{"meta":{"query_hash":"8458b7f752df","filters":{"venue":"Distributed and Parallel Databases"},"cohort_total":20,"direct_labels_cover":1,"predictions_cover":20,"exported":20,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/8458b7f752df","api":"https://metacan.xera.ac/api/v1/cohort?venue=Distributed+and+Parallel+Databases"},"results":[{"id":"W139568493","doi":"10.1023/a:1019258807615","title":"A Graphical Approach to Allocating Class Fragments in Distributed Objectbase Systems","year":2001,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"VLSI and FPGA Design Techniques","field":"Engineering","cited_by":11,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Computer science; Overhead (engineering); Scheme (mathematics); Mathematical optimization; Optimization algorithm; Mathematics","score_opus":0.025049053546780962,"score_gpt":0.24733264465686508,"score_spread":0.2222835911100841,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W139568493","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0024935673,0.00007578188,0.99337643,0.0001269259,0.000024633366,0.000041454772,0.00003834272,0.0023499415,0.0014728815],"genre_scores_gemma":[0.10226159,0.0002534956,0.89164066,0.00017380061,0.000042659398,0.00016297282,0.00014886641,0.0006202619,0.004695639],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988055,0.0003875619,0.000089328154,0.0001601942,0.0004130945,0.00014434174],"domain_scores_gemma":[0.9981844,0.00067673984,0.00010726298,0.0006432876,0.000248837,0.00013956909],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014325168,0.00079908397,0.0006992935,0.0013832005,0.0010864788,0.0032733185,0.004031214,0.0013663417,0.011188525],"category_scores_gemma":[0.004280206,0.0008957552,0.0011921149,0.0015735833,0.0018174854,0.0036192155,0.002444275,0.0018030179,0.0017880421],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008693324,0.00025997826,0.0010634952,0.00051711383,0.0001175873,0.00041394006,0.00082881696,0.13688557,0.028627174,0.40602264,0.01289436,0.4115],"study_design_scores_gemma":[0.00022812036,0.00019358216,0.00037236337,0.00010903036,0.0001934637,0.0003911131,0.00018965449,0.7390787,0.031751897,0.17828555,0.049111966,0.000094586576],"about_ca_topic_score_codex":0.004588507,"about_ca_topic_score_gemma":0.009088657,"teacher_disagreement_score":0.011188525,"about_ca_system_score_codex":0.0009799014,"about_ca_system_score_gemma":0.0009788806,"threshold_uncertainty_score":0.037429392},"labels":[],"label_agreement":null},{"id":"W1991438431","doi":"10.1007/s10619-006-6575-6","title":"The cgmCUBE project: Optimizing parallel data cube generation for ROLAP","year":2006,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":52,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University; Concordia University; Carleton University","funders":"","keywords":"Online analytical processing; Computer science; Data cube; Data warehouse; Cube (algebra); Very large database; Database; Data structure; Data mining; Programming language","score_opus":0.07826883404703516,"score_gpt":0.31089565877940334,"score_spread":0.23262682473236818,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1991438431","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.124522574,0.002088694,0.65030843,0.0024927973,0.001159609,0.0010779529,0.0125628095,0.1819978,0.023789389],"genre_scores_gemma":[0.17595337,0.0005217138,0.7820124,0.00047740282,0.00014717704,0.00083414675,0.020949084,0.010721168,0.008383492],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99781835,0.0005991581,0.00009332442,0.00033819923,0.0008870298,0.00026385306],"domain_scores_gemma":[0.99742234,0.0006357193,0.00007392431,0.00081805384,0.000834422,0.0002154405],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021921853,0.0024395392,0.0017856633,0.0014434205,0.0012924671,0.0021700913,0.0041486816,0.0013092596,0.0138794985],"category_scores_gemma":[0.00558696,0.0009932668,0.0014769455,0.0029446757,0.0011895683,0.0027504012,0.002956788,0.0022022275,0.0036423423],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0026571823,0.0008101522,0.0024170063,0.0007650937,0.00044830397,0.00032660112,0.00042881272,0.16359143,0.01839018,0.018209651,0.3368922,0.4550634],"study_design_scores_gemma":[0.001088609,0.0002558173,0.0006389589,0.00003160577,0.00011478155,0.00009585804,0.00017997019,0.93155736,0.026174337,0.013440469,0.026348047,0.00007417494],"about_ca_topic_score_codex":0.027344162,"about_ca_topic_score_gemma":0.023391921,"teacher_disagreement_score":0.027344162,"about_ca_system_score_codex":0.0018802277,"about_ca_system_score_gemma":0.004020597,"threshold_uncertainty_score":0.054369986},"labels":[],"label_agreement":null},{"id":"W1993755977","doi":"10.1007/s10619-010-7076-1","title":"Query processing in sensor networks","year":2010,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Energy Efficient Wireless Sensor Networks","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Data structure; Programming