{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":41,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":41,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"151acead7183","filters":{"venue":"2022 IEEE International Conference on Big Data (Big Data)"}},"results":[{"id":"W4318185132","doi":"10.1109/bigdata55660.2022.10020719","title":"From Data Warehouse to Lakehouse: A Comparative Review","year":2022,"lang":"en","type":"review","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":65,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Queen's University","funders":"","keywords":"Data warehouse; Computer science; Big data; Data science; Unstructured data; Data management; Data extraction; Analytics; Strengths and weaknesses; Data transformation; Data virtualization; Database; Data mining; Cloud computing","authors":[{"name":"Ahmed A. Harby","is_ca":true},{"name":"Farhana Zulkernine","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.9445247924110595,"gpt":0.5941263284283097,"spread":0.3503984639827498,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004710013,0.0008632672,0.001620962,0.008820729,0.000556712,0.002913747,0.002102344,0.001376379,0.003942612],"category_scores_gemma":[0.007918675,0.0007103677,0.001130129,0.01868758,0.001201794,0.005366818,0.001491017,0.001742148,0.001646591],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00184064,"about_ca_system_score_gemma":0.004729199,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004187888,"about_ca_topic_score_gemma":0.006539751,"domain_scores_codex":[0.9984487,0.0003991753,0.0002980063,0.0002031419,0.0005606585,0.00009023998],"domain_scores_gemma":[0.9920392,0.005211985,0.0005698972,0.0001644295,0.001786569,0.0002279222],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00009266986,0.00007779762,0.0005398619,0.04937923,0.0001740562,0.0001166471,0.0002862739,0.0004138372,0.0003335404,0.01412051,0.03354496,0.9009205],"study_design_scores_gemma":[0.00001772702,0.0001171816,0.001262711,0.02465578,0.0002568342,0.0005879574,0.0004180486,0.0001719663,0.0003118061,0.004335397,0.9678226,0.00004208005],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0001072173,0.9978928,0.0003070315,0.0004849539,0.0001070418,0.00001020102,0.00002907824,0.000008015,0.001053627],"genre_scores_gemma":[0.0007362419,0.9981734,0.000569218,0.0002526114,0.00009050578,0.00001209925,0.00004331318,0.000004038174,0.0001185109],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.008820729,"threshold_uncertainty_score":0.02490926,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024021","doi":"10.1109/bigdata55660.2022.10020313","title":"The Analysis and Development of an XAI Process on Feature Contribution Explanation","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":25,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Consistency (knowledge bases); Computer science; Feature (linguistics); Process (computing); Ranking (information retrieval); Data mining; Feature selection; Artificial intelligence; Information retrieval","authors":[{"name":"Jun Huang","is_ca":true},{"name":"Zerui Wang","is_ca":true},{"name":"Li Ding","is_ca":true},{"name":"Yan Liu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2539979832914212,"gpt":0.3711565148971464,"spread":0.1171585316057251,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05738936,0.0015796,0.001224924,0.008872311,0.002479532,0.006342391,0.003744287,0.001925717,0.00373332],"category_scores_gemma":[0.1681436,0.001480523,0.002627977,0.005268311,0.003590911,0.01009434,0.006267645,0.004891819,0.001096754],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004402451,"about_ca_system_score_gemma":0.008682775,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005007095,"about_ca_topic_score_gemma":0.003878372,"domain_scores_codex":[0.9621729,0.02107815,0.002617621,0.003429597,0.009943916,0.0007579183],"domain_scores_gemma":[0.7786217,0.1443537,0.01000092,0.03077756,0.03496775,0.001278412],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00021767,0.000527631,0.01933994,0.001457472,0.0002402644,0.0003930145,0.01592031,0.02868278,0.01213687,0.2535862,0.005108752,0.662389],"study_design_scores_gemma":[0.0001410145,0.0007250048,0.01361,0.001630101,0.0002804664,0.0009255852,0.005566709,0.5588713,0.05667326,0.2873986,0.07385308,0.0003247798],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008737051,0.0001684107,0.9856923,0.0008166755,0.00002006271,0.0009372047,0.0001973323,0.001250887,0.002180149],"genre_scores_gemma":[0.04842373,0.000109656,0.9495083,0.00008739282,0.00001705462,0.0006830896,0.00036965,0.0002139627,0.0005870603],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05738936,"threshold_uncertainty_score":0.3035076,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147647","doi":"10.1109/bigdata55660.2022.10020940","title":"Using CNN-LSTM Model for Weather Forecasting","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Energy Load and Power Forecasting","field":"Engineering","cited_by":22,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"","keywords":"Computer science; Deep learning; Generalization; Artificial neural network; Weather forecasting; Artificial intelligence; Feature (linguistics); Data modeling; Big data; Machine learning; Numerical weather prediction; Data mining; Meteorology; Database; Geography","authors":[{"name":"Michael K.H. Fan","is_ca":true},{"name":"Omar Imran","is_ca":true},{"name":"Arka Singh","is_ca":true},{"name":"Samuel