{"meta":{"query_hash":"183d8fa8b2bb","filters":{"venue":"IEEE Transactions on Big Data"},"cohort_total":22,"direct_labels_cover":0,"predictions_cover":22,"exported":22,"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/183d8fa8b2bb","api":"https://metacan.xera.ac/api/v1/cohort?venue=IEEE+Transactions+on+Big+Data"},"results":[{"id":"W2610514569","doi":"10.1109/tbdata.2017.2701816","title":"PPHOPCM: Privacy-Preserving High-Order Possibilistic c-Means Algorithm for Big Data Clustering with Cloud Computing","year":2017,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":123,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"St. Francis Xavier University","funders":"","keywords":"Cluster analysis; Computer science; Big data; Cloud computing; Data mining; Fuzzy clustering; Scheme (mathematics); Encryption; CURE data clustering algorithm; Clustering high-dimensional data; Algorithm; Artificial intelligence; Mathematics","score_opus":0.1454192020572554,"score_gpt":0.3165568846312504,"score_spread":0.17113768257399503,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2610514569","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":[0.0066608675,0.00012095397,0.99179804,0.00015581936,0.000030243798,0.00006342822,0.000059190545,0.00053606473,0.0005753694],"genre_scores_gemma":[0.29160276,0.00019396048,0.70521307,0.0002617986,0.00005840112,0.00025477735,0.0003470405,0.00014337664,0.0019248503],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983021,0.00039442317,0.000090467045,0.00032860858,0.000725151,0.00015934209],"domain_scores_gemma":[0.99846554,0.00043353572,0.00014772666,0.0004917432,0.00035448666,0.000106992884],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013772807,0.0007341663,0.001004658,0.0010354266,0.0016043643,0.0013330049,0.0025643725,0.0012803619,0.0014780711],"category_scores_gemma":[0.004182975,0.00044553363,0.0011723137,0.0019189204,0.0010758548,0.0019104177,0.0023250454,0.002060341,0.00047263788],"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":[0.0005114248,0.00021557377,0.0017132121,0.00024944922,0.00020963265,0.0002746088,0.00037964384,0.55189526,0.01711908,0.06774383,0.0089333085,0.350755],"study_design_scores_gemma":[0.000020774658,0.00003653839,0.00014857468,0.000005522251,0.000007811357,0.00009630439,0.000025851323,0.9723141,0.004068251,0.021655012,0.0016042019,0.000017061671],"about_ca_topic_score_codex":0.00663188,"about_ca_topic_score_gemma":0.0063149356,"teacher_disagreement_score":0.00663188,"about_ca_system_score_codex":0.0015733264,"about_ca_system_score_gemma":0.0038274897,"threshold_uncertainty_score":0.013186574},"labels":[],"label_agreement":null},{"id":"W2731419512","doi":"10.1109/tbdata.2017.2721444","title":"Achieving Efficient and Privacy-Preserving Cross-Domain Big Data Deduplication in Cloud","year":2017,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Cloud Data Security Solutions","field":"Computer Science","cited_by":58,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Data deduplication; Computer science; Cloud computing; Big data; Homomorphic encryption; Encryption; Cloud storage; Computer security; Information privacy; Domain (mathematical analysis); Data mining; Operating system","score_opus":0.16154579015690074,"score_gpt":0.3454990075471699,"score_spread":0.18395321739026915,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2731419512","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":[0.20699951,0.0013237421,0.785416,0.00075373787,0.00017391576,0.00033857496,0.00039453685,0.0011115085,0.0034884375],"genre_scores_gemma":[0.8895173,0.0003066808,0.10835647,0.00017087192,0.00004092448,0.00006965965,0.0002295709,0.000027582088,0.001280816],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9985084,0.00030964706,0.00017041399,0.00032871412,0.00043971834,0.00024304382],"domain_scores_gemma":[0.99471587,0.0009908589,0.00050031545,0.0027965938,0.00078775926,0.00020862953],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024698256,0.00037816787,0.00071443775,0.00057983113,0.0011966262,0.0016394007,0.001415607,0.00072979054,0.0007309957],"category_scores_gemma":[0.004557864,0.00033418427,0.0005488979,0.001250316,0.00079479854,0.0030295684,0.0033463417,0.0008577761,0.00031330445],"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":[0.0025088282,0.00086884556,0.018144751,0.00091719034,0.00039742966,0.001776812,0.0013967348,0.24621092,0.17838521,0.10332015,0.017200567,0.42887264],"study_design_scores_gemma":[0.00012754205,0.00037330828,0.0026791797,0.000045962413,0.000091318034,0.0020415762,0.0006583162,0.79794466,0.15228686,0.032476094,0.011181422,0.000093746414],"about_ca_topic_score_codex":0.00072732434,"about_ca_topic_score_gemma":0.0009689494,"teacher_disagreement_score":0.0024698256,"about_ca_system_score_codex":0.0006246727,"about_ca_system_score_gemma":0.001814253,"threshold_uncertainty_score":0.013061821},"labels":[],"label_agreement":null},{"id":"W2753315806","doi":"10.1109/tbdata.2017.2742530","title":"Euler Clustering on Large-Scale Dataset","year":2017,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"National Key Research and Development Program of China; Guangdong Science and Technology Department; Sun Yat-sen University; National Natural Science Foundation of China","keywords":"Cluster analysis; Euler's formula; Kernel (algebra); Spectral clustering; Mathematics; Variable kernel density estimation; Algorithm; Pattern recognition (psychology); Computer science; Kernel method; Artificial intelligence; Discrete mathematics; Mathematical analysis; Support vector