{"meta":{"query_hash":"ea3402be1ff5","filters":{"venue":"Very Large Data Bases"},"cohort_total":17,"direct_labels_cover":0,"predictions_cover":17,"exported":17,"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/ea3402be1ff5","api":"https://metacan.xera.ac/api/v1/cohort?venue=Very+Large+Data+Bases"},"results":[{"id":"W10721341","doi":"10.5694/j.1326-5377.1999.tb123814.x","title":"Letter from the Research Track Co-Chair.","year":2011,"lang":"en","type":"letter","venue":"Very Large Data Bases","topic":"","field":"","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 Toronto","funders":"","keywords":"Computer science; Track (disk drive); Operating system","score_opus":0.2888122786995828,"score_gpt":0.38252509556888176,"score_spread":0.09371281686929894,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W10721341","genre_codex":"commentary","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00043621234,0.0011962139,0.00031438484,0.94414103,0.019673321,0.000050769053,0.0008773917,0.0001488818,0.03316186],"genre_scores_gemma":[0.003988091,0.00083362457,0.0005520818,0.81093377,0.008468904,0.00011934682,0.00044018487,0.00012375937,0.17454028],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99746466,0.0005245795,0.00017268828,0.00044343254,0.0008663308,0.0005283472],"domain_scores_gemma":[0.99597436,0.0015254937,0.00016991507,0.00015489648,0.0012423721,0.00093296793],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036324817,0.00065644825,0.0009884194,0.000663878,0.0045488123,0.0053143096,0.0012500082,0.027955616,0.03298968],"category_scores_gemma":[0.012798113,0.00068762637,0.00093921076,0.000743677,0.0012568898,0.00393425,0.0014824683,0.02026253,0.040553503],"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":[0.000031170843,0.000008076226,0.000114864895,0.0000059403164,0.0000021873072,0.0000799016,0.000009893104,0.000014158858,0.000031075353,0.0005947556,0.99707615,0.0020319642],"study_design_scores_gemma":[0.00005957926,0.000036250793,0.00097579183,0.00007538638,0.000011682948,0.0002480699,0.00025926394,0.0003327095,0.00023563042,0.0030911716,0.99463934,0.00003514107],"about_ca_topic_score_codex":0.014554436,"about_ca_topic_score_gemma":0.032167956,"teacher_disagreement_score":0.03298968,"about_ca_system_score_codex":0.0035233167,"about_ca_system_score_gemma":0.0066801542,"threshold_uncertainty_score":0.11036146},"labels":[],"label_agreement":null},{"id":"W145033769","doi":"","title":"The Long-Term Preservation of Authentic Electronic Records","year":2001,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":26,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Term (time); Computer science; Phase (matter); Data science","score_opus":0.0536414918134093,"score_gpt":0.24818344492992914,"score_spread":0.19454195311651984,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W145033769","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":[0.2827854,0.012043978,0.4347984,0.04633716,0.0008463392,0.0005210074,0.000729156,0.0007547812,0.22118391],"genre_scores_gemma":[0.931049,0.0025472278,0.05291867,0.00075781165,0.00049729034,0.00018959877,0.0003626348,0.00008012361,0.011597676],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9698648,0.017519062,0.001515384,0.0018197774,0.00808911,0.0011918367],"domain_scores_gemma":[0.8688199,0.043271814,0.016417868,0.05060521,0.01719236,0.0036929063],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02908105,0.0002177698,0.0004086574,0.0024502524,0.0059623485,0.017290054,0.0035236727,0.0027367899,0.0034865127],"category_scores_gemma":[0.08869656,0.0004659056,0.00029707997,0.00422262,0.015603739,0.02252215,0.015776666,0.0034307907,0.0013644282],"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":[0.00014947013,0.00012213377,0.008690615,0.0005233974,0.00002511581,0.00029433356,0.021687781,0.0009300313,0.0017337591,0.755517,0.004481291,0.20584516],"study_design_scores_gemma":[0.000050469276,0.0005461091,0.012339986,0.0019366777,0.00010776146,0.0027939351,0.06421522,0.0052198083,0.010051208,0.5321887,0.37043452,0.00011562485],"about_ca_topic_score_codex":0.0010287537,"about_ca_topic_score_gemma":0.001096295,"teacher_disagreement_score":0.02908105,"about_ca_system_score_codex":0.0022997556,"about_ca_system_score_gemma":0.006041354,"threshold_uncertainty_score":0.15379709},"labels":[],"label_agreement":null},{"id":"W1489499991","doi":"","title":"Personalizing XML text search in PIMENT","year":2005,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Advanced Database Systems and Queries","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":"University