language","score_opus":0.014371540501215427,"score_gpt":0.2452749747759609,"score_spread":0.2309034342747455,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1993755977","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.025726534,0.013876883,0.9503101,0.0023698304,0.0005461208,0.00014439895,0.0002310034,0.0011005957,0.0056945863],"genre_scores_gemma":[0.6907648,0.011439794,0.27855822,0.0006763541,0.001286075,0.00038315094,0.00092653354,0.00029564337,0.015669461],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9969796,0.00080211874,0.0002858237,0.0005147551,0.0012184351,0.00019931232],"domain_scores_gemma":[0.9965293,0.002077899,0.00018744032,0.0005989973,0.00048290784,0.0001235151],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003061294,0.0007395448,0.002094516,0.001092676,0.00093195384,0.003188719,0.0020768049,0.0013286789,0.0027571793],"category_scores_gemma":[0.008178885,0.0006840996,0.00053122605,0.0030995305,0.0015046247,0.005013625,0.001318874,0.0014614544,0.00073694735],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007627958,0.0003243574,0.0018595102,0.0010918436,0.00021079341,0.00029297295,0.0005293209,0.35133156,0.01206548,0.23688254,0.024036774,0.37061208],"study_design_scores_gemma":[0.000045275327,0.000061880615,0.00036592156,0.000025061518,0.0000511055,0.00014728053,0.00010814188,0.85835594,0.0026008748,0.12710656,0.011114403,0.000017520972],"about_ca_topic_score_codex":0.0033932745,"about_ca_topic_score_gemma":0.0021909955,"teacher_disagreement_score":0.0033932745,"about_ca_system_score_codex":0.0012131539,"about_ca_system_score_gemma":0.0010840575,"threshold_uncertainty_score":0.016189873},"labels":[],"label_agreement":null},{"id":"W2008361063","doi":"10.1007/s10619-009-7052-9","title":"Guest editorial: special issue on ranking in databases","year":2009,"lang":"en","type":"editorial","venue":"Distributed and Parallel Databases","topic":"Web visibility and informetrics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Ranking (information retrieval); Database; Information retrieval; Data science","score_opus":0.024128819301910822,"score_gpt":0.2946601032873167,"score_spread":0.2705312839854059,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2008361063","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00003024686,0.004496959,0.00024244831,0.026199382,0.9674503,0.0000141117125,0.000062418665,0.00006217431,0.001441909],"genre_scores_gemma":[0.0004340225,0.0032417474,0.00019009328,0.00884772,0.97800624,0.0000150205515,0.000047737904,0.000058316815,0.0091591505],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99290365,0.001215162,0.00096675707,0.00084136537,0.003606891,0.0004661938],"domain_scores_gemma":[0.9702837,0.009283311,0.0020753255,0.00091082184,0.014116196,0.0033307185],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0095784655,0.0039309873,0.0052634855,0.007115558,0.002845708,0.010098059,0.0038556566,0.014240259,0.021243356],"category_scores_gemma":[0.0301631,0.0014806547,0.0026140022,0.0030825872,0.0022549874,0.004862401,0.0017861666,0.015112235,0.013118104],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000021341473,0.000008595854,0.000017692568,0.00011964913,0.00001635688,0.000063757936,0.0000033828708,0.000018357014,0.00003722999,0.00024790387,0.99612397,0.0033217603],"study_design_scores_gemma":[0.00006864019,0.00002891918,0.00027317338,0.00027764993,0.000076729586,0.00030649104,0.000022112119,0.00030613726,0.00017395316,0.0013764189,0.9970694,0.000020403731],"about_ca_topic_score_codex":0.0012579759,"about_ca_topic_score_gemma":0.00449617,"teacher_disagreement_score":0.021243356,"about_ca_system_score_codex":0.0032446,"about_ca_system_score_gemma":0.0026478171,"threshold_uncertainty_score":0.07106614},"labels":[],"label_agreement":null},{"id":"W2016326061","doi":"10.1023/b:dapd.0000013069.97679.62","title":"An Adaptive Data-Shipping Architecture for Client Caching Data Management Systems","year":2004,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Distributed computing; Cache; Robustness (evolution); Architecture; Data management; Database; Computer network","score_opus":0.09128449405044331,"score_gpt":0.32144718531403554,"score_spread":0.23016269126359223,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2016326061","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.13364714,0.0009117297,0.8096563,0.00070262415,0.0002857309,0.00045856385,0.0003254289,0.04567545,0.008336972],"genre_scores_gemma":[0.60574234,0.00040104662,0.37883437,0.00043606138,0.00013379105,0.0002734466,0.0009158886,0.0010794271,0.012183589],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990877,0.0001591205,0.00009090256,0.00014458437,0.00037086598,0.0001468957],"domain_scores_gemma":[0.9970958,0.00031601381,0.00012478884,0.0014734354,0.00071991904,0.00027001702],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013422768,0.0006612741,0.00079161907,0.000948699,0.001333092,0.0021501235,0.005543297,0.0010743502,0.0037464972],"category_scores_gemma":[0.0028930067,0.0007766071,0.0004470207,0.0016927257,0.0009025387,0.00247484,0.0020895735,0.0017530684,0.0011628456],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0033734827,0.0013023142,0.011845556,0.00040320505,0.0003231022,0.0007486968,0.0010879366,0.10793613,0.12394921,0.046522323,0.05546553,0.6470426],"study_design_scores_gemma":[0.00015762389,0.0002175635,0.001528961,0.000019085765,0.00013701004,0.0002197417,0.00009727926,0.9406896,0.033359386,0.010077896,0.013420449,0.00007540141],"about_ca_topic_score_codex":0.009625981,"about_ca_topic_score_gemma":0.0098641245,"teacher_disagreement_score":0.009625981,"about_ca_system_score_codex":0.0016801473,"about_ca_system_score_gemma":0.0019358828,"threshold_uncertainty_score":0.019139886},"labels":[],"label_agreement":null},{"id":"W2047284331","doi":"10.1007/s10619-007-7018-8","title":"Adaptive processing of historical spatial