A. Ajila","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5383621885771043,"gpt":0.3526381344140766,"spread":0.1857240541630277,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001800685,0.0006595976,0.0002951684,0.0002621771,0.0001637435,0.0004537435,0.0005878911,0.0005773486,0.002664972],"category_scores_gemma":[0.0005333254,0.0002271857,0.000384063,0.0004286467,0.0001173756,0.0008873217,0.000279298,0.0008568015,0.0009582448],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005353364,"about_ca_system_score_gemma":0.0005634428,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02104211,"about_ca_topic_score_gemma":0.02439829,"domain_scores_codex":[0.9999224,0.000007764602,0.000005146429,0.00002623907,0.0000189318,0.00001954716],"domain_scores_gemma":[0.9999069,0.00002664454,0.000009183639,0.00001144048,0.00004053097,0.000005248032],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002272127,0.0001232384,0.002657514,0.0001290707,0.0001138454,0.0001678819,0.00003235266,0.7532723,0.01847842,0.001979801,0.01025384,0.2125645],"study_design_scores_gemma":[0.000002746227,0.00001385205,0.0003100701,0.000004379395,0.000006458496,0.000009944314,0.000003499566,0.9964779,0.002093708,0.0006178847,0.0004556103,0.000003924533],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2713991,0.003424962,0.6833476,0.002032879,0.001223651,0.00009863278,0.005194104,0.01221071,0.02106832],"genre_scores_gemma":[0.9382578,0.000852547,0.04805961,0.0002524363,0.0000958981,0.00005657136,0.00366198,0.0001051743,0.008658162],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02104211,"threshold_uncertainty_score":0.04183924,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318185064","doi":"10.1109/bigdata55660.2022.10020982","title":"Mixers Detection in bitcoin network: a step towards detecting money laundering in crypto-currencies","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Crime, Illicit Activities, and Governance","field":"Social Sciences","cited_by":16,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of New Brunswick","funders":"Innovation Fund","keywords":"Money laundering; Computer science; Proof-of-work system; Anonymity; Database transaction; Computer security; Mixing (physics); Process (computing); Cryptocurrency; Communication source; Digital currency; Pruning; Tree (set theory); Decision tree; Computer network; Data mining; Business; World Wide Web; Operating system; Payment; Database","authors":[{"name":"M. Mazhar Rathore","is_ca":true},{"name":"Sushil S. Chaurasia","is_ca":true},{"name":"Dhirendra Shukla","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3417629011562336,"gpt":0.3747394520079122,"spread":0.03297655085167861,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00107952,0.0009716603,0.0009107617,0.002629901,0.0006217373,0.001374463,0.0008461496,0.001122523,0.00159872],"category_scores_gemma":[0.002326518,0.0002099688,0.0005820093,0.001149699,0.0003034434,0.002066804,0.0009095234,0.001390425,0.001023561],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005900656,"about_ca_system_score_gemma":0.0009321892,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003040433,"about_ca_topic_score_gemma":0.002693776,"domain_scores_codex":[0.9991102,0.0001442265,0.0000551813,0.0001665587,0.0003416286,0.0001823083],"domain_scores_gemma":[0.9988231,0.0003465113,0.0002238084,0.000117393,0.0003765688,0.0001125518],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001223604,0.001176638,0.06940215,0.0002738657,0.0001869919,0.001172708,0.0003201319,0.09410695,0.05243964,0.008521674,0.01387938,0.7572962],"study_design_scores_gemma":[0.00002967688,0.000137526,0.006763551,0.00002888824,0.00003374058,0.0002865579,0.0001136876,0.9621577,0.022852,0.003596965,0.003964792,0.00003491906],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3775644,0.001094537,0.6036886,0.001129035,0.0002540206,0.0004053303,0.001234744,0.008357229,0.006272116],"genre_scores_gemma":[0.9011925,0.0001887238,0.09359196,0.0001725539,0.00006294036,0.0001118804,0.00120898,0.00007634619,0.003394086],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003040433,"threshold_uncertainty_score":0.006045461,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147159","doi":"10.1109/bigdata55660.2022.10020959","title":"Adaptive Content Forwarding Mechanism for Platoon based Vehicular Named Data Networks","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Caching and Content Delivery","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Brandon University","funders":"","keywords":"Platoon; Computer science; Computer network; Overhead (engineering); Scheme (mathematics); Throughput; The Internet; Vehicular ad hoc network; Data exchange; Electronic data interchange; Protocol (science); Distributed computing; Wireless ad hoc network; Telecommunications; Control (management); Wireless; Database; Operating system","authors":[{"name":"Anu Kaushik","is_ca":false},{"name":"Deepanshu Garg","is_ca":false},{"name":"Anushka Nehra","is_ca":false},{"name":"Rasmeet Singh Bali","is_ca":false},{"name":"Mohamed Baza","is_ca":false},{"name":"Gautam Srivastava","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4758896369626464,"gpt":0.3314948332699191,"spread":0.1443948036927272,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008518143,0.0003954249,0.0004712314,0.001084133,0.0009482296,0.0005839454,0.001434562,0.0005344014,0.0009663384],"category_scores_gemma":[0.001842672,0.0001698763,0.0003430872,0.0007013552,0.0004452537,0.001126997,0.0008420768,0.0003769977,0.0002461619],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001165417,"about_ca_system_score_gemma":0.0009251614,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005108405,"about_ca_topic_score_gemma":0.004444557,"domain_scores_codex":[0.9995936,0.00008512679,0.00003147707,0.00008297485,0.0001288791,0.00007800842],"domain_scores_gemma":[0.9989766,0.0002747771,0.0001313871,0.0001884543,0.0003422638,0.00008654725],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001748874,0.0004955978,0.01050596,0.0004233566,0.0002854751,0.001470641,0.0014171,0.3277828,0.1272301,0.1245509,0.01950021,0.384589],"study_design_scores_gemma":[0.00005163688,0.0001432011,0.0007404296,0.00001253057,0.0000490057,0.000238583,0.00009561182,0.96709,0.01305425,0.009558567,0.008929804,0.00003645568],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1124007,0.0007753872,0.8764396,0.0004357568,0.0003765091,0.0003537768,0.0002591461,0.002384034,0.006575055],"genre_scores_gemma":[0.9280716,0.0002947884,0.06892936,0.0001007718,0.00004899004,0.0001144945,0.0003419272,0.00003447886,0.002063518],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005108405,"threshold_uncertainty_score":0.01015735,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318185035","doi":"10.1109/bigdata55660.2022.10020256","title":"Neuro-heuristic Pallet Detection for Automated Guided Vehicle Navigation","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Advanced Manufacturing and Logistics Optimization","field":"Engineering","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Brandon