of British Columbia","funders":"","keywords":"Computer science; Information retrieval; Web search query; Ranking (information retrieval); Personalization; Query expansion; Web query classification; XML; World Wide Web; Search engine","score_opus":0.056066286513882985,"score_gpt":0.31407353981900304,"score_spread":0.2580072533051201,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1489499991","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":[0.12130383,0.0008991062,0.79181707,0.0006876909,0.00009227661,0.0007707912,0.0017662148,0.06979152,0.012871568],"genre_scores_gemma":[0.4514003,0.0005329584,0.521467,0.0007933828,0.00013963191,0.00041754107,0.0070129894,0.0015708441,0.016665295],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9972198,0.0008176783,0.00026537597,0.00053525297,0.0009824252,0.00017948393],"domain_scores_gemma":[0.9971547,0.0009407081,0.00017884553,0.0011512142,0.00046037458,0.00011408275],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029846674,0.0005597528,0.00094831095,0.002111426,0.00076843385,0.0018688046,0.0014406822,0.00082679914,0.0023989426],"category_scores_gemma":[0.008237071,0.00050192483,0.0006928549,0.001721826,0.0005085737,0.0029845678,0.0020000022,0.0007569488,0.0018312093],"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":[0.0017181006,0.00068668334,0.008517252,0.0004789566,0.00026235895,0.0012357589,0.00251842,0.023845015,0.079916365,0.028452527,0.048951413,0.8034171],"study_design_scores_gemma":[0.00019926035,0.00044388504,0.0075951484,0.000072063536,0.00024343838,0.0019115862,0.0008297662,0.6380879,0.16082951,0.03739826,0.15213831,0.00025078873],"about_ca_topic_score_codex":0.0025018468,"about_ca_topic_score_gemma":0.0041519315,"teacher_disagreement_score":0.0029846674,"about_ca_system_score_codex":0.00066206994,"about_ca_system_score_gemma":0.0007425213,"threshold_uncertainty_score":0.015784621},"labels":[],"label_agreement":null},{"id":"W2041180570","doi":"10.5555/1182635.1164189","title":"Lazy database replication with snapshot isolation","year":2006,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":146,"is_retracted":false,"has_abstract":true,"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; Snapshot (computer storage); Serializability; Distributed computing; Concurrency; Database; Database transaction; Concurrency control; Isolation (microbiology); Data integrity; Transaction processing; Distributed transaction","score_opus":0.025532778176718715,"score_gpt":0.26121851579626315,"score_spread":0.23568573761954442,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2041180570","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":[0.017526342,0.0011260249,0.97454923,0.00032842887,0.00013440967,0.0001614823,0.00004915777,0.0033714417,0.0027534608],"genre_scores_gemma":[0.58713806,0.0008472968,0.40535644,0.00038987695,0.00038442318,0.00035175568,0.00023976974,0.00041342492,0.004879017],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99356985,0.0015869244,0.000612619,0.0010539084,0.0026969712,0.00047975674],"domain_scores_gemma":[0.9839048,0.0022397668,0.0013274051,0.010437664,0.0015736514,0.0005166863],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0039539086,0.0006781252,0.0011520978,0.0010296901,0.0015288164,0.0028977217,0.0034569583,0.0012402235,0.002114192],"category_scores_gemma":[0.013251941,0.0007166082,0.00082903134,0.0017139226,0.0018663798,0.006081867,0.006638477,0.0021408552,0.0013037812],"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":[0.001020331,0.00038877712,0.0066314917,0.0010265983,0.00046989665,0.0009515239,0.001329683,0.13637203,0.070636906,0.31635746,0.015575776,0.44923955],"study_design_scores_gemma":[0.0003658333,0.0008035468,0.0015572178,0.00014714927,0.0004168316,0.0019194215,0.00042203238,0.588038,0.080245495,0.26885822,0.057027258,0.00019892535],"about_ca_topic_score_codex":0.0007328423,"about_ca_topic_score_gemma":0.0008262879,"teacher_disagreement_score":0.0039539086,"about_ca_system_score_codex":0.00084154546,"about_ca_system_score_gemma":0.002071357,"threshold_uncertainty_score":0.020910501},"labels":[],"label_agreement":null},{"id":"W2108075439","doi":"","title":"Approximate joins: concepts and techniques","year":2005,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Joins; Multitude; Computer science; Database; Quality (philosophy); Data quality; Character (mathematics); Service (business); Relational database; Missing data; Data science; Data mining; World Wide