range queries in peer-to-peer sensor networks","year":2007,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Energy Efficient Wireless Sensor Networks","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Athabasca University; University of Alberta","funders":"","keywords":"Computer science; Query optimization; Flooding (psychology); Spatial query; Online aggregation; Data mining; Distributed computing; Query expansion; Energy consumption; Web search query; Web query classification; Information retrieval; Search engine","score_opus":0.026885808651503078,"score_gpt":0.26124794385443934,"score_spread":0.23436213520293625,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2047284331","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1779164,0.0018015134,0.8164073,0.0005285833,0.00027670755,0.00012601832,0.00015171789,0.0012429878,0.001548804],"genre_scores_gemma":[0.929632,0.00052952586,0.06772701,0.0001086142,0.0001957186,0.00007674369,0.00023864598,0.000094471456,0.0013973638],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99789894,0.00034698384,0.00019234745,0.000453234,0.00094412547,0.00016435725],"domain_scores_gemma":[0.99478817,0.0027549192,0.00045826868,0.00081637,0.0009674601,0.00021482134],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002225367,0.0006350565,0.0019236753,0.0011150892,0.0010540878,0.0014436088,0.0026068748,0.0010613382,0.0007990592],"category_scores_gemma":[0.011292258,0.00066974515,0.00031933875,0.001658902,0.00085883145,0.0029062321,0.0015301016,0.00089618284,0.00027639524],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012287238,0.00041725853,0.0072011705,0.0002888331,0.00017914928,0.00059545983,0.00092890067,0.57223094,0.045080595,0.010791087,0.004588774,0.3564691],"study_design_scores_gemma":[0.000023910734,0.00010785376,0.0010452943,0.0000062421955,0.000034988654,0.00020589269,0.00016036635,0.9850552,0.0058788275,0.006501196,0.00096297654,0.000017211913],"about_ca_topic_score_codex":0.0026570738,"about_ca_topic_score_gemma":0.0032110128,"teacher_disagreement_score":0.0026570738,"about_ca_system_score_codex":0.00060419174,"about_ca_system_score_gemma":0.0006477019,"threshold_uncertainty_score":0.011768997},"labels":[],"label_agreement":null},{"id":"W2070626055","doi":"10.1007/s10619-005-4257-4","title":"Preventive Replication in a Database Cluster","year":2005,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":39,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Replica; Computer science; Replication (statistics); Database transaction; Benchmark (surveying); Distributed database; Database; Throughput; Cluster (spacecraft); Consistency (knowledge bases); Data consistency; Distributed computing; Parallel computing; Computer network; Operating system; Mathematics; Artificial intelligence; Geography","score_opus":0.022992260274848044,"score_gpt":0.2855527655536087,"score_spread":0.2625605052787607,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2070626055","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.665119,0.0010902104,0.3153157,0.0015548621,0.000402871,0.00020755817,0.00017281756,0.0071416,0.008995437],"genre_scores_gemma":[0.94873804,0.00013380445,0.047484554,0.00006507943,0.000067313194,0.000058130405,0.000103399914,0.00010111501,0.0032485598],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99827206,0.00036432274,0.00010693992,0.00034258375,0.0006448559,0.00026919637],"domain_scores_gemma":[0.99284756,0.0013434512,0.00043280204,0.0032817188,0.0013254778,0.0007689146],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0022535557,0.00034117338,0.000983835,0.0009174201,0.0023527348,0.0022202588,0.002826814,0.0010308621,0.001878176],"category_scores_gemma":[0.006117551,0.0006690131,0.00035875133,0.001592368,0.001040887,0.0017512776,0.0019969768,0.0009913741,0.000552555],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.005734737,0.0014740296,0.029233722,0.00049888494,0.00029728335,0.0020864322,0.002092231,0.4561655,0.092112124,0.13894469,0.023200693,0.24815968],"study_design_scores_gemma":[0.00026971995,0.00036860124,0.0022699055,0.000011754745,0.00012445616,0.0003905015,0.00034245636,0.94942486,0.013546498,0.026628926,0.0065715355,0.000050815648],"about_ca_topic_score_codex":0.0044888747,"about_ca_topic_score_gemma":0.0036591792,"teacher_disagreement_score":0.0044888747,"about_ca_system_score_codex":0.0011384516,"about_ca_system_score_gemma":0.002497375,"threshold_uncertainty_score":0.011918128},"labels":[],"label_agreement":null},{"id":"W2072466548","doi":"10.1007/s10619-006-6284-1","title":"Incremental click-stream tree model: Learning from new users for web page prediction","year":2006,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Clickstream; Tree (set theory); Web log analysis software; Data mining; Process (computing); Online model; Web server; Recommender system; Data stream; Web page; Document Object Model; Machine learning; Database; World Wide Web; Information retrieval; Static web page; The Internet; Web