University","funders":"Silesian University of Technology","keywords":"Heuristics; Computer science; Pallet; Focus (optics); Automation; Heuristic; Artificial intelligence; Computer vision; Point (geometry); Robot; Real-time computing; Engineering","authors":[{"name":"Katarzyna Prokop","is_ca":false},{"name":"Dawid Połap","is_ca":false},{"name":"Gautam Srivastava","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2412269997103307,"gpt":0.3397395132856229,"spread":0.0985125135752922,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005522354,0.000992346,0.0008796859,0.001180248,0.0005068327,0.001106768,0.00157526,0.001389667,0.001768557],"category_scores_gemma":[0.001362081,0.0005541209,0.0006002557,0.0008341101,0.0006014222,0.0008782027,0.0006344344,0.0007478626,0.0003790363],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001506178,"about_ca_system_score_gemma":0.00156486,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01914746,"about_ca_topic_score_gemma":0.01539123,"domain_scores_codex":[0.9996866,0.00006564661,0.00001745648,0.00009708948,0.00005998653,0.00007335118],"domain_scores_gemma":[0.9994672,0.000253171,0.00008186465,0.00002868069,0.0001265151,0.00004260405],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001490407,0.0001044328,0.0006794676,0.00004568798,0.00004382634,0.0000606751,0.00002953122,0.9381211,0.001057358,0.001977157,0.001401395,0.05633032],"study_design_scores_gemma":[0.000005603355,0.00001850048,0.00008515077,0.000002898373,0.000004400482,0.000006773623,0.000006149482,0.9988344,0.0002699729,0.0006140065,0.0001490759,0.000003085405],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0763492,0.0008982603,0.9126941,0.0004046042,0.0001528597,0.0001732742,0.0002471645,0.001532283,0.007548164],"genre_scores_gemma":[0.8457364,0.000246806,0.1483043,0.0002390278,0.00005823951,0.0001408174,0.0003186951,0.0000693331,0.004886358],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01914746,"threshold_uncertainty_score":0.03807199,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147587","doi":"10.1109/bigdata55660.2022.10021025","title":"Handwritten Word Recognition using Deep Learning Approach: A Novel Way of Generating Handwritten Words","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Handwritten Text Recognition Techniques","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba","funders":"","keywords":"Computer science; Word (group theory); Artificial intelligence; Word error rate; Bengali; Natural language processing; Process (computing); Speech recognition; Recall rate; Deep learning; Precision and recall; Language model; Pattern recognition (psychology); Mathematics","authors":[{"name":"Mst Shapna Akter","is_ca":false},{"name":"Hossain Shahriar","is_ca":false},{"name":"Alfredo Cuzzocrea","is_ca":false},{"name":"Nova Ahmed","is_ca":false},{"name":"Carson K. Leung","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3210273366955822,"gpt":0.3386155405561522,"spread":0.01758820386056997,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003108816,0.0006081464,0.0003566104,0.0004232879,0.0001867307,0.0005349625,0.000819704,0.0005243961,0.001722932],"category_scores_gemma":[0.0007860306,0.0002343303,0.0004961052,0.0005109179,0.0003065483,0.001012566,0.0005326018,0.0008558689,0.001048249],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004120735,"about_ca_system_score_gemma":0.0006035463,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002488135,"about_ca_topic_score_gemma":0.003555537,"domain_scores_codex":[0.9996643,0.00004229293,0.00002597068,0.0001169568,0.0001117292,0.00003870727],"domain_scores_gemma":[0.9996165,0.00008458299,0.00003634679,0.00009141416,0.0001469334,0.00002414616],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001887415,0.0001792211,0.001492523,0.000194295,0.00006797798,0.0003602381,0.0001317388,0.07599393,0.1189753,0.003144873,0.00528243,0.7939887],"study_design_scores_gemma":[0.00002760754,0.0002131945,0.001178922,0.00002342152,0.00003424296,0.000419721,0.00004093221,0.8588541,0.1279823,0.003350697,0.007843582,0.00003122439],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07856534,0.0005364357,0.9082318,0.0002518286,0.0001903224,0.0001460274,0.0003950692,0.006392518,0.00529057],"genre_scores_gemma":[0.5356234,0.0004183974,0.4485623,0.0003514343,0.00004950159,0.0001420805,0.00171249,0.0003115068,0.012829],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002488135,"threshold_uncertainty_score":0.005763769,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024165","doi":"10.1109/bigdata55660.2022.10020320","title":"Preserving Privacy Integration and Mining for Big Temporal Co-occurrence Patterns","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba","funders":"University of Manitoba","keywords":"Transparency (behavior); Big data; Information privacy; Computer science; Internet privacy; Data integration; Privacy law; Hierarchy; Open data; Computer security; Data science; Privacy policy; Data mining; World Wide Web; Political science","authors":[{"name":"Anifat M. Olawoyin","is_ca":true},{"name":"Carson K. Leung","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.386741268485755,"gpt":0.384995275204468,"spread":0.001745993281286995,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006513974,0.0008664614,0.001381871,0.003173425,0.002080364,0.004000867,0.002311541,0.001365278,0.000626558],"category_scores_gemma":[0.02919683,0.0006266861,0.002400998,0.007289378,0.001655455,0.008381825,0.004383034,0.002128145,0.0004844019],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001418713,"about_ca_system_score_gemma":0.003848201,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00502288,"about_ca_topic_score_gemma":0.007413277,"domain_scores_codex":[0.987518,0.002659992,0.001733105,0.003265847,0.003966359,0.000856732],"domain_scores_gemma":[0.9728187,0.008427052,0.003727136,0.01241785,0.002037608,0.0005717207],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001634751,0.001112779,0.1205327,0.0009510055,0.0008412235,0.002044701,0.004080347,0.1147995,0.02934864,0.113539,0.01439856,0.5967168],"study_design_scores_gemma":[0.00004934139,0.0001595453,0.0107516,0.00008301705,0.0001768782,0.001436923,0.001473385,0.7745705,0.02199157,0.1756327,0.01361318,0.00006133097],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1178151,0.0009352259,0.8713895,0.001568032,0.00008023129,0.0002851998,0.004067831,0.001850423,0.002008397],"genre_scores_gemma":[0.6058532,0.0005190584,0.3811342,0.0004100922,0.0001365248,0.0002633028,0.01016651,0.0001536654,0.001363521],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006513974,"threshold_uncertainty_score":0.03444958,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318187137","doi":"10.1109/bigdata55660.2022.10020293","title":"Entity