Web; Business; Marketing","score_opus":0.2365840025725053,"score_gpt":0.4607702635409333,"score_spread":0.224186260968428,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2108075439","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":[0.0026884517,0.010040408,0.9739386,0.00075435895,0.00026033336,0.000109252695,0.0003528027,0.0007315455,0.011124185],"genre_scores_gemma":[0.09598676,0.016111223,0.8736335,0.00069232815,0.0020302222,0.00049263315,0.0013824686,0.0004471985,0.009223702],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9927654,0.0016660726,0.0006580524,0.0014117584,0.003171111,0.00032763864],"domain_scores_gemma":[0.99400604,0.0031927156,0.0005770659,0.0012163812,0.000822228,0.0001855755],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0060401033,0.0013723798,0.0017882354,0.0055304132,0.0020785113,0.007529389,0.004463063,0.0021236117,0.007152562],"category_scores_gemma":[0.014311964,0.0013455267,0.0020865074,0.01196514,0.003817465,0.0141124325,0.006193004,0.0039420747,0.0038183809],"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":[0.000082302715,0.000057099416,0.0006265952,0.0005938731,0.00009858916,0.00015687889,0.0006541516,0.011403973,0.0007031186,0.8232879,0.011162531,0.15117304],"study_design_scores_gemma":[0.000019219538,0.000044858672,0.00019723272,0.00013402534,0.000051671865,0.00046390796,0.0001446096,0.02946806,0.00071584183,0.9121346,0.05660031,0.00002572359],"about_ca_topic_score_codex":0.0018656141,"about_ca_topic_score_gemma":0.0010323875,"teacher_disagreement_score":0.007529389,"about_ca_system_score_codex":0.0016497046,"about_ca_system_score_gemma":0.0016040823,"threshold_uncertainty_score":0.03194344},"labels":[],"label_agreement":null},{"id":"W2118326624","doi":"10.5555/1182635.1164140","title":"Efficient secure query evaluation over encrypted XML databases","year":2006,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Cryptography and Data Security","field":"Computer Science","cited_by":126,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Database; Encryption; Database security; Metadata; View; XML database; XML Encryption; Information retrieval; XML; Computer security; Database design; World Wide Web","score_opus":0.036339330316432605,"score_gpt":0.2968752340972579,"score_spread":0.2605359037808253,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2118326624","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":[0.2920678,0.000589723,0.70038605,0.0009532759,0.000046257683,0.00024881374,0.00040247772,0.0027441634,0.0025613923],"genre_scores_gemma":[0.88170767,0.00021395854,0.11600208,0.0001301677,0.00004457253,0.000074874515,0.00032559634,0.00013458569,0.0013665378],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9903555,0.0022402348,0.0013234554,0.0009771566,0.004263247,0.00084046595],"domain_scores_gemma":[0.9826225,0.008430656,0.0014325131,0.005415978,0.0017425771,0.0003558203],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0049346685,0.0007589792,0.0015178842,0.0007413275,0.00092343835,0.002481638,0.0018535851,0.0010536155,0.0014502835],"category_scores_gemma":[0.013776188,0.0005003082,0.0009924505,0.0012729615,0.0011453623,0.0052298787,0.0032850637,0.001410986,0.00047317252],"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":[0.0057080504,0.0010512121,0.015619588,0.0006394408,0.00035343616,0.0018301662,0.0023544158,0.3993197,0.13205251,0.08125668,0.0077032833,0.35211158],"study_design_scores_gemma":[0.00011239023,0.00016449594,0.00075209566,0.000015683976,0.000045098634,0.0004090489,0.00025496102,0.9330346,0.039456617,0.024343006,0.001390265,0.000021710246],"about_ca_topic_score_codex":0.0018911185,"about_ca_topic_score_gemma":0.001627897,"teacher_disagreement_score":0.0049346685,"about_ca_system_score_codex":0.0016323886,"about_ca_system_score_gemma":0.0016782066,"threshold_uncertainty_score":0.026097357},"labels":[],"label_agreement":null},{"id":"W2144006311","doi":"","title":"MIX: a meta-data indexing system for XML","year":2005,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"XPath; Computer science; XML database; Information retrieval; Streaming XML; XML validation; Efficient XML Interchange; XML Encryption; Document Structure Description; XML Schema (W3C); Database; XML Schema Editor; Search engine indexing; Simple API for XML; XML Signature; XML; Data mining; World Wide