API","score_opus":0.040724674281577715,"score_gpt":0.262250113213566,"score_spread":0.22152543893198828,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2072466548","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.114395864,0.0019444042,0.8758866,0.0009805956,0.00026931876,0.00024153254,0.0019551797,0.0028654893,0.0014610142],"genre_scores_gemma":[0.80521435,0.0012160905,0.18272403,0.00031347867,0.00048608446,0.0003607777,0.0028357343,0.00013905796,0.006710448],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99920315,0.00022395058,0.00005954769,0.00020719708,0.0002204157,0.000085795786],"domain_scores_gemma":[0.99510854,0.0030349134,0.00025109499,0.00057782227,0.00077432295,0.0002533389],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027428796,0.0010086668,0.0029571871,0.002593286,0.00060376234,0.0011137293,0.0041434625,0.0021669588,0.0023621786],"category_scores_gemma":[0.007665153,0.0006919852,0.001000067,0.0036711292,0.0005264181,0.0026259378,0.0008054379,0.002020273,0.0014923079],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012634892,0.0014416811,0.017458614,0.00030755182,0.00038559627,0.0001635407,0.0001875563,0.50039285,0.0027533732,0.010079391,0.016159926,0.44940645],"study_design_scores_gemma":[0.0000130085955,0.00002630717,0.00029022477,0.0000032721152,0.00002116138,0.000014050629,0.0000035149562,0.9973558,0.00017853717,0.0019184496,0.00017070596,0.000005020778],"about_ca_topic_score_codex":0.016670158,"about_ca_topic_score_gemma":0.028402103,"teacher_disagreement_score":0.016670158,"about_ca_system_score_codex":0.00080061174,"about_ca_system_score_gemma":0.0013537748,"threshold_uncertainty_score":0.033146262},"labels":[],"label_agreement":null},{"id":"W2078842512","doi":"10.1007/s10619-005-3296-1","title":"Stream Cube: An Architecture for Multi-Dimensional Analysis of Data Streams","year":2005,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":155,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"National Science Foundation","keywords":"Computer science; Data cube; Data stream mining; Data stream; Data structure; Volume (thermodynamics); Online analytical processing; Computation; Stream processing; Data mining; Algorithm; Parallel computing; Data warehouse","score_opus":0.10197947995303348,"score_gpt":0.3616370861324058,"score_spread":0.25965760617937234,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2078842512","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0108086085,0.00067170954,0.9556461,0.0004456774,0.00019523046,0.00018670212,0.0013707546,0.028293975,0.0023812922],"genre_scores_gemma":[0.16158845,0.0022511766,0.8191268,0.00069580623,0.0002826515,0.0006530093,0.0065284213,0.0017969046,0.007076782],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991358,0.00014937647,0.00009149202,0.000118145676,0.00042507687,0.00008014939],"domain_scores_gemma":[0.99819165,0.00046482543,0.00007923576,0.0004784862,0.00061611616,0.0001696427],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014162146,0.00087929453,0.0011973183,0.0016507809,0.0007687686,0.002640029,0.002882491,0.0007365436,0.0069296677],"category_scores_gemma":[0.0032746505,0.00063577935,0.00084637053,0.003502876,0.000649701,0.0027895083,0.0021787074,0.0014302207,0.0025723677],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0035346486,0.00045627294,0.0067023193,0.000945957,0.00067504187,0.0007056603,0.0008050214,0.057004437,0.042656466,0.05215815,0.1454743,0.68888175],"study_design_scores_gemma":[0.00031214696,0.00035007513,0.0012764444,0.00010586349,0.00030377106,0.0005064694,0.00021288927,0.8270568,0.049704667,0.06654923,0.053470727,0.00015100834],"about_ca_topic_score_codex":0.0054131607,"about_ca_topic_score_gemma":0.0047933105,"teacher_disagreement_score":0.0069296677,"about_ca_system_score_codex":0.00077785173,"about_ca_system_score_gemma":0.0022261655,"threshold_uncertainty_score":0.023182034},"labels":[],"label_agreement":null},{"id":"W2080139625","doi":"10.1007/s10619-011-7085-8","title":"Scaling XML query processing: distribution, localization and pruning","year":2011,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Query optimization; XML database; Query plan; Sargable; Scalability; Query expansion; Query language; Information retrieval; Database; Data mining; XML; Web search query; Search engine; World Wide Web","score_opus":0.04172179025022446,"score_gpt":0.25388811319574134,"score_spread":0.21216632294551688,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2080139625","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.24064867,0.0061815097,0.7250597,0.002591463,0.00039495522,0.0003173503,0.000617764,0.015361886,0.008826747],"genre_scores_gemma":[0.6888909,0.0010031975,0.30320737,0.00031528904,0.00032680423,0.0002179364,0.0010192282,0.0013685183,0.0036508585],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9933683,0.0015643644,0.00055700185,0.000995848,0.0028389315,0.0006754935],"domain_scores_gemma":[0.97683245,0.011428139,0.0010471989,0.006909888,0.0031536908,0.000628538],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0048868386,0.0014331978,0.0026056627,0.0032511172,0.001650663,0.0038636515,0.004108101,0.0014872306,0.0035795271],"category_scores_gemma":[0.028652439,0.0012025412,0.0009095823,0.0053054444,0.0014051488,0.0063958494,0.00313172,0.0021088507,0.0011180115],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016578225,0.0010616846,0.011299197,0.00038373965,0.00015329347,0.00023399266,0.0006536415,0.16132991,0.028289348,0.02445721,0.0194799,0.75100034],"study_design_scores_gemma":[0.00014443322,0.00013946845,0.0019725938,0.000022455099,0.00009167016,0.0002495499,0.00025965102,0.96259177,0.013676189,0.017900411,0.0029218164,0.000029962743],"about_ca_topic_score_codex":0.009337986,"about_ca_topic_score_gemma":0.012078158,"teacher_disagreement_score":0.009337986,"about_ca_system_score_codex":0.0024123508,"about_ca_system_score_gemma":0.0031101466,"threshold_uncertainty_score":0.025844395},"labels":[],"label_agreement":null},{"id":"W2121550967","doi":"10.1007/s10619-006-0445-0","title":"Parallel