Matching with AUC-Based Fairness","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Computer science; Matching (statistics); Statistics; Mathematics","authors":[{"name":"Soudeh Nilforoushan","is_ca":true},{"name":"Qianfan Wu","is_ca":true},{"name":"Mostafa Milani","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6664563814827509,"gpt":0.4557380255016488,"spread":0.2107183559811021,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02678535,0.001245359,0.002721226,0.003180128,0.002043812,0.005003613,0.003926276,0.00319431,0.002434625],"category_scores_gemma":[0.1364511,0.0006941913,0.001489522,0.004596928,0.003615809,0.008819884,0.006642721,0.003006437,0.0008215393],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003000777,"about_ca_system_score_gemma":0.003236826,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003730386,"about_ca_topic_score_gemma":0.002042626,"domain_scores_codex":[0.9803161,0.009162597,0.001278789,0.005053448,0.003230415,0.0009585649],"domain_scores_gemma":[0.9307194,0.04092634,0.004897801,0.01620779,0.005929904,0.001318798],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008998867,0.0002436894,0.02193455,0.0002381851,0.0002861011,0.0002996953,0.000722184,0.5717043,0.002300524,0.1328978,0.005417662,0.2630554],"study_design_scores_gemma":[0.00002857802,0.00008540067,0.001347073,0.00002921148,0.00003251544,0.0001679258,0.00006326329,0.8687935,0.003171737,0.1237694,0.002472335,0.00003892709],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03713205,0.0008125136,0.9573455,0.0007539403,0.00009623414,0.0001486396,0.0002644452,0.001055056,0.002391632],"genre_scores_gemma":[0.7072269,0.0003566648,0.2873195,0.000518674,0.0002046467,0.0002284506,0.0007127209,0.0003759531,0.003056445],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02678535,"threshold_uncertainty_score":0.1416562,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147167","doi":"10.1109/bigdata55660.2022.10020867","title":"Emergence of an Autonomous Vehicle Secondary Data Market for Breakthrough Applications","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"","keywords":"Computer science; Data science; Cloud computing; Variety (cybernetics); Architecture; Data collection; Key (lock); Telecommunications; Systems engineering; Artificial intelligence; Engineering; Computer security; Geography","authors":[{"name":"Kevin Dick","is_ca":true},{"name":"James R. Green","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3288503865665113,"gpt":0.4098341532725872,"spread":0.08098376670607593,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004696312,0.0004138353,0.0004908085,0.001011092,0.001127217,0.006888398,0.002587081,0.002708508,0.007780433],"category_scores_gemma":[0.007741979,0.0003323995,0.0005935127,0.00162238,0.001856261,0.01366185,0.004030952,0.003268908,0.002515695],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002729588,"about_ca_system_score_gemma":0.002747904,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002788791,"about_ca_topic_score_gemma":0.002107126,"domain_scores_codex":[0.9984611,0.0002879878,0.00005704956,0.0002818945,0.0006452512,0.0002667416],"domain_scores_gemma":[0.9925458,0.001779625,0.0004152989,0.001155577,0.002916387,0.001187307],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002793883,0.0002257619,0.004929757,0.0002814386,0.00003379345,0.00049716,0.0008591732,0.008742684,0.009613982,0.7930433,0.0574206,0.124073],"study_design_scores_gemma":[0.00007628367,0.000443763,0.003502334,0.0002194282,0.0000318455,0.0005161604,0.002386285,0.2145515,0.009551401,0.3282207,0.4403999,0.0001003979],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1914573,0.004743397,0.5315946,0.1345815,0.002858722,0.0006190997,0.002923498,0.00432396,0.1268979],"genre_scores_gemma":[0.8197573,0.003907285,0.1396094,0.0061817,0.001697572,0.0002842665,0.002300225,0.0005847424,0.02567746],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007780433,"threshold_uncertainty_score":0.02602816,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147794","doi":"10.1109/bigdata55660.2022.10020299","title":"Temporal Graph Representation Learning via Maximal Cliques","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Clique; Computer science; Graph; Node (physics); Theoretical computer science; Embedding; Representation (politics); Graph embedding; Feature learning; Artificial intelligence; Mathematics; Combinatorics","authors":[{"name":"Shima Khoshraftar","is_ca":true},{"name":"Aijun An","is_ca":true},{"name":"Nastaran Babanejad","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.304841879149363,"gpt":0.3731440942115506,"spread":0.06830221506218759,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007572331,0.0009151363,0.0009099173,0.002196542,0.0007691778,0.000856853,0.001792654,0.001156081,0.003151304],"category_scores_gemma":[0.004555995,0.0005481379,0.00131201,0.002474522,0.0007372856,0.00308604,0.001288154,0.001467728,0.0006382394],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001273045,"about_ca_system_score_gemma":0.0009455015,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009070338,"about_ca_topic_score_gemma":0.02029452,"domain_scores_codex":[0.9993224,0.0001693816,0.00002819148,0.0002695223,0.0001361195,0.00007430583],"domain_scores_gemma":[0.9982273,0.0008627985,0.0002361142,0.0003003655,0.0002588323,0.0001146229],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002769576,0.0003053334,0.007614385,0.0003610276,0.0002719382,0.0003002688,0.0003826895,0.5407816,0.005163216,0.06339067,0.018522,0.3626299],"study_design_scores_gemma":[0.00001062005,0.00001545646,0.000344723,0.00001242331,0.00001554451,0.00002955282,0.00002656566,0.9699661,0.0004170141,0.02783577,0.001320209,0.000006079487],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0626722,0.0006402356,0.9301938,0.0005754046,0.00008400137,0.0001256476,0.001098976,0.001596072,0.003013665],"genre_scores_gemma":[0.6790603,0.000725839,0.3056372,0.0004025857,0.0001749831,0.0003163344,0.006678321,0.0004133111,0.00659115],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009070338,"threshold_uncertainty_score":0.01803505,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318186065","doi":"10.1109/bigdata55660.2022.10020878","title":"Cough