Web","score_opus":0.1406926542850825,"score_gpt":0.32553850099715964,"score_spread":0.18484584671207713,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2144006311","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":[0.0071364837,0.0010771449,0.83837867,0.00040318322,0.00017698982,0.00035672632,0.0048828973,0.13935123,0.008236703],"genre_scores_gemma":[0.11146884,0.0012670263,0.8288536,0.0011669444,0.00041755583,0.0012515475,0.023518326,0.013148443,0.018907692],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9974757,0.0003117606,0.00035413625,0.00038783601,0.0013289327,0.00014172807],"domain_scores_gemma":[0.9976005,0.0005757527,0.00022352955,0.0010214783,0.00036101014,0.00021764527],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027528284,0.00096826785,0.0011809561,0.0033249378,0.0010659483,0.005422274,0.0035735455,0.0013044184,0.011839228],"category_scores_gemma":[0.0055264076,0.0012455905,0.0010675524,0.0034969887,0.0009825493,0.009650578,0.005644654,0.0022196053,0.0073918975],"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":[0.0020765746,0.00030258863,0.0053681172,0.001578296,0.0003509223,0.00061815087,0.0015276207,0.007286623,0.0719251,0.12301332,0.14867045,0.63728225],"study_design_scores_gemma":[0.00047873362,0.00043897418,0.0020279125,0.0002882265,0.00030523227,0.0015711854,0.0003870929,0.14924979,0.13529475,0.08170447,0.62792546,0.00032829287],"about_ca_topic_score_codex":0.0013402937,"about_ca_topic_score_gemma":0.0012619777,"teacher_disagreement_score":0.011839228,"about_ca_system_score_codex":0.0012417039,"about_ca_system_score_gemma":0.0014229247,"threshold_uncertainty_score":0.039606214},"labels":[],"label_agreement":null},{"id":"W2169424245","doi":"10.5555/1182635.1164177","title":"Answering tree pattern queries using views","year":2006,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":70,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Rewriting; Schema (genetic algorithms); Computer science; Time complexity; Tree (set theory); Theoretical computer science; Mathematics; Algorithm; Combinatorics; Programming language; Information retrieval","score_opus":0.06735454525073631,"score_gpt":0.29709100551204315,"score_spread":0.22973646026130684,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2169424245","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":[0.06990268,0.000971986,0.92400825,0.0006985263,0.00004466183,0.00012596998,0.00041417527,0.0012251046,0.0026087363],"genre_scores_gemma":[0.53073156,0.0013664652,0.46160507,0.0005156415,0.00021520571,0.00024010152,0.0021716172,0.0003971329,0.0027572692],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99623424,0.0013086679,0.00026471628,0.00079693896,0.0010893695,0.00030603402],"domain_scores_gemma":[0.98807794,0.008647798,0.00068630744,0.0016355102,0.00077119557,0.00018121282],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0023971493,0.0005957028,0.0011964315,0.0009056213,0.000649592,0.0022405568,0.0017320508,0.0013661499,0.0016821304],"category_scores_gemma":[0.0125670545,0.00056865666,0.0012519641,0.0023282242,0.0011491337,0.008404446,0.0020251407,0.0018821806,0.0004800105],"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":[0.00093084143,0.0004938488,0.008150309,0.0010558042,0.00030355988,0.0010037287,0.0024136389,0.19154291,0.042370945,0.33952588,0.013674659,0.3985339],"study_design_scores_gemma":[0.000091807364,0.00029309443,0.00090466853,0.0000461338,0.00010786954,0.0006821129,0.00042462465,0.6291592,0.021055127,0.33566114,0.011526421,0.00004773899],"about_ca_topic_score_codex":0.0024204613,"about_ca_topic_score_gemma":0.0018390074,"teacher_disagreement_score":0.0024204613,"about_ca_system_score_codex":0.00092530646,"about_ca_system_score_gemma":0.0006984763,"threshold_uncertainty_score":0.012677491},"labels":[],"label_agreement":null},{"id":"W2398890672","doi":"","title":"Towards Adaptive Resource Allocation for Database Workloads.","year":2015,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Cloud Computing and Resource Management","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"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; Workload; Scalability; Resource allocation; Distributed computing; Performance tuning; Performance metric; Real-time computing; Database; Operating