Bifold: Large-scale parallel pattern mining with constraints","year":2006,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Data mining; Set (abstract data type); Scale (ratio); Cluster (spacecraft); Space (punctuation); Geography","score_opus":0.01781232993566086,"score_gpt":0.2489065911592234,"score_spread":0.23109426122356252,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2121550967","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.019468617,0.0008050703,0.9437988,0.0006642881,0.00019621196,0.0004731791,0.0038345319,0.02704843,0.003710785],"genre_scores_gemma":[0.110424764,0.00040448748,0.8775632,0.00024410782,0.00009568212,0.00063596136,0.006436954,0.0013270769,0.0028677096],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9986268,0.00028451154,0.00014897497,0.00027066434,0.00054857566,0.00012048103],"domain_scores_gemma":[0.99631953,0.0015542813,0.00020480377,0.0010275074,0.0006662662,0.00022757701],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019613293,0.0017196579,0.0016444339,0.002001039,0.001206923,0.0024106037,0.004040366,0.0007619838,0.010435264],"category_scores_gemma":[0.009506075,0.0010920162,0.0009877778,0.0041488,0.0006855054,0.0034221772,0.0024251381,0.0017678519,0.0026872645],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0022353253,0.0005070019,0.0052324906,0.0012375221,0.00036473095,0.00044202304,0.00015020896,0.09284389,0.012479699,0.024504054,0.09135471,0.7686483],"study_design_scores_gemma":[0.00043839813,0.0001251833,0.0006974468,0.000035681667,0.00008355484,0.0002540932,0.0000601481,0.9346517,0.011299043,0.037888125,0.014425162,0.000041527706],"about_ca_topic_score_codex":0.0064712507,"about_ca_topic_score_gemma":0.012514701,"teacher_disagreement_score":0.010435264,"about_ca_system_score_codex":0.0006957521,"about_ca_system_score_gemma":0.003569142,"threshold_uncertainty_score":0.034909427},"labels":[],"label_agreement":null},{"id":"W2126220026","doi":"10.1007/s10619-009-7043-x","title":"Continuously monitoring top-k uncertain data streams: a probabilistic threshold method","year":2009,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Data stream mining; Computer science; Probabilistic logic; Uncertain data; Quantile; Data mining; Data structure; Algorithm; Artificial intelligence; Mathematics; Statistics","score_opus":0.07410263996446681,"score_gpt":0.33747829743305896,"score_spread":0.26337565746859215,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2126220026","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.003978627,0.00017406272,0.9952342,0.0000863295,0.000028258812,0.000019152587,0.000029753497,0.00021058317,0.00023905202],"genre_scores_gemma":[0.4719142,0.00061205,0.524624,0.00018237748,0.00028290635,0.00015563151,0.0002432449,0.00019151231,0.0017941537],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9968671,0.0005829437,0.0002857062,0.0008765584,0.0011154978,0.00027213985],"domain_scores_gemma":[0.9928019,0.0044576004,0.00056759996,0.00085528195,0.0010315066,0.00028603012],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038330804,0.0009202566,0.002570082,0.0016785525,0.0008430126,0.0027603002,0.0042544194,0.0017054002,0.00165694],"category_scores_gemma":[0.01366929,0.00080945116,0.001399651,0.0028563482,0.0013104716,0.0049600936,0.002616903,0.0024531411,0.00039469995],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000903664,0.00019251734,0.0041209804,0.00032578877,0.00035838285,0.00016365823,0.00024572274,0.5255739,0.010735191,0.044198092,0.002801724,0.41038042],"study_design_scores_gemma":[0.000010785816,0.000024752679,0.00015338151,0.000005665315,0.000029942003,0.00003307959,0.0000105895315,0.98849213,0.0010614841,0.009845556,0.00032237783,0.000010188254],"about_ca_topic_score_codex":0.0040516607,"about_ca_topic_score_gemma":0.002489069,"teacher_disagreement_score":0.0042544194,"about_ca_system_score_codex":0.0013972656,"about_ca_system_score_gemma":0.0017986806,"threshold_uncertainty_score":0.02027154},"labels":[],"label_agreement":null},{"id":"W2133443743","doi":"10.1007/s10619-007-7023-y","title":"PnP: sequential, external memory, and parallel iceberg cube computation","year":2008,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University; Concordia University; Carleton University","funders":"","keywords":"Computer science; Speedup; Cube (algebra); Scalability; Iceberg; Auxiliary memory; Data cube; Computation; Parallel computing; Pruning; Data structure; Algorithm; Data mining; Mathematics; Computer hardware; Database; Operating