machine","score_opus":0.129188872121244,"score_gpt":0.3208625784651777,"score_spread":0.1916737063439337,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2753315806","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":[0.07280857,0.0025744312,0.889791,0.0015909952,0.0009778094,0.00054197834,0.0139322765,0.013407832,0.004375204],"genre_scores_gemma":[0.19961864,0.00093615113,0.73096657,0.0007182951,0.00024191086,0.00067259587,0.06153864,0.0010524386,0.0042548506],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99498737,0.0010879346,0.00046953475,0.0018536098,0.0012686074,0.00033300652],"domain_scores_gemma":[0.9924655,0.0015346914,0.00044477172,0.003025896,0.0022574903,0.0002717482],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0055030305,0.0018385226,0.0016591944,0.00416259,0.0018823437,0.0025194115,0.0030622198,0.002520881,0.002404041],"category_scores_gemma":[0.01977602,0.0004751506,0.0021822439,0.004460905,0.0010865601,0.0035856445,0.0029579266,0.0026146201,0.00252265],"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":[0.0010522248,0.0005698679,0.015086383,0.0015611215,0.00086036406,0.00087429746,0.00063828717,0.30745742,0.015611745,0.029033495,0.14416155,0.48309326],"study_design_scores_gemma":[0.000060979146,0.00009808082,0.005879808,0.00010595057,0.00006466235,0.00035800954,0.0003935019,0.91983366,0.009294522,0.036136653,0.027675547,0.00009853871],"about_ca_topic_score_codex":0.008343728,"about_ca_topic_score_gemma":0.014323688,"teacher_disagreement_score":0.008343728,"about_ca_system_score_codex":0.0016338014,"about_ca_system_score_gemma":0.0023131114,"threshold_uncertainty_score":0.02910316},"labels":[],"label_agreement":null},{"id":"W2920772632","doi":"10.1109/tbdata.2019.2903092","title":"Incremental Deep Computation Model for Wireless Big Data Feature Learning","year":2019,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Machine Learning and ELM","field":"Computer Science","cited_by":31,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"St. Francis Xavier University","funders":"","keywords":"Computer science; Big data; Artificial intelligence; Deep learning; Machine learning; Wireless network; Computation; Wireless; Feature (linguistics); Data modeling; Algorithm; Data mining","score_opus":0.1011934834914356,"score_gpt":0.3053111633478779,"score_spread":0.2041176798564423,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2920772632","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":[0.01824291,0.0003054935,0.9788831,0.00028477432,0.0000690097,0.000039711904,0.00010401468,0.00051097875,0.0015600753],"genre_scores_gemma":[0.809831,0.0006546335,0.18302563,0.00027751987,0.00008569726,0.00023424295,0.000527632,0.00009698648,0.005266538],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99969363,0.000042260224,0.000019007965,0.000087201544,0.00010178262,0.000056263278],"domain_scores_gemma":[0.99954396,0.00014962076,0.00004329224,0.00007085258,0.00016223808,0.000029936338],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005465632,0.0006144201,0.00056892337,0.000426833,0.00032360057,0.0006704965,0.0017275162,0.0005683416,0.001519229],"category_scores_gemma":[0.001875266,0.00030692265,0.0006251965,0.0006036498,0.00056426344,0.0017590199,0.0008474345,0.0015550532,0.0003227625],"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":[0.00012111915,0.000103248545,0.0024911845,0.00009489278,0.000062689076,0.00014471986,0.000101016594,0.7866994,0.0044615115,0.03541032,0.0040793186,0.16623059],"study_design_scores_gemma":[0.0000018918859,0.000010158586,0.00007688801,0.0000018671772,0.00000507808,0.000012020005,0.0000029016455,0.9953524,0.00043887793,0.0037232758,0.0003719677,0.000002685601],"about_ca_topic_score_codex":0.007901137,"about_ca_topic_score_gemma":0.009478655,"teacher_disagreement_score":0.007901137,"about_ca_system_score_codex":0.0008694557,"about_ca_system_score_gemma":0.0012722409,"threshold_uncertainty_score":0.015710294},"labels":[],"label_agreement":null},{"id":"W2923023146","doi":"10.1109/tbdata.2019.2907624","title":"Hierarchical Density-Based Clustering Using MapReduce","year":2019,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":55,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Fundação de Amparo à Pesquisa do Estado de Minas Gerais; Conselho Nacional de Desenvolvimento Científico e Tecnológico; Natural Sciences and Engineering Research Council of Canada; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior; Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Computer science; Cluster analysis; Scalability; Automatic summarization; Correlation clustering; Hierarchical clustering; CURE data clustering algorithm; Data mining; Canopy clustering algorithm; Constrained clustering; Artificial intelligence; Database","score_opus":0.1308743166587772,"score_gpt":0.3353534078399405,"score_spread":0.2044790911811633,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2923023146","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":[0.0083112195,0.00022392608,0.97952205,0.00018376303,0.00005726308,0.00018962704,0.0006036907,0.008026351,0.0028821218],"genre_scores_gemma":[0.22214764,0.00043100645,0.76855195,0.0002062581,0.000071874565,0.00052930485,0.0036675017,0.0008128196,0.0035816047],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9986765,0.00020312484,0.00006268795,0.00025830374,0.000673063,0.00012616446],"domain_scores_gemma":[0.9989784,0.00020314344,0.000055294597,0.00030946534,0.00037179954,0.00008202947],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008565807,0.0009837473,0.0011493777,0.0013896544,0.0012649423,0.0012851341,0.002279169,0.00062480516,0.0023056823],"category_scores_gemma":[0.002722541,0.0005978104,0.0013282425,0.0024006115,0.000498505,0.0016837929,0.0019854116,0.0011610343,0.0014940738],"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":[0.00035229264,0.00028244813,0.0023695792,0.00053776853,0.00028968882,0.00029899,0.0006579127,0.5395772,0.02003823,0.037400436,0.04763719,0.3505583],"study_design_scores_gemma":[0.00004875009,0.000035702265,0.0007136447,0.000011392667,0.000021921041,0.00011559032,0.00014544571,0.94820964,0.0081555415,0.028044434,0.014454353,0.000043592783],"about_ca_topic_score_codex":0.016656213,"about_ca_topic_score_gemma":0.01581814,"teacher_disagreement_score":0.016656213,"about_ca_system_score_codex":0.0012099238,"about_ca_system_score_gemma":0.0025483766,"threshold_uncertainty_score":0.033118546},"labels":[],"label_agreement":null},{"id":"W2982169658","doi":"10.1109/tbdata.2019.2948613","title":"PES: Priority Edge Sampling in Streaming Triangle Estimation","year":2019,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","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 Windsor","funders":"","keywords":"Computer science; Sampling (signal processing); Estimator; Enhanced Data Rates for GSM Evolution; Streaming algorithm; Metric (unit); Cluster analysis; Algorithm; Variance (accounting); Clustering