system","score_opus":0.09405781546455771,"score_gpt":0.29614691187237346,"score_spread":0.20208909640781575,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2398890672","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":[0.01238081,0.00046531632,0.9843531,0.00016523938,0.000042159496,0.00007163581,0.000018358582,0.0006037851,0.0018996777],"genre_scores_gemma":[0.6872973,0.00048089254,0.30959687,0.00025091541,0.000076382625,0.00019760388,0.00007277625,0.00011429685,0.0019129615],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999138,0.00022063433,0.000047063426,0.0001710849,0.0003461598,0.0000770854],"domain_scores_gemma":[0.99858594,0.0006914492,0.00018533571,0.0001956242,0.0002850061,0.000056723355],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011180877,0.00086017046,0.0005408033,0.0006871102,0.00033713886,0.0010004706,0.0011878726,0.0006456757,0.00081246806],"category_scores_gemma":[0.00498716,0.00038435968,0.00027251657,0.0005515374,0.00061170134,0.0010648977,0.00089277804,0.0011908863,0.00041263644],"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":[0.00024381849,0.00020952035,0.0027222822,0.00025056532,0.000069405214,0.00013726293,0.00024263423,0.53081936,0.079237476,0.031879365,0.003607346,0.350581],"study_design_scores_gemma":[0.000007950195,0.00003431517,0.00024958496,0.000011399321,0.00000523573,0.00003948028,0.000017425356,0.987633,0.0042921994,0.0064956397,0.0012053618,0.000008544881],"about_ca_topic_score_codex":0.0016558259,"about_ca_topic_score_gemma":0.0018062296,"teacher_disagreement_score":0.0016558259,"about_ca_system_score_codex":0.0007273798,"about_ca_system_score_gemma":0.0007755676,"threshold_uncertainty_score":0.0059131384},"labels":[],"label_agreement":null},{"id":"W2400133291","doi":"","title":"Robust concurrency control in main-memory DBMS: What main memory giveth, the application taketh away.","year":2014,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":1,"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 Toronto","funders":"","keywords":"Computer science; Concurrency control; Multiversion concurrency control; Non-lock concurrency control; Concurrency; Database; Optimistic concurrency control; Memory management; Distributed concurrency control; Parallel computing; Programming language; Operating system; Database transaction; Overlay","score_opus":0.027956875898266287,"score_gpt":0.25220511764092207,"score_spread":0.22424824174265579,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2400133291","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":[0.06354129,0.03829704,0.8436385,0.030155472,0.0029400932,0.00010340091,0.0001815319,0.0045832316,0.016559528],"genre_scores_gemma":[0.8988123,0.007834743,0.07454379,0.0026172919,0.0021700212,0.00010195866,0.00015533902,0.0011068665,0.012657665],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99720687,0.00073618273,0.00015204136,0.0003682455,0.001130802,0.0004059456],"domain_scores_gemma":[0.99245304,0.002724254,0.0003474038,0.002682819,0.0012908374,0.0005016149],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003904853,0.0005264324,0.000865498,0.00043304576,0.00078254903,0.0044812844,0.0025428713,0.0017760491,0.0041165296],"category_scores_gemma":[0.020115884,0.0007530269,0.0005249267,0.0005178711,0.003653837,0.012297507,0.0028419949,0.0047403965,0.0010836578],"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":[0.0009655048,0.00025480907,0.0046198457,0.00095779344,0.00024057279,0.00027534578,0.0014014539,0.07101655,0.02079394,0.44645157,0.034536064,0.4184866],"study_design_scores_gemma":[0.00011459263,0.0002389335,0.0011375223,0.00035198862,0.00013103249,0.0002844229,0.0005678135,0.3165948,0.01956395,0.6127646,0.048104614,0.00014563547],"about_ca_topic_score_codex":0.002900196,"about_ca_topic_score_gemma":0.0020035792,"teacher_disagreement_score":0.0044812844,"about_ca_system_score_codex":0.00092840526,"about_ca_system_score_gemma":0.0014771038,"threshold_uncertainty_score":0.020651042},"labels":[],"label_agreement":null},{"id":"W2400495833","doi":"","title":"Composing Scalability for Transactions on Multicore Platforms.","year":2013,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Distributed and Parallel Computing Systems","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 Toronto","funders":"","keywords":"Multi-core processor; Scalability; Computer science; Distributed computing; Computer architecture; Parallel computing; Database","score_opus":0.05550861582558886,"score_gpt":0.28469288155287087,"score_spread":0.229184265727282,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2400495833","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":[0.12647398,0.0028622104,0.83925843,0.0015918345,0.00092785456,0.00050517014,0.00047098173,0.008802064,0.019107329],"genre_scores_gemma":[0.61990035,0.0009074069,0.36719665,0.00034682456,0.00049123075,0.00051090145,0.0013615895,0.0011638276,0.008121309],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9960532,0.0010389152,0.00039716627,0.00055621506,0.0013820067,0.0005724023],"domain_scores_gemma":[0.98996747,0.0033047174,0.0003384571,0.004390729,0.0013651982,0.00063351466],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0044501014,0.0010720826,0.0012501612,0.0007368455,0.0017079902,0.0023844573,0.0026