Classification with Deep Derived Features using Audio Spectrogram Transformer","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Respiratory and Cough-Related Research","field":"Medicine","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa; Carleton University; National Research Council Canada","funders":"National Research Council","keywords":"Spectrogram; Transformer; Computer science; Artificial intelligence; Feature engineering; Feature selection; Machine learning; Nonlinear system; Feature extraction; Deep learning; Data mining; Pattern recognition (psychology); Engineering","authors":[{"name":"Julio J. Valdés","is_ca":true},{"name":"Karim Habashy","is_ca":true},{"name":"Pengcheng Xi","is_ca":true},{"name":"Madison Cohen-McFarlane","is_ca":true},{"name":"Bruce Wallace","is_ca":true},{"name":"Rafik Goubran","is_ca":true},{"name":"Frank Knoefel","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4486312601254973,"gpt":0.407795632811187,"spread":0.04083562731431023,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004396398,0.0007051156,0.0004146312,0.0006107163,0.0001391455,0.0006013428,0.0004858831,0.0004536359,0.001002802],"category_scores_gemma":[0.001339346,0.0001625289,0.0006497673,0.0003696386,0.0002048549,0.0007215026,0.0006248275,0.000698486,0.0004690738],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003180068,"about_ca_system_score_gemma":0.0003913396,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002317742,"about_ca_topic_score_gemma":0.002407375,"domain_scores_codex":[0.9998729,0.00002749244,0.000009788362,0.00003281273,0.00003196071,0.00002506196],"domain_scores_gemma":[0.9996983,0.000161851,0.00002534908,0.00003047982,0.00006605963,0.00001803157],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0009353724,0.0004677353,0.01539025,0.000184345,0.0001805312,0.0004718184,0.0001272049,0.2976134,0.06998735,0.001919129,0.003478921,0.6092439],"study_design_scores_gemma":[0.000007905277,0.00007871338,0.001655526,0.000006687727,0.00001764987,0.00005936857,0.00002024662,0.9904171,0.00658406,0.0007731886,0.0003710256,0.00000857174],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2714895,0.0007474406,0.7218077,0.0003983752,0.000165425,0.00008846479,0.0005988473,0.002640143,0.002064226],"genre_scores_gemma":[0.9548069,0.0002351524,0.04242563,0.00007819542,0.00004156374,0.00004699022,0.000684419,0.00004392569,0.001637203],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002317742,"threshold_uncertainty_score":0.004608452,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318186088","doi":"10.1109/bigdata55660.2022.10020426","title":"A Convergence Theory for Federated Average: Beyond Smoothness","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; University of British Columbia Hospital","funders":"","keywords":"Smoothness; Computer science; Lipschitz continuity; Convergence (economics); Bounded function; Function (biology); Gradient descent; Mathematical optimization; Scheme (mathematics); Stochastic gradient descent; Applied mathematics; Mathematics; Artificial intelligence; Artificial neural network; Mathematical analysis","authors":[{"name":"Xiaoxiao Li","is_ca":true},{"name":"Zhao Song","is_ca":false},{"name":"Runzhou Tao","is_ca":false},{"name":"Guangyi Zhang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2782544865332073,"gpt":0.3560469689397069,"spread":0.07779248240649961,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01287062,0.002972136,0.003308933,0.003186401,0.001812927,0.003198568,0.00425962,0.002964424,0.00538445],"category_scores_gemma":[0.05411834,0.001068776,0.003835653,0.002938734,0.004820506,0.007911771,0.005842644,0.008657575,0.00121361],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003395754,"about_ca_system_score_gemma":0.003131595,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003876083,"about_ca_topic_score_gemma":0.002029018,"domain_scores_codex":[0.9953828,0.001574914,0.0002561368,0.001074141,0.001269455,0.0004424611],"domain_scores_gemma":[0.9679183,0.02243644,0.001540015,0.002539256,0.004609753,0.0009562347],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0003343218,0.0001450914,0.002304552,0.0005230221,0.0001855576,0.0002721119,0.0003971181,0.3924459,0.002248563,0.5508825,0.007086073,0.04317515],"study_design_scores_gemma":[0.00001720768,0.00006944041,0.0002072194,0.0000732327,0.00003350392,0.00008327382,0.0000380433,0.8609276,0.000734864,0.1362446,0.001545739,0.000025321],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006333806,0.0008519553,0.9888055,0.0008049782,0.00009117497,0.00006134855,0.0001168696,0.0002871802,0.002647219],"genre_scores_gemma":[0.5751334,0.005236943,0.4015481,0.001833173,0.0007500913,0.001163878,0.0008822639,0.0011376,0.01231459],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01287062,"threshold_uncertainty_score":0.06806713,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318185157","doi":"10.1109/bigdata55660.2022.10020564","title":"A Novel Rigorous Measurement Model for Big Data Quality Characteristics","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Big Data and Business Intelligence","field":"Business, Management and Accounting","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Big data; Computer science; Data modeling; Quality (philosophy); Data mining; Database; Physics","authors":[{"name":"Haochen Zou","is_ca":true},{"name":"Kun