coefficient; Theoretical computer science; Mathematics; Statistics; Artificial intelligence; Upper and lower bounds","score_opus":0.08562440375598687,"score_gpt":0.32429078161761027,"score_spread":0.2386663778616234,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2982169658","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":[0.010351597,0.00013422804,0.98776025,0.00007649899,0.000049337417,0.000052643267,0.00013878783,0.00081746996,0.0006191451],"genre_scores_gemma":[0.2827477,0.0002868348,0.71224743,0.00016499711,0.00013964955,0.00019540483,0.0012359209,0.00036249997,0.0026196816],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995951,0.00009110968,0.000028754253,0.00010234733,0.00013728385,0.000045504883],"domain_scores_gemma":[0.99862015,0.00066197116,0.0000975861,0.00024276222,0.00029306964,0.000084385465],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00083153177,0.0006060625,0.00070631533,0.0008714329,0.00037998884,0.00079935044,0.0016888871,0.00081750815,0.0031289493],"category_scores_gemma":[0.0063292338,0.0003641038,0.00047953855,0.0011133421,0.00045447252,0.001637394,0.0013162518,0.0008847597,0.00089975796],"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":[0.000662074,0.00014170243,0.00668468,0.00021200141,0.000106089035,0.00022359611,0.00023734858,0.35897887,0.0150519665,0.0344077,0.009995475,0.5732985],"study_design_scores_gemma":[0.000014225058,0.00002089031,0.00019719343,0.0000054556936,0.000004960086,0.00004357503,0.000018306231,0.9905567,0.00227617,0.0057595344,0.0010967782,0.000006169718],"about_ca_topic_score_codex":0.003997435,"about_ca_topic_score_gemma":0.005393278,"teacher_disagreement_score":0.003997435,"about_ca_system_score_codex":0.00042205944,"about_ca_system_score_gemma":0.0008241027,"threshold_uncertainty_score":0.0104673505},"labels":[],"label_agreement":null},{"id":"W3005586683","doi":"10.1109/tbdata.2020.2972887","title":"Least Squares Approximation via Sparse Subsampled Randomized Hadamard Transform","year":2020,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Sparse and Compressive Sensing Techniques","field":"Engineering","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 Alberta","funders":"","keywords":"Hadamard transform; Computer science; Algorithm; Subspace topology; Overdetermined system; Randomized algorithm; Embedding; Mathematics; Artificial intelligence; Applied mathematics","score_opus":0.0974922001653938,"score_gpt":0.25270262016454015,"score_spread":0.15521041999914637,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3005586683","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":[0.003483697,0.00011004333,0.99558014,0.00009987648,0.000028211916,0.000016123642,0.000036601527,0.00024004062,0.0004052967],"genre_scores_gemma":[0.2019349,0.0003116222,0.7950767,0.00020885242,0.00010462373,0.0001717521,0.0003629581,0.00012836972,0.0017002593],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9986739,0.0004931153,0.00005989268,0.00020815815,0.00047747995,0.000087457716],"domain_scores_gemma":[0.9972025,0.0017230358,0.00022762713,0.00042998875,0.0003436441,0.000073156276],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015489049,0.00084412465,0.0011215621,0.00054484647,0.00036145066,0.00085935777,0.0010963306,0.0011911608,0.001789597],"category_scores_gemma":[0.008678102,0.0004420758,0.0008539578,0.0008760785,0.0010439929,0.0016152266,0.0013528481,0.001991628,0.0006850192],"study_design_candidate":"theoretical_or_conceptual","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":[0.00022786869,0.00007952553,0.00087784085,0.00018976585,0.00009948496,0.0001478944,0.00011134926,0.8090063,0.012081444,0.06286389,0.0037600829,0.11055454],"study_design_scores_gemma":[0.000012966755,0.000023958253,0.000049015704,0.0000047306394,0.0000037987295,0.000032693435,0.000008006731,0.99177015,0.0013512487,0.0061240834,0.0006126432,0.0000067445367],"about_ca_topic_score_codex":0.0028879247,"about_ca_topic_score_gemma":0.0033121363,"teacher_disagreement_score":0.0028879247,"about_ca_system_score_codex":0.00047786458,"about_ca_system_score_gemma":0.00141725,"threshold_uncertainty_score":0.008191466},"labels":[],"label_agreement":null},{"id":"W3033628195","doi":"10.1109/tbdata.2020.2998770","title":"Dynamic Entity-Based Named Entity Recognition Under Unconstrained Tagging Schemes","year":2020,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Topic Modeling","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"St. Francis Xavier University","funders":"National Natural Science Foundation of China","keywords":"Computer science; Named-entity recognition; Natural language processing; Sentence; Artificial intelligence; Word (group theory); Entity linking; Language model; Precision and recall; Named entity; Task (project management); Knowledge base; Linguistics","score_opus":0.15396775093685824,"score_gpt":0.2866244407588694,"score_spread":0.13265668982201115,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3033628195","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":[0.022297017,0.00044935342,0.9721168,0.00018145116,0.000078425415,0.0000731808,0.00040359012,0.002621623,0.0017785871],"genre_scores_gemma":[0.4020863,0.0010658387,0.57872856,0.00040398445,0.0001171558,0.00022871663,0.006466553,0.00033089795,0.01057207],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99779403,0.0005097963,0.0002137182,0.0010212352,0.0003226384,0.0001385832],"domain_scores_gemma":[0.9965301,0.0010559794,0.00033857967,0.0014086957,0.00058785063,0.00007884087],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025215717,0.0011937324,0.0010258075,0.0014394568,0.0007237849,0.001801053,0.0022127735,0.0016776618,0.0014308508],"category_scores_gemma":[0.0056272894,0.0005242612,0.0012396672,0.0019563604,0.0008225462,0.009331742,0.0024102277,0.0018675979,0.0022774132],"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":[0.00047074963,0.00027470972,0.005512995,0.0003539926,0.000259734,0.0010689896,0.000722148,0.30737412,0.04089523,0.04587005,0.010546132,0.58665115],"study_design_scores_gemma":[0.000015038031,0.00007374464,0.001211498,0.000031599066,0.00007658976,0.0003326508,0.00010105183,0.948422,0.020062424,0.02021273,0.009387717,0.000072963936],"about_ca_topic_score_codex":0.004072802,"about_ca_topic_score_gemma":0.0052909893,"teacher_disagreement_score":0.004072802,"about_ca_system_score_codex":0.00069233385,"about_ca_system_score_gemma":0.00092821725,"threshold_uncertainty_score":0.013335526},"labels":[],"label_agreement":null},{"id":"W3089041741","doi":"10.1109/tbdata.2020.3026318","title":"On