508046,0.0012157406,0.004334752],"category_scores_gemma":[0.01310141,0.00094997435,0.0010898405,0.001028913,0.0015072433,0.005898638,0.0033807307,0.002621416,0.0013665148],"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":[0.0018650347,0.0007306879,0.010291702,0.0010829957,0.0007152211,0.0009832988,0.0015648303,0.23065268,0.08514099,0.18355414,0.045231245,0.43818712],"study_design_scores_gemma":[0.00017066314,0.0002846106,0.0012703472,0.000081470214,0.0001717717,0.00022978689,0.00035934278,0.77859634,0.029274996,0.17075674,0.018746251,0.000057634883],"about_ca_topic_score_codex":0.0033009816,"about_ca_topic_score_gemma":0.0053139864,"teacher_disagreement_score":0.0044501014,"about_ca_system_score_codex":0.0010818348,"about_ca_system_score_gemma":0.0019696204,"threshold_uncertainty_score":0.023534656},"labels":[],"label_agreement":null},{"id":"W2400761516","doi":"","title":"Towards Dynamic Green-Sizing for Database Servers.","year":2015,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Cloud Computing and Resource Management","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Server; Computer science; Frequency scaling; Memory management; Exploit; Workload; Database server; Power (physics); Dynamic demand; Database; Transaction processing; Operating system; Database transaction; Semiconductor memory","score_opus":0.07489147399189085,"score_gpt":0.3030708951274435,"score_spread":0.22817942113555267,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2400761516","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":[0.04580961,0.0021165505,0.93136406,0.0008195053,0.00016381103,0.000107175096,0.00016884744,0.002131717,0.017318647],"genre_scores_gemma":[0.614623,0.001005819,0.37342218,0.0004725941,0.00008202497,0.000094229006,0.0002862412,0.0005457956,0.009468121],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99964774,0.00005001058,0.000012777951,0.00007149611,0.00016998286,0.00004803186],"domain_scores_gemma":[0.9995946,0.00014224875,0.0000407349,0.000080933954,0.00010805431,0.000033407043],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00047221663,0.00076743535,0.00053795404,0.0005513064,0.00055917475,0.0014114798,0.0014835099,0.0003854945,0.0039728945],"category_scores_gemma":[0.0011836101,0.00033898803,0.00026143473,0.00066926243,0.00048230373,0.0018872406,0.00084258127,0.00075634633,0.00103343],"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":[0.00030868602,0.0003589224,0.0023340075,0.000339982,0.00007097162,0.00015493836,0.00024331785,0.31123266,0.12641458,0.05829107,0.012371398,0.4878795],"study_design_scores_gemma":[0.000041076,0.00010444156,0.0007912749,0.000027416323,0.000026737944,0.00015030365,0.00015082271,0.9059783,0.02709703,0.042661246,0.022947632,0.00002369126],"about_ca_topic_score_codex":0.0015578972,"about_ca_topic_score_gemma":0.0043827216,"teacher_disagreement_score":0.0039728945,"about_ca_system_score_codex":0.0008829962,"about_ca_system_score_gemma":0.0010432185,"threshold_uncertainty_score":0.013290644},"labels":[],"label_agreement":null},{"id":"W2784053634","doi":"10.1145/3164135.3164139","title":"The ubiquity of large graphs and surprising challenges of graph processing","year":2017,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":89,"is_retracted":false,"has_abstract":true,"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; Scalability; Suite; Visualization; Graph; Software; Data science; Call graph; Theoretical computer science; Data visualization; World Wide Web; Data mining; Programming language; Database","score_opus":0.07621659304314052,"score_gpt":0.3534706649777041,"score_spread":0.27725407193456353,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2784053634","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.54175943,0.023182487,0.35357445,0.0526655,0.00072672934,0.00036231132,0.0018969739,0.003624843,0.022207255],"genre_scores_gemma":[0.7660667,0.012304638,0.20947888,0.0037973549,0.00081540435,0.00028507021,0.0018639418,0.002015052,0.0033730036],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.981712,0.00960679,0.00071879313,0.0022336657,0.005285201,0.0004436028],"domain_scores_gemma":[0.79851323,0.17282747,0.006979656,0.010916785,0.0085304305,0.0022324221],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.016650068,0.0005647455,0.0006027501,0.00383631,0.0019432108,0.005126248,0.0018750937,0.0020010408,0.0025251808],"category_scores_gemma":[0.09797047,0.0009414402,0.00070686114,0.0051398478,0.0038713636,0.013913522,0.0030829683,0.0027056155,0.00076533423],"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":[0.00037940266,0.00017343843,0.0684664,0.007079237,0.00031238783,0.0030487694,0.091828816,0.011050892,0.01797997,0.10343374,0.059555523,0.6366914],"study_design_scores_gemma":[0.000050844617,0.00033568766,0.058322188,0.0017203998,0.00012435348,0.009872593,0.08603596,0.037051443,0.010269688,0.31714505,0.4786808,0.0003910153],"about_ca_topic_score_codex":0.0014195004,"about_ca_topic_score_gemma":0.0029274279,"teacher_disagreement_score":0.016650068,"about_ca_system_score_codex":0.0013299602,"about_ca_system_score_gemma":0.0010868686,"threshold_uncertainty_score":0.088055015},"labels":[],"label_agreement":null},{"id":"W2791526170","doi":"10.1145/3164135.3164147","title":"Bztree: