275,0.006756993,0.0004036657,0.002018052,0.001690155,0.0003037944],"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":[0.001869318,0.0008376754,0.004778986,0.001165181,0.0002097138,0.000816551,0.001438316,0.2329872,0.0390308,0.009083797,0.04471408,0.6630684],"study_design_scores_gemma":[0.0001648005,0.0002496626,0.0007474704,0.00002982591,0.00005935042,0.00009229151,0.0003186364,0.9578103,0.02260318,0.009819552,0.008062644,0.00004219076],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0931278,0.0006280868,0.7810889,0.0005263693,0.0002143479,0.000959839,0.002766921,0.1145603,0.006127271],"genre_scores_gemma":[0.5247341,0.0001654682,0.4511634,0.0002994257,0.00006845826,0.001105446,0.01424265,0.004989411,0.003231558],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00489074,"threshold_uncertainty_score":0.01803488,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}