Security of an Identity-Based Dynamic Data Auditing Protocol for Big Data Storage","year":2020,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Cloud Data Security Solutions","field":"Computer Science","cited_by":9,"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 New Brunswick","funders":"Natural Science Foundation of Hunan Province; National Natural Science Foundation of China","keywords":"Computer science; Big data; Protocol (science); Audit; Computer security; Cryptographic protocol; Cryptography; Data mining","score_opus":0.22039325091568981,"score_gpt":0.37304155498988695,"score_spread":0.15264830407419713,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3089041741","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":[0.038208727,0.0017138784,0.926151,0.007196273,0.0008222181,0.00064781314,0.00019923958,0.0007666152,0.024294194],"genre_scores_gemma":[0.813696,0.0021660144,0.1673488,0.0020541605,0.00047762392,0.0008655006,0.00023550505,0.00017685676,0.012979536],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98452383,0.005116978,0.001381374,0.001490294,0.0060866424,0.0014008757],"domain_scores_gemma":[0.9684691,0.013715228,0.0023550477,0.010278522,0.0044807224,0.00070131803],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011233777,0.00076337956,0.0011434768,0.0018115193,0.0035318043,0.004983234,0.0026084075,0.003200937,0.0023247278],"category_scores_gemma":[0.028052835,0.0008085284,0.0009581507,0.0019029694,0.005398785,0.018558,0.00875642,0.005925449,0.0009777467],"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":[0.00040911537,0.00009006678,0.00096996163,0.0002761397,0.000051504325,0.00049987854,0.0015229762,0.008963812,0.008748633,0.9342877,0.004091983,0.040088125],"study_design_scores_gemma":[0.0001946226,0.00050761725,0.00063254376,0.00041199249,0.00017618472,0.0022055262,0.0010802045,0.27063584,0.06449901,0.5748559,0.08445006,0.00035053474],"about_ca_topic_score_codex":0.0010362817,"about_ca_topic_score_gemma":0.00039371915,"teacher_disagreement_score":0.011233777,"about_ca_system_score_codex":0.0027205,"about_ca_system_score_gemma":0.0042658597,"threshold_uncertainty_score":0.059410572},"labels":[],"label_agreement":null},{"id":"W3173588045","doi":"10.1109/tbdata.2021.3089986","title":"SOREL: Efficient and Secure ORE-based Range Query over Outsourced Data","year":2021,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Cryptography and Data Security","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 New Brunswick","funders":"","keywords":"Computer science; Scalability; Encryption; Cloud computing; Range query (database); Outsourcing; Scheme (mathematics); Big data; Query optimization; Sargable; Database; Distributed computing; Computer security; Data mining; Web search query; Operating system; Information retrieval; Search engine","score_opus":0.08136779334271013,"score_gpt":0.27963554450835715,"score_spread":0.19826775116564702,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3173588045","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":[0.100972876,0.0013898859,0.8851867,0.00061139,0.00011823277,0.0004499845,0.0005622723,0.0032788594,0.0074298447],"genre_scores_gemma":[0.8246819,0.00070251466,0.16766404,0.0003250054,0.00009691065,0.00017299486,0.0007806113,0.00012845849,0.005447646],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99731785,0.00042875632,0.000268591,0.00024914864,0.0012757768,0.0004599134],"domain_scores_gemma":[0.99687344,0.0005824864,0.00039295756,0.001572997,0.0004480835,0.00012999942],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0018359587,0.00044535956,0.0012605187,0.00084520556,0.0006953113,0.001367863,0.0016114671,0.00068469637,0.0022897644],"category_scores_gemma":[0.0035252823,0.00031974423,0.0007482861,0.0013032666,0.0010587999,0.005856037,0.00405325,0.0011904371,0.00082752627],"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":[0.0033258577,0.00071961404,0.0043351823,0.00095987745,0.0002695294,0.0013820641,0.0014245966,0.0889556,0.15341033,0.19643271,0.021757398,0.52702725],"study_design_scores_gemma":[0.0004375731,0.0011303324,0.0019514817,0.00008555227,0.00012696245,0.0023378474,0.0006187803,0.7703734,0.11920422,0.06757429,0.03594564,0.0002138881],"about_ca_topic_score_codex":0.0009486538,"about_ca_topic_score_gemma":0.00088511565,"teacher_disagreement_score":0.0022897644,"about_ca_system_score_codex":0.00082496385,"about_ca_system_score_gemma":0.0016384225,"threshold_uncertainty_score":0.009709597},"labels":[],"label_agreement":null},{"id":"W4286656147","doi":"10.1109/tbdata.2022.3186857","title":"Efficient Learned Spatial Index With Interpolation Function Based Learned Model","year":2022,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":12,"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 New Brunswick","funders":"","keywords":"Computer science; Leverage (statistics); Spatial database; Data mining; Spatial analysis; Spatial query; Index (typography); Overhead (engineering); Interpolation (computer graphics); Multivariate interpolation; Function (biology); Machine learning; Artificial intelligence; Information retrieval; Sargable; Web search query; Bilinear