a high-performance latch-free range index for non-volatile memory","year":2018,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":83,"is_retracted":false,"has_abstract":true,"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; Non-volatile memory; Block (permutation group theory); Embedded system; Throughput; Computer hardware; Tree (set theory); Semiconductor memory; Code (set theory); Data retention; Parallel computing; Operating system","score_opus":0.024349013921160394,"score_gpt":0.2591629562395728,"score_spread":0.2348139423184124,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2791526170","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":[0.10211343,0.008068644,0.8074611,0.00051357923,0.0006007947,0.00037901112,0.0032832497,0.05490519,0.022675041],"genre_scores_gemma":[0.5429686,0.0023322357,0.41308543,0.0006378849,0.00021445307,0.0006226764,0.011410766,0.003927645,0.024800362],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995858,0.000028586379,0.0000335848,0.000040532683,0.00025614904,0.00005530733],"domain_scores_gemma":[0.9994018,0.000095405136,0.00007582967,0.0001638327,0.00020185232,0.00006136617],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00026611178,0.00036020164,0.00039229618,0.0010443452,0.00051621103,0.0011831502,0.0019511488,0.0004372377,0.0056242216],"category_scores_gemma":[0.0010082782,0.00031900738,0.000247334,0.0013721433,0.00029028507,0.0023210228,0.0012947761,0.00065309094,0.0027030844],"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":[0.0016292457,0.00025554735,0.004006941,0.0010192015,0.00014629694,0.000510006,0.000395564,0.009983262,0.23651907,0.029464679,0.10342148,0.6126488],"study_design_scores_gemma":[0.00070206093,0.0017454068,0.0051034708,0.00018384894,0.00016499078,0.0016926908,0.00027817767,0.22746651,0.40816092,0.024687236,0.3294839,0.00033081925],"about_ca_topic_score_codex":0.0014116378,"about_ca_topic_score_gemma":0.0019002435,"teacher_disagreement_score":0.0056242216,"about_ca_system_score_codex":0.0005346498,"about_ca_system_score_gemma":0.00082502404,"threshold_uncertainty_score":0.018814921},"labels":[],"label_agreement":null},{"id":"W2889053599","doi":"10.5555/3236187.3269462","title":"Efficient construction of approximate ad-hoc ML models through materialization and reuse","year":2018,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Online analytical processing; Reuse; Dimension (graph theory); Cluster analysis; Data warehouse; Data mining; Construct (python library); Variety (cybernetics); Mixture model; Data modeling; Machine learning; Artificial intelligence; Database; Programming language","score_opus":0.04435581347020066,"score_gpt":0.26769465447678786,"score_spread":0.2233388410065872,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2889053599","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":[0.013759134,0.00012380215,0.9825075,0.00024825372,0.000012676904,0.00008296494,0.00013101815,0.0023571386,0.00077748415],"genre_scores_gemma":[0.2288459,0.00019853332,0.7668522,0.00021109308,0.00006104561,0.00024637682,0.0011871207,0.0008229409,0.0015747581],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9944278,0.0018889277,0.00041980195,0.0009655513,0.0018007179,0.00049721333],"domain_scores_gemma":[0.9821231,0.009989939,0.0009999617,0.005292078,0.0012315968,0.00036341546],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005564487,0.0021294847,0.0025678093,0.001993426,0.0013171071,0.0046274224,0.004590313,0.0019059279,0.0033487598],"category_scores_gemma":[0.026633224,0.0018187516,0.0031465027,0.0032994747,0.0022787722,0.008489148,0.0066871885,0.0037005646,0.0016268742],"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":[0.00016978827,0.00019492739,0.0023633768,0.00012327352,0.00011078065,0.00020262609,0.00036859015,0.8129876,0.002688573,0.03874538,0.0039741923,0.13807087],"study_design_scores_gemma":[0.000012505122,0.000018548486,0.000057628542,0.0000040647233,0.000010354343,0.000027586631,0.000051890944,0.97926694,0.0010922985,0.018641096,0.00080889306,0.000008232639],"about_ca_topic_score_codex":0.009461256,"about_ca_topic_score_gemma":0.011842938,"teacher_disagreement_score":0.009461256,"about_ca_system_score_codex":0.002661781,"about_ca_system_score_gemma":0.0038007316,"threshold_uncertainty_score":0.029428124},"labels":[],"label_agreement":null},{"id":"W3204354173","doi":"","title":"Enabling