system; Geology; Combinatorics","score_opus":0.04727396136730296,"score_gpt":0.2760956854569131,"score_spread":0.22882172408961013,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2133443743","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027323527,0.0008679874,0.94570416,0.0011996208,0.00045327062,0.000094609546,0.0006018129,0.0028608611,0.020894172],"genre_scores_gemma":[0.43474853,0.0013182582,0.54185563,0.00047536925,0.0004100715,0.0003111874,0.0016013172,0.00067866594,0.018601062],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994141,0.0001250216,0.00002928832,0.000092135495,0.000252494,0.00008705304],"domain_scores_gemma":[0.9990388,0.00025735132,0.00004377695,0.00033601926,0.00024089267,0.00008302947],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008272703,0.00051446253,0.0007908778,0.00062199956,0.0010025233,0.0019354518,0.0017979104,0.000610955,0.004949782],"category_scores_gemma":[0.003623422,0.0003409097,0.0004900451,0.0016717361,0.0011544114,0.0030580638,0.0028675445,0.0014541253,0.0012192648],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00063801144,0.000117147836,0.0016161753,0.00021857394,0.000055533543,0.00018995423,0.0001839041,0.20341241,0.00277407,0.50001436,0.052880727,0.23789914],"study_design_scores_gemma":[0.000047065936,0.00002976144,0.00021565963,0.000015251333,0.000014909337,0.00008510893,0.00003935875,0.6142029,0.0023657575,0.36864862,0.014322256,0.000013201232],"about_ca_topic_score_codex":0.0042358832,"about_ca_topic_score_gemma":0.0045939963,"teacher_disagreement_score":0.004949782,"about_ca_system_score_codex":0.0008661708,"about_ca_system_score_gemma":0.0017616005,"threshold_uncertainty_score":0.016558647},"labels":[],"label_agreement":null},{"id":"W2135133299","doi":"10.1023/b:dapd.0000018572.20283.e0","title":"Parallel ROLAP Data Cube Construction on Shared-Nothing Multiprocessors","year":2004,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":58,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University; Dalhousie University","funders":"","keywords":"Computer science; Online analytical processing; Row; Parallel computing; Data cube; Speedup; Data warehouse; Data structure; Cube (algebra); Database; Data mining; Programming language; Mathematics","score_opus":0.05168987859294967,"score_gpt":0.2929623151742056,"score_spread":0.24127243658125597,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2135133299","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.099168725,0.00068716245,0.87551373,0.0006726076,0.0001985528,0.00024746422,0.0012450911,0.011774555,0.010492015],"genre_scores_gemma":[0.40411124,0.00030217945,0.5849125,0.0001875427,0.00005345553,0.00030325894,0.0033342212,0.00080845517,0.005987238],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9985838,0.00026181547,0.00009482989,0.00019022211,0.00058945065,0.00027986718],"domain_scores_gemma":[0.99746215,0.00046386046,0.00009571444,0.0012301704,0.0005924018,0.00015564133],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00097553956,0.00054132316,0.001256119,0.0010833394,0.0014250489,0.0021226832,0.0031038148,0.00066572736,0.0048035914],"category_scores_gemma":[0.0029629134,0.0005584235,0.0011035869,0.0024617293,0.00091903773,0.0031119108,0.0043629827,0.0011623324,0.0013178656],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.001936885,0.00043478582,0.004097954,0.00092013483,0.0002735665,0.00088057306,0.0008853112,0.20394652,0.037065312,0.078645684,0.05212998,0.6187833],"study_design_scores_gemma":[0.00022898523,0.00025277675,0.0010542274,0.00006448712,0.00011729076,0.0004597613,0.0008296918,0.78787965,0.07214467,0.10830128,0.028574953,0.00009231224],"about_ca_topic_score_codex":0.004628518,"about_ca_topic_score_gemma":0.0056627323,"teacher_disagreement_score":0.0048035914,"about_ca_system_score_codex":0.0007098239,"about_ca_system_score_gemma":0.0018945914,"threshold_uncertainty_score":0.01606965},"labels":[],"label_agreement":null},{"id":"W2299020861","doi":"10.1007/s10619-017-7196-y","title":"TAPER: query-aware, partition-enhancement for large, heterogenous graphs","year":2017,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"Newcastle University","keywords":"Computer science; Graph partition; Scalability; Partition (number theory); Workload; Space partitioning; Hash function; Parallel computing; USable; Theoretical computer science; Graph; Algorithm; Database; Mathematics","score_opus":0.03487313379500431,"score_gpt":0.29416619769974,"score_spread":0.2592930639047357,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2299020861","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.16025329,0.0013263997,0.67429435,0.00070074265,0.00022201124,0.00065569876,0.002019888,0.14870286,0.0118247615],"genre_scores_gemma":[0.45027322,0.00047130717,0.53403234,0.0003476655,0.000091445,0.00025647925,0.005056041,0.004598227,0.004873315],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99880433,0.0002936354,0.000064922795,0.00025546277,0.00044357587,0.00013820732],"domain_scores_gemma":[0.99606687,0.0012481328,0.00021663155,0.0018856608,0.00038933224,0.0001933747],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014216948,0.0013071185,0.00084722176,0.00128349,0.00085103104,0.0015022369,0.0031049629,0.0007546573,0.0032864395],"category_scores_gemma":[0.0057904273,0.0006189782,0.001001679,0.0014879693,0.0008986771,0.003459591,0.0028887417,0.0009066815,0.001661609],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.002215709,0.0007030598,0.006097755,0.0009581133,0.00027905192,0.0005196759,0.0011321434,0.24885133,0.10563132,0.015683308,0.07224961,0.545679],"study_design_scores_gemma":[0.00029019156,0.00052260194,0.0024334267,0.000031550426,0.000065814645,0.0005055854,0.00040409548,0.9222855,0.04014053,0.012280303,0.020971293,0.00006902803],"about_ca_topic_score_codex":0.0066344882,"about_ca_topic_score_gemma":0.009993817,"teacher_disagreement_score":0.0066344882,"about_ca_system_score_codex":0.0009187161,"about_ca_system_score_gemma":0.0010310138,"threshold_uncertainty_score":0.0131917},"labels":[],"label_agreement":null},{"id":"W3088726914","doi":"10.1007/s10619-020-07313-y","title":"Introduction