interpolation; Mathematics","score_opus":0.08700016958677916,"score_gpt":0.2643861053845783,"score_spread":0.17738593579779915,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4286656147","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":[0.047436558,0.00081880635,0.94240725,0.00034375268,0.00007100628,0.00010137102,0.0011246451,0.00436884,0.0033277355],"genre_scores_gemma":[0.67958385,0.0010351791,0.30721098,0.0002595745,0.00010199863,0.00025556676,0.0047150804,0.0003460663,0.0064916653],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992341,0.0000807385,0.000057316855,0.00018547563,0.00034335058,0.000099128],"domain_scores_gemma":[0.9988945,0.0003197611,0.000088194996,0.00031851936,0.0003196297,0.0000593547],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005431248,0.0008841226,0.0012721283,0.00087588455,0.00043879746,0.0015455832,0.0021384393,0.0008407664,0.0020238948],"category_scores_gemma":[0.0034898724,0.00037210787,0.00090842525,0.0022065411,0.00049054285,0.0044937227,0.0016383382,0.001298862,0.0009831383],"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":[0.0002786351,0.00021093566,0.0058471393,0.00018085232,0.000069278234,0.00024626424,0.00017760127,0.7080814,0.009042935,0.018486306,0.010061802,0.24731684],"study_design_scores_gemma":[0.000005001153,0.000020117393,0.00015053824,0.0000022901752,0.0000054826596,0.000037775517,0.000013037326,0.9959694,0.0011812521,0.0019181056,0.0006903001,0.000006692008],"about_ca_topic_score_codex":0.015455055,"about_ca_topic_score_gemma":0.011722804,"teacher_disagreement_score":0.015455055,"about_ca_system_score_codex":0.0012003062,"about_ca_system_score_gemma":0.0022026522,"threshold_uncertainty_score":0.030730188},"labels":[],"label_agreement":null},{"id":"W4323896842","doi":"10.1109/tbdata.2023.3255003","title":"Approximate Clustering Ensemble Method for Big Data","year":2023,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":31,"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":"National Natural Science Foundation of China","keywords":"Cluster analysis; Computer science; Disjoint sets; Big data; Correlation clustering; Data mining; Component (thermodynamics); Cluster (spacecraft); Artificial intelligence; Mathematics; Combinatorics","score_opus":0.3223194882196442,"score_gpt":0.40855782386772044,"score_spread":0.08623833564807626,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4323896842","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":[0.0031196969,0.0003664529,0.9942153,0.000123945,0.0000798064,0.00004489962,0.00013835046,0.0009362323,0.00097533205],"genre_scores_gemma":[0.21615851,0.00082984153,0.7743323,0.0003058753,0.00036167452,0.00047192012,0.00200196,0.0005813942,0.0049564647],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9977398,0.0005044582,0.000100516416,0.00056045054,0.0009542068,0.00014058096],"domain_scores_gemma":[0.9975706,0.0006663449,0.00017333413,0.0006122164,0.00084850995,0.00012896463],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00227493,0.0012398384,0.00194772,0.0020161567,0.0011663318,0.0016603841,0.0028685578,0.0013274091,0.002977135],"category_scores_gemma":[0.0063682,0.00047465312,0.0015429836,0.00296577,0.0006738655,0.0028853086,0.00205053,0.0021511079,0.0014916344],"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":[0.00022873803,0.00008809231,0.0020339806,0.00020360801,0.00031967097,0.00013896347,0.00018597495,0.70526505,0.00422882,0.047043182,0.011025988,0.22923793],"study_design_scores_gemma":[0.000005956249,0.000013225837,0.00018107108,0.000005652746,0.000012451001,0.000021358972,0.000017002018,0.9845483,0.0007616364,0.012343407,0.0020805753,0.000009413166],"about_ca_topic_score_codex":0.006591692,"about_ca_topic_score_gemma":0.006989103,"teacher_disagreement_score":0.006591692,"about_ca_system_score_codex":0.0013264163,"about_ca_system_score_gemma":0.0014711154,"threshold_uncertainty_score":0.013106644},"labels":[],"label_agreement":null},{"id":"W4387757671","doi":"10.1109/tbdata.2023.3325746","title":"A Federated Convolution Transformer for Fake News Detection","year":2023,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Misinformation and Its Impacts","field":"Social Sciences","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":"Brandon University","funders":"","keywords":"Computer science; Cluster analysis; Federated learning; Embedding; Data mining; The Internet; World Wide Web; Machine learning; Artificial intelligence","score_opus":0.2312254761613512,"score_gpt":0.37149474252581516,"score_spread":0.14026926636446396,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4387757671","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":[0.03946567,0.0002317588,0.954169,0.00020225815,0.000072849536,0.00006645138,0.00015412793,0.0038159962,0.0018219969],"genre_scores_gemma":[0.77677244,0.0002525246,0.2164372,0.00020229054,0.000100071105,0.00006398396,0.00055683503,0.000117924945,0.0054967273],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99882764,0.00024759604,0.00006511319,0.00028803851,0.00040670723,0.00016484741],"domain_scores_gemma":[0.9982956,0.00047107076,0.0002158219,0.0004950107,0.00041434247,0.00010802756],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014793208,0.0008675911,0.0011456667,0.00163078,0.00064653903,0.001388692,0.001415442,0.0011634473,0.002349385],"category_scores_gemma":[0.004313795,0.0004102556,0.0009151267,0.0012043156,0.0008250614,0.002989095,0.0014143116,0.001223908,0.0015362537],"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":[0.0010897832,0.0005604595,0.008779252,0.00013253329,0.00022086091,0.00049381104,0.0003101427,0.10950439,0.034652483,0.012457549,0.0070051993,0.8247935],"study_design_scores_gemma":[0.000009888587,0.000050379083,0.0007223438,0.0000050602766,0.000020623113,0.0002122127,0.000034441826,0.9809713,0.013427615,0.0034500065,0.0010810309,0.000015044682],"about_ca_topic_score_codex":0.004602882,"about_ca_topic_score_gemma":0.0047500143,"teacher_disagreement_score":0.004602882,"about_ca_system_score_codex":0.0010485542,"about_ca_system_score_gemma":0.0010764074,"threshold_uncertainty_score":0.009152174},"labels":[],"label_agreement":null},{"id":"W4390738572","doi":"10.1109/tbdata.2024.3352978","title":"GAT-COBO: Cost-Sensitive Graph Neural Network for Telecom Fraud Detection","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":33,"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":"National Natural Science Foundation of China","keywords":"Computer science; Graph; Artificial intelligence; Machine learning; Classifier (UML); Theoretical computer