NUMA-aware Main Memory Spatial Join Processing: An Experimental Study.","year":2020,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":1,"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 New Brunswick","funders":"","keywords":"Computer science; Join (topology); Parallel computing; Computer architecture","score_opus":0.08098202661924112,"score_gpt":0.3075131336643204,"score_spread":0.2265311070450793,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3204354173","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":[0.92650664,0.0014961886,0.049681313,0.0005412762,0.00039384916,0.00023653034,0.0007536702,0.0068169613,0.013573582],"genre_scores_gemma":[0.9594269,0.00024072807,0.03630263,0.000118102806,0.000046158908,0.000097944,0.0004826852,0.00026441074,0.0030203394],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99904364,0.00019061324,0.000044017754,0.0001789809,0.00037111886,0.00017151977],"domain_scores_gemma":[0.9967919,0.0013797256,0.00018143139,0.0008735945,0.00049667014,0.00027660918],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00095951464,0.0005570193,0.00058331934,0.00038037467,0.0009271509,0.0014601592,0.002264723,0.0007407466,0.005949197],"category_scores_gemma":[0.004254949,0.00029083053,0.00021958291,0.0008764104,0.00065119716,0.002041416,0.0010636532,0.00089733215,0.0012236299],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.015300618,0.0070385775,0.00943735,0.0015852869,0.00041026573,0.0007484042,0.0008917885,0.07948508,0.6688377,0.01288812,0.018690204,0.18468662],"study_design_scores_gemma":[0.00085425767,0.0043096496,0.007810457,0.00004895022,0.00025285437,0.0004154318,0.0007222652,0.5452675,0.4193237,0.0061873193,0.014713295,0.00009423065],"about_ca_topic_score_codex":0.0025578325,"about_ca_topic_score_gemma":0.0023929693,"teacher_disagreement_score":0.005949197,"about_ca_system_score_codex":0.0005484652,"about_ca_system_score_gemma":0.00095204834,"threshold_uncertainty_score":0.01990205},"labels":[],"label_agreement":null},{"id":"W3204354299","doi":"","title":"A Data Discovery Platform Empowered by Knowledge GraphTechnologies: Challenges and Opportunities.","year":2021,"lang":"en","type":"article","venue":"Very Large Data Bases","topic":"Semantic Web and Ontologies","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":"Concordia University","funders":"","keywords":"Knowledge extraction; Computer science; Data science; Knowledge management; Data mining","score_opus":0.22418996122957144,"score_gpt":0.32163684355728805,"score_spread":0.09744688232771662,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3204354299","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":[0.0066812676,0.0017465435,0.9553997,0.010260713,0.0005118823,0.0002858862,0.0026211997,0.012875864,0.009616979],"genre_scores_gemma":[0.048484582,0.0021642984,0.93125105,0.0019046431,0.00022113821,0.0002742599,0.009752192,0.001114258,0.004833531],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9951439,0.0017991074,0.0005061556,0.00065861666,0.0016653057,0.00022684416],"domain_scores_gemma":[0.98116106,0.007809654,0.00067077554,0.006627701,0.002302397,0.0014283698],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011214727,0.0006647691,0.0011342695,0.005595871,0.002279375,0.009866868,0.0040002954,0.0025360787,0.0040123737],"category_scores_gemma":[0.019309977,0.00093716144,0.0011827147,0.0073228204,0.0021718908,0.020570645,0.008293849,0.0036486313,0.0034789233],"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":[0.0003453619,0.0006385803,0.0034860764,0.00096165156,0.00032663083,0.00072640856,0.000825921,0.008588339,0.009651171,0.44174218,0.095716685,0.43699104],"study_design_scores_gemma":[0.000053388056,0.00006572019,0.00061280245,0.00034951995,0.00009962491,0.0003862936,0.00047182315,0.07500941,0.006944548,0.6422229,0.27370587,0.000078125675],"about_ca_topic_score_codex":0.0036921198,"about_ca_topic_score_gemma":0.0075094295,"teacher_disagreement_score":0.011214727,"about_ca_system_score_codex":0.0013021481,"about_ca_system_score_gemma":0.0050603515,"threshold_uncertainty_score":0.0593099},"labels":[],"label_agreement":null}]}