Xiang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7505477363862263,"gpt":0.4080711581157322,"spread":0.3424765782704941,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01265433,0.001531168,0.001163578,0.004748272,0.001090823,0.005786613,0.003666724,0.001810335,0.002058089],"category_scores_gemma":[0.03952235,0.0008043307,0.002152782,0.004451491,0.002717114,0.01262358,0.003507426,0.003203054,0.0006707771],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004878956,"about_ca_system_score_gemma":0.005008986,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01017129,"about_ca_topic_score_gemma":0.005695426,"domain_scores_codex":[0.9797633,0.006247068,0.001900213,0.004127363,0.00700897,0.0009531234],"domain_scores_gemma":[0.9761101,0.009616345,0.003194085,0.002748327,0.007781511,0.0005496112],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001348605,0.0002771312,0.01519927,0.0003709331,0.0002345721,0.0002456427,0.001013069,0.2595353,0.004844608,0.5952786,0.004046548,0.1188195],"study_design_scores_gemma":[0.00001706421,0.00007747099,0.001859001,0.00003453978,0.00004966335,0.00008547423,0.0001010027,0.8937955,0.001132375,0.1005244,0.002281063,0.00004238749],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00421806,0.00007427587,0.9933792,0.0003140524,0.00002063806,0.0001261163,0.0001069494,0.0003306751,0.001429943],"genre_scores_gemma":[0.4177821,0.000395852,0.5774072,0.0002587934,0.0001234524,0.00114412,0.0007260084,0.0001278544,0.002034652],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01265433,"threshold_uncertainty_score":0.06692332,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147859","doi":"10.1109/bigdata55660.2022.10020507","title":"Discovering Limitations of Image Quality Assessments with Noised Deep Learning Image Sets","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Image and Video Quality Assessment","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Canadian Institute for Advanced Research; Google","keywords":"MNIST database; Computer science; Artificial intelligence; Pixel; Image quality; Image (mathematics); Image processing; Computer vision; Noise (video); Image resolution; Image restoration; Pattern recognition (psychology); Deep learning","authors":[{"name":"Wei Dai","is_ca":false},{"name":"Daniel Berleant","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4654309621175568,"gpt":0.4333924151649992,"spread":0.03203854695255764,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01108315,0.001261805,0.001031581,0.002080553,0.0007885604,0.002437231,0.002681056,0.001659436,0.0009850285],"category_scores_gemma":[0.05345085,0.0005625257,0.0009565289,0.00114088,0.002080703,0.003599929,0.002579435,0.002413427,0.000592606],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001887231,"about_ca_system_score_gemma":0.001109502,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006132843,"about_ca_topic_score_gemma":0.005925033,"domain_scores_codex":[0.9932733,0.0016991,0.0006020333,0.001270499,0.002835417,0.0003197021],"domain_scores_gemma":[0.9813405,0.00810902,0.001503067,0.003795022,0.004845674,0.0004069028],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.002227839,0.0005525739,0.03007684,0.001543997,0.001102688,0.0004288758,0.0007824429,0.2905378,0.0635741,0.0121872,0.01318472,0.583801],"study_design_scores_gemma":[0.00008872569,0.0004946391,0.01073687,0.0002469119,0.0001875597,0.0004483945,0.0002326954,0.9001076,0.06316436,0.0196896,0.004501458,0.0001012064],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3446874,0.007715698,0.6301368,0.002990888,0.0006851706,0.0004949262,0.001940046,0.004990181,0.006358837],"genre_scores_gemma":[0.7444268,0.001150958,0.2485544,0.001045357,0.0001491136,0.0002173892,0.002401865,0.0004168204,0.001637322],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01108315,"threshold_uncertainty_score":0.05861402,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147653","doi":"10.1109/bigdata55660.2022.10020557","title":"Multimodal Deep Homography Estimation Using a Domain Adaptation Generative Adversarial Network","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Advanced Image and Video Retrieval Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Defence Research and Development Canada; National Research Council Canada; Thales (Canada)","funders":"","keywords":"Artificial intelligence; Computer science; Homography; Computer vision; Parallax; Deep learning; Feature (linguistics); Generative grammar; Modalities; Domain (mathematical analysis); Adversarial system; Pattern recognition (psychology); Mathematics; Projective test","authors":[{"name":"T. Pouplin","is_ca":true},{"name":"Hughes Perreault","is_ca":true},{"name":"Benoît Debaque","is_ca":true},{"name":"M-A. Drouin","is_ca":true},{"name":"N. Duclos-Hindie","is_ca":true},{"name":"Sébastien Roy","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2705887029607451,"gpt":0.3671188911458947,"spread":0.09653018818514963,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007842469,0.001231287,0.001328496,0.0006141839,0.0003337089,0.0007927343,0.001726102,0.001293058,0.004018029],"category_scores_gemma":[0.001731839,0.0006153238,0.001174738,0.0006810514,0.0008733773,0.001300793,0.002002326,0.002857846,0.001398758],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008499872,"about_ca_system_score_gemma":0.0009685078,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005216438,"about_ca_topic_score_gemma":0.006933577,"domain_scores_codex":[0.9995747,0.0001064899,0.00001359782,0.0001397034,0.0001064991,0.0000589826],"domain_scores_gemma":[0.9995098,0.0001941725,0.00005634689,0.0001243228,0.00007580359,0.00003956401],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001294847,0.00008919177,0.0005863612,0.00006623165,0.0001124727,0.0001519681,0.00006971301,0.7954578,0.00641463,0.009274618,0.005865452,0.181782],"study_design_scores_gemma":[0.000003973404,0.00001197017,0.00005985097,0.000003921588,0.000004412429,0.00002749211,0.000004094208,0.9963257,0.0008384239,0.002246701,0.0004693451,0.000004187856],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0112148,0.0002960716,0.9846196,0.0002459543,0.00005212087,0.00004252448,0.0001100099,0.001798352,0.001620726],"genre_scores_gemma":[0.6229854,0.0006525422,0.355367,0.0008780361,0.0001869406,0.00021329,0.00152367,0.0006675881,0.01752562],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005216438,"threshold_uncertainty_score":0.01344168,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318148360","doi":"10.1109/bigdata55660.2022.10020874","title":"LSTM-based Pulmonary Air Leak Forecasting for Chest Tube Management","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"COVID-19 diagnosis using AI","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital","funders":"","keywords":"Context (archaeology); Empyema; Chest tube; Computer science; Leak; Autoregressive model; Medicine; Intensive care medicine; Surgery; Engineering","authors":[{"name":"Roberto Corizzo","is_ca":false},{"name":"Rodrigo Yepez-Lopez","is_ca":false},{"name":"Sébastien Gilbert","is_ca":true},{"name":"Nathalie Japkowicz","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4472536916586746,"gpt":0.3922616421471449,"spread":0.05499204951152969,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004170867,0.0007136454,0.0004885659,0.000470808,0.0002120499,0.0005162762,0.0007438046,0.0008189158,0.001474954],"category_scores_gemma":[0.001549841,0.000228674,0.0004133756,0.0005207977,0.0001586237,0.0007426795,0.0003927578,0.001008059,0.0005521683],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005592936,"about_ca_system_score_gemma":0.0007360314,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006949858,"about_ca_topic_score_gemma":0.006893394,"domain_scores_codex":[0.9998785,0.00002232141,0.00001235563,0.00003393279,0.00002732149,0.00002553382],"domain_scores_gemma":[0.9996579,0.0001704936,0.00003495728,0.0000205584,0.00009568693,0.00002049564],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000421488,0.0002964929,0.005473239,0.0001759423,0.0001217113,0.0002738053,0.00009903064,0.6764566,0.01354673,0.001423019,0.007190294,0.2945217],"study_design_scores_gemma":[0.000003431463,0.00001930446,0.0002717075,0.000006112888,0.000007870009,0.00001215928,0.000006173507,0.9979742,0.0008783244,0.0006159975,0.0002012325,0.000003543741],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2146544,0.003852866,0.7637084,0.002212487,0.000783473,0.0001291842,0.001669446,0.007079919,0.005909872],"genre_scores_gemma":[0.9515593,0.00056207,0.04465699,0.0002690925,0.0001302184,0.00007523856,0.0008037821,0.00004096017,0.00190225],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006949858,"threshold_uncertainty_score":0.01381886,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024352","doi":"10.1109/bigdata55660.2022.10020258","title":"Review of Publically Available Health Big Data Sets","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Victoria","funders":"University of Victoria","keywords":"Health informatics; Public health informatics; Computer science; Informatics; Usability; Data science; Dashboard; Data mining; Information retrieval; World Wide Web; Public health; Medicine; Health policy; International health; Engineering","authors":[{"name":"Dillon Chrimes","is_ca":true},{"name":"Chanhee Kim","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.8077064389291911,"gpt":0.5061207953285485,"spread":0.3015856436006427,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03155651,0.001052886,0.002251909,0.05159301,0.001735999,0.005988922,0.004058817,0.001607262,0.005779488],"category_scores_gemma":[0.1717749,0.00143258,0.002924943,0.055131,0.002227793,0.007366271,0.005693455,0.002508166,0.001540937],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004370212,"about_ca_system_score_gemma":0.01835864,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0154184,"about_ca_topic_score_gemma":0.02595694,"domain_scores_codex":[0.9697468,0.008623609,0.006593123,0.002592647,0.01149096,0.0009527735],"domain_scores_gemma":[0.6788823,0.2396056,0.01760661,0.01765033,0.04310214,0.003153044],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003414069,0.0001208881,0.0233556,0.1068988,0.001682664,0.0005018684,0.002406664,0.002688971,0.001206787,0.03216124,0.2672195,0.5614157],"study_design_scores_gemma":[0.00003262492,0.00006647575,0.02779986,0.06288251,0.0006769891,0.0005831387,0.002037218,0.0009506827,0.001090139,0.007728277,0.896001,0.0001510462],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.01396789,0.7586977,0.01516241,0.03616301,0.003414442,0.0009325997,0.1467903,0.001214182,0.02365757],"genre_scores_gemma":[0.06999286,0.674496,0.03367609,0.01158832,0.00287934,0.001815049,0.2027586,0.0008822184,0.001911607],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.05159301,"threshold_uncertainty_score":0.1668888,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318148204","doi":"10.1109/bigdata55660.2022.10020871","title":"Predicting Deduplication Performance: An Analytical Model and Empirical Evaluation","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Advanced Data Storage Technologies","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Data deduplication; Computer science; Data mining; Software versioning; Kernel (algebra); Database; Software; Operating system","authors":[{"name":"Owen Randall","is_ca":true},{"name":"Paul Lu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5054255254777182,"gpt":0.4297697773039301,"spread":0.07565574817378812,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005261728,0.00243712,0.001340329,0.002538593,0.0008354508,0.002308005,0.002598772,0.002269363,0.001273411],"category_scores_gemma":[0.0354397,0.0009677552,0.0008685465,0.003184803,0.00114926,0.004590977,0.001050563,0.002639818,0.001201887],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003419048,"about_ca_system_score_gemma":0.002290427,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01831678,"about_ca_topic_score_gemma":0.01292454,"domain_scores_codex":[0.9972665,0.000566824,0.0001925051,0.0007603342,0.0008203724,0.0003934156],"domain_scores_gemma":[0.9695179,0.02132299,0.001875607,0.002342374,0.004431116,0.0005100341],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000428867,0.0006129332,0.02592646,0.0003909711,0.0000989893,0.0001636365,0.000200661,0.9233492,0.003610071,0.002394123,0.006268713,0.03655542],"study_design_scores_gemma":[0.000008198243,0.00005717379,0.001195128,0.0000134008,0.000008916923,0.00004644111,0.00003031295,0.9963342,0.001223965,0.0007555789,0.0003149938,0.00001167374],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7658339,0.004568174,0.2065862,0.003062312,0.0002286048,0.0007240386,0.005420077,0.008162227,0.005414494],"genre_scores_gemma":[0.8925783,0.001670466,0.09767785,0.0004612896,0.00009718475,0.0004259814,0.004873342,0.0003659511,0.001849739],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01831678,"threshold_uncertainty_score":0.03642035,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024124","doi":"10.1109/bigdata55660.2022.10020422","title":"Dynamic