science","score_opus":0.08922606455742202,"score_gpt":0.3055968200315244,"score_spread":0.21637075547410237,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4390738572","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":[0.049331296,0.0034464824,0.9303028,0.0024076793,0.0005740049,0.00033206714,0.00094203884,0.0046940865,0.007969588],"genre_scores_gemma":[0.7263609,0.0016134788,0.2505314,0.0014285947,0.00033298062,0.0005084677,0.0027174782,0.00027080093,0.016235797],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994505,0.00016443204,0.000027475344,0.00012909854,0.00012890699,0.000099581775],"domain_scores_gemma":[0.9990803,0.00034445376,0.0001195496,0.000085073836,0.0003098052,0.000060833514],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00109233,0.0015247551,0.0014304925,0.001902465,0.00075514003,0.0011182463,0.0024777374,0.0019961342,0.0030507052],"category_scores_gemma":[0.0033835035,0.0005227801,0.000843934,0.0017247201,0.00060972816,0.0014338237,0.0010490572,0.00179774,0.0007264346],"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":[0.00046552694,0.00042743317,0.0033921835,0.00019935981,0.00025404585,0.00013308373,0.000044608292,0.6150465,0.002821215,0.009797115,0.019462567,0.34795633],"study_design_scores_gemma":[0.000008869451,0.000015744066,0.000190461,0.0000060868742,0.000010141816,0.000010605236,0.0000040849795,0.9967379,0.0003365071,0.0021262735,0.000548518,0.000004701496],"about_ca_topic_score_codex":0.025531504,"about_ca_topic_score_gemma":0.029229239,"teacher_disagreement_score":0.025531504,"about_ca_system_score_codex":0.0021528106,"about_ca_system_score_gemma":0.0021013413,"threshold_uncertainty_score":0.050765753},"labels":[],"label_agreement":null},{"id":"W4398152538","doi":"10.1109/tbdata.2024.3403390","title":"Accelerating Blockchain-Enabled Federated Learning With Clustered Clients","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Basic and Applied Basic Research Foundation of Guangdong Province; National Natural Science Foundation of China","keywords":"Computer science; Single point of failure; Blockchain; Asynchronous communication; Distributed computing; Enhanced Data Rates for GSM Evolution; Cluster analysis; Distributed learning; Edge device; Convergence (economics); Artificial intelligence; Machine learning; Data mining; Computer security; Computer network","score_opus":0.10089906378633208,"score_gpt":0.2952834803242316,"score_spread":0.1943844165378995,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4398152538","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":[0.16976838,0.0004598931,0.818583,0.0008397935,0.000117644355,0.0002551869,0.00031241952,0.0038297246,0.005833871],"genre_scores_gemma":[0.96116114,0.00009599792,0.035985153,0.00008982112,0.000016455668,0.000105464664,0.00020144078,0.00004504905,0.002299436],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9985996,0.00039154207,0.000065459026,0.00029951226,0.00037446385,0.0002694337],"domain_scores_gemma":[0.9964954,0.0013209005,0.00023543499,0.0010514989,0.0006077808,0.00028902903],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0023336,0.0004942008,0.00086541224,0.00043857287,0.00092394487,0.0012422439,0.0018793195,0.0010347512,0.0030539043],"category_scores_gemma":[0.0065086125,0.00027019848,0.00031612662,0.0007385031,0.00087546685,0.0025915476,0.0028426773,0.0013494658,0.00071865105],"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":[0.0007844164,0.00026676588,0.003907613,0.00013413231,0.000056876932,0.00040097974,0.00034252062,0.79083383,0.009914912,0.029564971,0.004328331,0.15946463],"study_design_scores_gemma":[0.000037326394,0.00004066332,0.00013638813,0.0000063704833,0.0000043179443,0.000036718524,0.000027969834,0.98647445,0.0026345863,0.009735266,0.0008586627,0.0000072622674],"about_ca_topic_score_codex":0.0043391176,"about_ca_topic_score_gemma":0.004329419,"teacher_disagreement_score":0.0043391176,"about_ca_system_score_codex":0.000980928,"about_ca_system_score_gemma":0.0027949002,"threshold_uncertainty_score":0.01234138},"labels":[],"label_agreement":null},{"id":"W4398152819","doi":"10.1109/tbdata.2024.3403383","title":"Harnessing the Power of Local Supervision in Federated Learning","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Educational Curriculum and Learning Methods","field":"Social Sciences","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":"Computer science; Power (physics); Data science","score_opus":0.1214020775678456,"score_gpt":0.39488234107125697,"score_spread":0.27348026350341137,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4398152819","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":[0.041959044,0.00014215652,0.95335513,0.00043686,0.000033030246,0.000041188585,0.00004273103,0.002560778,0.0014290639],"genre_scores_gemma":[0.893892,0.00009587152,0.104018636,0.00024710796,0.00004975086,0.00008545859,0.00012846435,0.00012191271,0.0013608057],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9973247,0.0010449605,0.00013321429,0.00067113765,0.0005519025,0.000274148],"domain_scores_gemma":[0.98974025,0.0041151033,0.00050005823,0.0043273275,0.0009262475,0.00039107254],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004339443,0.0010202442,0.0015863978,0.0006455317,0.0010864028,0.0017249673,0.0031796515,0.0016111088,0.0015926438],"category_scores_gemma":[0.013334517,0.0006840905,0.0006081843,0.0009251468,0.0023310857,0.005746464,0.0048749456,0.002730371,0.0006024316],"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":[0.0007379219,0.0004374737,0.003456165,0.000116721356,0.00012970778,0.00029504145,0.00037770832,0.76825297,0.008279903,0.031035913,0.0030240202,0.18385647],"study_design_scores_gemma":[0.000017483226,0.00003781656,0.00009349286,0.000004295912,0.000008632369,0.000024217443,0.000016041931,0.9806552,0.0023372315,0.016437516,0.00036157173,0.0000064276405],"about_ca_topic_score_codex":0.0033907082,"about_ca_topic_score_gemma":0.0041753487,"teacher_disagreement_score":0.004339443,"about_ca_system_score_codex":0.0010644064,"about_ca_system_score_gemma":0.0021279065,"threshold_uncertainty_score":0.022949457},"labels":[],"label_agreement":null},{"id":"W4405907170","doi":"10.1109/tbdata.2024.3524105","title":"Data Reconstruction and Protection in Federated Learning for Fine-Tuning Large Language