to the special issue on Self-managing and Hardware-Optimized Database Systems 2019","year":2020,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Cloud Computing and Resource Management","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"University of Waterloo; University of California, Davis; Institute of Computing Technology, Chinese Academy of Sciences; University of Pittsburgh; Chinese Academy of Sciences; Oracle","keywords":"Computer science; Database; Data structure; Computer architecture; Operating system","score_opus":0.01679472336439665,"score_gpt":0.23503290242983793,"score_spread":0.2182381790654413,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3088726914","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0005575067,0.15906054,0.019584425,0.042787377,0.7325895,0.00010384843,0.0006221026,0.0005268244,0.04416788],"genre_scores_gemma":[0.0031166233,0.10062878,0.005694346,0.015513726,0.76429456,0.000068048124,0.0011159541,0.0006768332,0.10889118],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99785334,0.0002589503,0.00024104067,0.0004353919,0.0010336273,0.00017769038],"domain_scores_gemma":[0.9937583,0.0021112657,0.0003005687,0.00048681843,0.002476268,0.0008667824],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00245995,0.0016738264,0.00184699,0.0033079218,0.0011488124,0.0063997717,0.002056568,0.0031691154,0.05384265],"category_scores_gemma":[0.006667159,0.00085853203,0.0011493225,0.0038558508,0.0011351418,0.006367992,0.002142086,0.006295943,0.02956954],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000026813219,0.00004150885,0.000104593164,0.00038212605,0.00001684473,0.000063775806,0.0000144659425,0.00026147606,0.00041457167,0.0046540815,0.9303562,0.0636635],"study_design_scores_gemma":[0.0000058367755,0.000030779433,0.00021180467,0.00018313112,0.000011048718,0.00021706893,0.000011688663,0.0004155105,0.00016680625,0.0033040114,0.9954276,0.000014648862],"about_ca_topic_score_codex":0.0009519695,"about_ca_topic_score_gemma":0.002144462,"teacher_disagreement_score":0.05384265,"about_ca_system_score_codex":0.0016604973,"about_ca_system_score_gemma":0.0015819359,"threshold_uncertainty_score":0.1801216},"labels":[],"label_agreement":null},{"id":"W4287307637","doi":"10.1007/s10619-022-07415-9","title":"Scalable probabilistic truss decomposition using central limit theorem and H-index","year":2022,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"","keywords":"Truss; Probabilistic logic; Computer science; Scalability; Algorithm; Discrete mathematics; Mathematical optimization; Mathematics; Theoretical computer science; Artificial intelligence; Structural engineering","score_opus":0.023529447674205426,"score_gpt":0.2592946132143147,"score_spread":0.23576516554010926,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4287307637","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0116692195,0.0001551098,0.98419166,0.00011772312,0.000038122704,0.000042592474,0.00012020217,0.0008191321,0.0028462696],"genre_scores_gemma":[0.38318563,0.00027687388,0.6091279,0.00014259967,0.00010322515,0.00018460158,0.0006213535,0.00044346842,0.0059144325],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991653,0.00016242072,0.000040710027,0.00015863706,0.00037847218,0.00009438367],"domain_scores_gemma":[0.99797624,0.0006090135,0.00017646405,0.0006898076,0.00041358173,0.00013497594],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00096137734,0.0006913165,0.0018176957,0.0014826257,0.0008380573,0.001635729,0.0022094261,0.0009503908,0.0054688053],"category_scores_gemma":[0.004161977,0.0006647219,0.00088928017,0.0019362587,0.00088679173,0.0037726148,0.002561056,0.0013098976,0.0012193432],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003462147,0.00023686628,0.00080935797,0.00018887167,0.00011281367,0.00011707229,0.00013954003,0.5953601,0.009983049,0.14922418,0.012343562,0.23113836],"study_design_scores_gemma":[0.00001181578,0.000016751383,0.000060683466,0.000003713433,0.0000072447037,0.000023684866,0.00001213093,0.9549565,0.0008092101,0.043369185,0.00072244136,0.000006628179],"about_ca_topic_score_codex":0.0028293033,"about_ca_topic_score_gemma":0.0045773457,"teacher_disagreement_score":0.0054688053,"about_ca_system_score_codex":0.0010281355,"about_ca_system_score_gemma":0.0013837208,"threshold_uncertainty_score":0.01829499},"labels":[{"model":"gemma","categories":[],"domain":null,"study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"simulation_or_modeling","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4405880184","doi":"10.1007/s10619-024-07450-8","title":"Infomod: information-theoretic machine learning model diagnostics","year":2024,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; National Science Foundation","keywords":"Automatic summarization; Computer science; SPARK (programming language); Debugging; Process (computing); Machine learning; Artificial intelligence; Ranging; Test data; Data mining; Programming