Models","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":8,"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","score_opus":0.13940838244706683,"score_gpt":0.3164687905736269,"score_spread":0.1770604081265601,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405907170","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":[0.042373635,0.00023912475,0.95224565,0.0009203559,0.00005698984,0.00009518514,0.00013561992,0.0023584648,0.0015750135],"genre_scores_gemma":[0.90499943,0.00011461634,0.092484094,0.0004988169,0.000034784047,0.00015949868,0.00017371222,0.00015713763,0.0013780535],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98947173,0.0046751457,0.0007157394,0.0017382153,0.002493035,0.00090601924],"domain_scores_gemma":[0.9793602,0.007160964,0.0009969601,0.011027198,0.001131213,0.00032354085],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008904543,0.00093150284,0.001294378,0.0007121806,0.0013419369,0.0028415287,0.0023866643,0.0022597623,0.0019068724],"category_scores_gemma":[0.03603735,0.000706057,0.0012637002,0.0010668626,0.0035819719,0.007387907,0.0077886204,0.0045548747,0.0009201628],"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":[0.001990489,0.00047030294,0.0076940395,0.00031548867,0.00024206743,0.00091073365,0.0015012952,0.4239017,0.021163957,0.25402823,0.006686278,0.28109542],"study_design_scores_gemma":[0.00006836686,0.00012941055,0.0003953212,0.00004191916,0.000023738416,0.0003248809,0.0001412577,0.80115265,0.021176238,0.17406112,0.0024479602,0.000037144127],"about_ca_topic_score_codex":0.0010725397,"about_ca_topic_score_gemma":0.00077789434,"teacher_disagreement_score":0.008904543,"about_ca_system_score_codex":0.0015308514,"about_ca_system_score_gemma":0.0024831113,"threshold_uncertainty_score":0.04709232},"labels":[],"label_agreement":null},{"id":"W4405907376","doi":"10.1109/tbdata.2024.3524101","title":"Is Split Learning Privacy-Preserving for Fine-Tuning Large Language Models?","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":5,"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; Theoretical computer science; Artificial intelligence","score_opus":0.11959491806294341,"score_gpt":0.32876354956127773,"score_spread":0.2091686314983343,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405907376","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":[0.09271978,0.0009146498,0.89155763,0.006956846,0.0001871839,0.00021522948,0.0005093085,0.0020109846,0.0049283504],"genre_scores_gemma":[0.92925036,0.00030456707,0.06686832,0.0011188908,0.0001602226,0.00016714464,0.00028563963,0.00018588628,0.0016589683],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9836766,0.0076181935,0.0008644729,0.0028630167,0.003119278,0.0018583761],"domain_scores_gemma":[0.931595,0.027728822,0.0027382467,0.034999244,0.002076367,0.0008622736],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.016289636,0.0012451451,0.0022144543,0.0007394965,0.0022547417,0.0044381814,0.0035233367,0.0034800838,0.0026714976],"category_scores_gemma":[0.06636793,0.0012187358,0.00193087,0.001277186,0.006171495,0.018447638,0.007857104,0.007461462,0.0012319067],"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":[0.0026710553,0.000664006,0.014957878,0.00060087943,0.0006677886,0.0010080751,0.0026469992,0.3069505,0.01813866,0.41558108,0.01642221,0.21969084],"study_design_scores_gemma":[0.00012187176,0.00018488668,0.0005940986,0.00006001133,0.00006069522,0.0005055512,0.00032537794,0.55991054,0.011601691,0.4235606,0.0030106676,0.00006407233],"about_ca_topic_score_codex":0.0012075301,"about_ca_topic_score_gemma":0.001075509,"teacher_disagreement_score":0.016289636,"about_ca_system_score_codex":0.001972809,"about_ca_system_score_gemma":0.0038324706,"threshold_uncertainty_score":0.08614886},"labels":[],"label_agreement":null},{"id":"W4409077140","doi":"10.1109/tbdata.2025.3556615","title":"Utility-Driven Data Analytics Algorithm for Transaction Modifications Using Pre-Large Concept With Single Database Scan","year":2025,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Software Testing and Debugging Techniques","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 Alberta","funders":"","keywords":"Computer science; Database transaction; Analytics; Database; Transaction processing; Data mining; Transaction data; Algorithm; Transaction log; Distributed transaction","score_opus":0.2046491776236223,"score_gpt":0.3506634177596448,"score_spread":0.14601424013602252,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4409077140","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":[0.04734599,0.00033877013,0.9472261,0.0002213486,0.000038810358,0.0003241708,0.0002791006,0.0032785304,0.0009471577],"genre_scores_gemma":[0.3791113,0.00019772779,0.61731994,0.00015405078,0.00003030028,0.00039893689,0.00087789516,0.00016247932,0.0017474272],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99852705,0.00015339778,0.0001309513,0.0003731103,0.00071433774,0.000101103156],"domain_scores_gemma":[0.9971077,0.0009975873,0.00032884613,0.00061805535,0.0008220632,0.00012573284],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012811383,0.00090202544,0.0009769729,0.0028216722,0.000715422,0.0012220534,0.0023933346,0.0005688545,0.0013407149],"category_scores_gemma":[0.0055409134,0.0004801899,0.00090466347,0.0029576425,0.0007664399,0.002725137,0.0015618935,0.0013973714,0.000629562],"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":[0.00046458986,0.00037748492,0.018226322,0.00023798071,0.0001222485,0.0005216516,0.00052636117,0.07138486,0.028379908,0.0143111,0.006551265,0.85889626],"study_design_scores_gemma":[0.00004886886,0.0001866706,0.0018600794,0.000018277999,0.00003184807,0.0006298905,0.00017843943,0.9632093,0.016483132,0.0135178035,0.003804764,0.000030996358],"about_ca_topic_score_codex":0.0025108783,"about_ca_topic_score_gemma":0.0036594067,"teacher_disagreement_score":0.0028216722,"about_ca_system_score_codex":0.00066956517,"about_ca_system_score_gemma":0.001929707,"threshold_uncertainty_score":0.006775379},"labels":[],"label_agreement":null},{"id":"W4413822057","doi":"10.1109/tbdata.2025.3604171","title":"Optimizing Deduplication Parameters via a Change-Estimation Analytical Model","year":2025,"lang":"en","type":"article","venue":"IEEE Transactions on Big Data","topic":"Innovation Diffusion and Forecasting","field":"Decision Sciences","cited_by":0,"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 Alberta","funders":"","keywords":"Computer science; Data deduplication; Database","score_opus":0.44762231411228676,"score_gpt":0.4268905043208662,"score_spread":0.02073180979142053,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413822057","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":[0.14132416,0.0014848632,0.84380686,0.0011483593,0.00012172725,0.0003640494,0.0007221981,0.004978835,0.006048815],"genre_scores_gemma":[0.8018524,0.0005613284,0.19213225,0.00028959537,0.00003855314,0.00035208394,0.00069654273,0.00039713157,0.0036801244],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99874413,0.00023469735,0.0000827096,0.0003523888,0.00043010985,0.00015586751],"domain_scores_gemma":[0.9949856,0.00307235,0.00036659345,0.0005266922,0.0009490116,0.00009972387],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0018731715,0.0010267636,0.0006979965,0.0007834862,0.0005630388,0.001584105,0.001316173,0.00086034305,0.0013201438],"category_scores_gemma":[0.012540437,0.0005840353,0.0004806084,0.0008961025,0.0006573998,0.002609877,0.00077668694,0.0013557832,0.0007188106],"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":[0.00015509324,0.00012070947,0.002672918,0.00012519723,0.00002677067,0.000083664054,0.000114481234,0.93893963,0.008377264,0.005632885,0.0027199045,0.041031513],"study_design_scores_gemma":[0.000004990475,0.000020996245,0.0001365657,0.0000052370006,0.0000051309885,0.000024068362,0.000010902986,0.99503976,0.0034503443,0.0007702663,0.0005264763,0.0000051472207],"about_ca_topic_score_codex":0.005080025,"about_ca_topic_score_gemma":0.0053284536,"teacher_disagreement_score":0.005080025,"about_ca_system_score_codex":0.0017860407,"about_ca_system_score_gemma":0.0022548211,"threshold_uncertainty_score":0.012958646},"labels":[],"label_agreement":null},{"id":"W4417002497","doi":"10.1109/tbdata.2025.3639977","title":"TrustChain: A Blockchain Framework for Auditing and Verifying Aggregators in Decentralized Federated Learning","year":2025,"lang":"","type":"article","venue":"IEEE Transactions on Big Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of New Brunswick; University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"News aggregator; Node (physics); Trust management (information system); Audit; Independence (probability theory); Federated learning; Anomaly detection; Scheme (mathematics)","score_opus":0.08314352619012509,"score_gpt":0.3241899130853899,"score_spread":0.2410463868952648,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417002497","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":[0.006512966,0.0001268337,0.9890187,0.00023759578,0.000043699405,0.00022310588,0.00017326027,0.0022088105,0.0014550633],"genre_scores_gemma":[0.5932177,0.00036095615,0.40076038,0.00013271137,0.00009467211,0.00050010864,0.00067679764,0.00026383143,0.0039928053],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99542,0.0019665088,0.00036643902,0.0005398634,0.0013233436,0.00038393782],"domain_scores_gemma":[0.9916916,0.003370942,0.00075666665,0.002472601,0.001175036,0.0005331827],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0072809597,0.0007431223,0.0011512114,0.001376568,0.0016317461,0.0026694492,0.0023963603,0.0015629989,0.0041012745],"category_scores_gemma":[0.015050684,0.00074791926,0.00065642945,0.0013912255,0.0016674807,0.0061871465,0.003789299,0.0018438621,0.001090632],"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":[0.0008614867,0.0003241511,0.0037630005,0.0003598014,0.0001377313,0.0008325492,0.00074675865,0.48176795,0.005657238,0.23191828,0.009929725,0.2637012],"study_design_scores_gemma":[0.000056499495,0.000059287297,0.00017370073,0.00003093237,0.000014412544,0.00008682797,0.000052748415,0.9093578,0.0022000067,0.08231935,0.005623089,0.00002544447],"about_ca_topic_score_codex":0.0058509535,"about_ca_topic_score_gemma":0.006642431,"teacher_disagreement_score":0.0072809597,"about_ca_system_score_codex":0.001566064,"about_ca_system_score_gemma":0.0048498376,"threshold_uncertainty_score":0.038505852},"labels":[],"label_agreement":null},{"id":"W4417002646","doi":"10.1109/tbdata.2025.3639954","title":"Dual-Channel Learning Framework for miRNA-Drug Interaction Prediction Based on Structural Features and Signed Bipartite Graph Neural Network","year":2025,"lang":"","type":"article","venue":"IEEE Transactions on Big Data","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":0,"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 Saskatchewan","funders":"Fundamental Research Funds for the Central Universities; Shaanxi Normal University; National Natural Science Foundation of China","keywords":"Bipartite graph; Graph isomorphism; Graph; Artificial neural network; Biological network; Convolutional neural network; Complex network; Matching (statistics); Sequence (biology)","score_opus":0.06206073388846581,"score_gpt":0.3363125870963738,"score_spread":0.274251853207908,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417002646","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":[0.07996214,0.0014858594,0.9110887,0.0006515104,0.00012172497,0.00009135159,0.0004182207,0.0024841353,0.0036962852],"genre_scores_gemma":[0.8987654,0.00070561306,0.0944848,0.00042474648,0.000086761356,0.00019335227,0.0009600502,0.00008212715,0.004297062],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996946,0.00005976705,0.000018274706,0.00008594701,0.000090880356,0.000050604776],"domain_scores_gemma":[0.9996537,0.00013973175,0.000050918232,0.000020859188,0.00010468229,0.000030192059],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00063705613,0.00089730514,0.0008681716,0.0009258082,0.00026996725,0.0006734157,0.0013202582,0.0010045978,0.0013890046],"category_scores_gemma":[0.0013097205,0.00030773837,0.00082154636,0.0006462516,0.00041130156,0.0010279955,0.0007495183,0.0010098472,0.000337171],"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":[0.00023356803,0.0002668611,0.004488907,0.0001422597,0.00016046053,0.0002534153,0.0000526502,0.80443084,0.0073512336,0.0063903537,0.0031650614,0.17306449],"study_design_scores_gemma":[0.0000034341256,0.000019860994,0.0001339105,0.000001802748,0.000008983318,0.000013014754,0.0000023980167,0.99778986,0.0005511369,0.0013235033,0.00014873916,0.0000034877496],"about_ca_topic_score_codex":0.007586541,"about_ca_topic_score_gemma":0.008845988,"teacher_disagreement_score":0.007586541,"about_ca_system_score_codex":0.0009097514,"about_ca_system_score_gemma":0.0015115524,"threshold_uncertainty_score":0.0150847435},"labels":[],"label_agreement":null}]}