language","score_opus":0.01595322642870192,"score_gpt":0.25958549389043345,"score_spread":0.24363226746173153,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405880184","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004871241,0.00028060365,0.98217505,0.00072840194,0.000095109266,0.00008129604,0.0011022156,0.008340497,0.0023255283],"genre_scores_gemma":[0.40786648,0.00060904643,0.57742554,0.0009063344,0.00035418582,0.00053180946,0.0037520938,0.003021488,0.005533024],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9972691,0.0011525647,0.00014211106,0.00038001893,0.00093453313,0.000121643454],"domain_scores_gemma":[0.9890215,0.00745992,0.0006270899,0.0018928584,0.0008359735,0.00016276746],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051384945,0.0014186971,0.0013636831,0.0020811893,0.00055662636,0.0027328804,0.0028368037,0.0019631856,0.012406537],"category_scores_gemma":[0.029934924,0.00087194046,0.0017063657,0.0008952031,0.0012285379,0.0033383553,0.003251799,0.0032481735,0.0032717942],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007947109,0.0003470592,0.0034789445,0.00049068837,0.00038696153,0.00043877307,0.00008515013,0.5007887,0.003502041,0.18463641,0.041871455,0.2631791],"study_design_scores_gemma":[0.000023491884,0.00003391631,0.00012975364,0.000016242724,0.000015332613,0.000066870925,0.000006629534,0.93052775,0.0023089452,0.064952396,0.0019069101,0.000011813644],"about_ca_topic_score_codex":0.0013174533,"about_ca_topic_score_gemma":0.0012231936,"teacher_disagreement_score":0.012406537,"about_ca_system_score_codex":0.001093608,"about_ca_system_score_gemma":0.0022909446,"threshold_uncertainty_score":0.041503966},"labels":[],"label_agreement":null},{"id":"W4406162498","doi":"10.1007/s10619-024-07452-6","title":"Indexing temporal relations for range-duration queries","year":2025,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Sherbrooke","funders":"Università degli Studi di Padova","keywords":"Search engine indexing; Computer science; Data structure; Range (aeronautics); Duration (music); Information retrieval; Data mining; Programming language","score_opus":0.02640828025020287,"score_gpt":0.2848948208440601,"score_spread":0.25848654059385723,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4406162498","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.086216345,0.003514922,0.8790755,0.000987962,0.00030921004,0.0004990244,0.0052321707,0.01514415,0.009020805],"genre_scores_gemma":[0.40816247,0.0013671694,0.57499754,0.00039100606,0.00045030063,0.0004453617,0.010063549,0.0013005134,0.0028220634],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99399924,0.0007638817,0.0010406843,0.0008517045,0.0029226865,0.00042179017],"domain_scores_gemma":[0.9857918,0.0051540215,0.0014843871,0.0050045443,0.0019455869,0.0006197222],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029160744,0.0010077587,0.0016233888,0.0029134192,0.0011840701,0.004017375,0.0024341538,0.0010374428,0.003914461],"category_scores_gemma":[0.01822824,0.00054214464,0.0009431298,0.006167846,0.0010926516,0.007940743,0.0034152549,0.0016164145,0.001588207],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0024987801,0.001032905,0.022997243,0.0017001067,0.00028278475,0.00093219656,0.002383182,0.103198245,0.07632829,0.16739516,0.06934532,0.55190575],"study_design_scores_gemma":[0.00024494974,0.000403966,0.0031826126,0.00016708711,0.00010503838,0.0011305999,0.00072633865,0.7711482,0.031794455,0.11376171,0.07720368,0.0001312678],"about_ca_topic_score_codex":0.003733999,"about_ca_topic_score_gemma":0.003590401,"teacher_disagreement_score":0.004017375,"about_ca_system_score_codex":0.0012684303,"about_ca_system_score_gemma":0.0021219896,"threshold_uncertainty_score":0.015421808},"labels":[],"label_agreement":null},{"id":"W7117103606","doi":"10.1007/s10619-025-07455-x","title":"Revisiting shared registers and leaderless consensus in WAN environments","year":2025,"lang":"en","type":"article","venue":"Distributed and Parallel Databases","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Latency (audio); Throughput; Overhead (engineering); Protocol (science); Process (computing); Communications protocol; Distributed database","score_opus":0.029226334190379503,"score_gpt":0.2739984755017218,"score_spread":0.24477214131134228,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W7117103606","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07197771,0.0042459145,0.87895226,0.0070049143,0.0013063747,0.00009089119,0.00011262388,0.00085004506,0.03545916],"genre_scores_gemma":[0.86510223,0.003205983,0.10871863,0.0009429807,0.0010498363,0.00010629943,0.00015202284,0.0005842589,0.020137824],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9976089,0.0009834323,0.000121618425,0.0003996225,0.00055595674,0.00033035237],"domain_scores_gemma":[0.98896885,0.0062864115,0.00032809147,0.0028007592,0.0013490252,0.0002668202],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0050325952,0.00051981077,0.0009316981,0.0008503266,0.0017758168,0.0036953748,0.0031865116,0.0018463606,0.004867916],"category_scores_gemma":[0.019453669,0.00057252665,0.0006594996,0.0014078886,0.0035257128,0.016054278,0.003926725,0.0033398247,0.0010140539],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00038064568,0.00009140014,0.0009647949,0.0002445801,0.0000409106,0.00024238694,0.0010294041,0.062644884,0.0022302822,0.825875,0.0053498778,0.10090588],"study_design_scores_gemma":[0.00006976415,0.00010760469,0.00022263297,0.00008318076,0.000057114252,0.000072635456,0.0005185762,0.21522391,0.0031261384,0.75428593,0.026195757,0.00003662724],"about_ca_topic_score_codex":0.003191395,"about_ca_topic_score_gemma":0.0037084485,"teacher_disagreement_score":0.0050325952,"about_ca_system_score_codex":0.001196941,"about_ca_system_score_gemma":0.0019202388,"threshold_uncertainty_score":0.026615262},"labels":[],"label_agreement":null}]}