Graph Summarization: Optimal and Scalable","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Victoria","funders":"","keywords":"Automatic summarization; Computer science; Scalability; Lossless compression; Graph; Theoretical computer science; Algorithm; Artificial intelligence; Data compression","authors":[{"name":"Mahdi Hajiabadi","is_ca":true},{"name":"Venkatesh Srinivasan","is_ca":true},{"name":"Alex Thomo","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.158207502154851,"gpt":0.3135448753019634,"spread":0.1553373731471124,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008821317,0.001703022,0.0012322,0.002614337,0.0008841509,0.001901187,0.002209421,0.001252063,0.003004466],"category_scores_gemma":[0.006183981,0.0006758538,0.0007796665,0.002856929,0.0006424216,0.004319291,0.001733615,0.001352093,0.001989365],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009780048,"about_ca_system_score_gemma":0.001165059,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003365779,"about_ca_topic_score_gemma":0.006822084,"domain_scores_codex":[0.9985663,0.000197518,0.0001071954,0.0005095589,0.0005012594,0.0001180694],"domain_scores_gemma":[0.9969386,0.0008733531,0.0003294024,0.001073228,0.0006387477,0.0001466661],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004363509,0.0002393771,0.00142735,0.0005648188,0.0001312992,0.000265557,0.0003872375,0.1249993,0.04047303,0.01178611,0.048517,0.7707726],"study_design_scores_gemma":[0.0001423235,0.000272836,0.001219701,0.00004553685,0.0001144343,0.0005548588,0.000358511,0.8907195,0.03627177,0.04323395,0.02700343,0.00006319459],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02312491,0.001462699,0.9536057,0.0006598174,0.0002050695,0.0002816366,0.001594976,0.01581874,0.003246416],"genre_scores_gemma":[0.1717051,0.0006688097,0.811233,0.0003198661,0.0002308019,0.000238693,0.008927389,0.001463594,0.005212753],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003365779,"threshold_uncertainty_score":0.01005095,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147168","doi":"10.1109/bigdata55660.2022.10020831","title":"A Real-time, Scalable Monitoring and User Analytics Solution for Microservices-based Software Applications","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Software System Performance and Reliability","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Microservices; Computer science; Scalability; Analytics; Big data; Database transaction; Transaction log; Software; Distributed computing; Database; Cloud computing; Data mining; Operating system","authors":[{"name":"Cathy H Zhang","is_ca":true},{"name":"M. Omair Shafiq","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1826437870504777,"gpt":0.3440523884764013,"spread":0.1614086014259236,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007179425,0.0007325864,0.0005518082,0.0005647579,0.0004490183,0.001078658,0.001589356,0.0006544187,0.001198278],"category_scores_gemma":[0.001575259,0.0004058286,0.0003366014,0.0005104844,0.0003279386,0.001636566,0.001484671,0.001270337,0.0006407963],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004480394,"about_ca_system_score_gemma":0.000876137,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001863343,"about_ca_topic_score_gemma":0.002328838,"domain_scores_codex":[0.9990319,0.0001140134,0.00007846726,0.000260596,0.0003945459,0.0001204646],"domain_scores_gemma":[0.9991585,0.0001677438,0.0001207117,0.0001972841,0.0002137304,0.0001420968],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001519496,0.001235422,0.0262916,0.0004669323,0.0001963845,0.001676473,0.00161912,0.04944352,0.1774513,0.01277749,0.02318209,0.7041401],"study_design_scores_gemma":[0.00005819717,0.0004203802,0.01220231,0.00007482078,0.00006572353,0.0004607217,0.0004220665,0.8970878,0.06447021,0.007778035,0.01684912,0.000110567],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2114588,0.000554506,0.7273628,0.001404889,0.0001378455,0.0004683922,0.0005656847,0.05198727,0.006059853],"genre_scores_gemma":[0.7983428,0.0001868398,0.1949842,0.0003866736,0.00006108742,0.0002478119,0.0006208827,0.000401878,0.004767688],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001863343,"threshold_uncertainty_score":0.00400871,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318148225","doi":"10.1109/bigdata55660.2022.10020386","title":"Adaptive Method for Machine Learning Model Selection in Data Science Projects","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Software Engineering Research","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Selection (genetic algorithm); Machine learning; Process (computing); Heuristics; Feature selection; Model selection; Artificial intelligence; Data mining","authors":[{"name":"Cristina Tavares","is_ca":true},{"name":"Nathalia Nascimento","is_ca":true},{"name":"Paulo Alencar","is_ca":true},{"name":"Donald Cowan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4726800557815008,"gpt":0.4217361885031067,"spread":0.05094386727839412,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01356869,0.00109503,0.0009174866,0.002928711,0.0009752389,0.001830918,0.002637245,0.001639127,0.00381531],"category_scores_gemma":[0.04898661,0.0007623716,0.001702372,0.002277857,0.0009944342,0.002137985,0.002343253,0.003348596,0.0006640197],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001568343,"about_ca_system_score_gemma":0.00286122,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004615558,"about_ca_topic_score_gemma":0.007426095,"domain_scores_codex":[0.9905923,0.006097058,0.0004585738,0.001169952,0.001412179,0.0002699872],"domain_scores_gemma":[0.9591773,0.03396612,0.001570697,0.002282236,0.002600614,0.0004029306],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000418662,0.0003794291,0.01200561,0.0003297956,0.0004161185,0.0004100631,0.0008298467,0.5824137,0.003129953,0.06175525,0.004448804,0.3334629],"study_design_scores_gemma":[0.00002606363,0.00003024467,0.000264733,0.0000160844,0.00001652133,0.00002964246,0.00004130354,0.983095,0.0006266172,0.01473688,0.001105469,0.00001143177],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005203799,0.00004400319,0.9935132,0.0001016857,0.0000131216,0.0001056363,0.00004133253,0.0005492142,0.000428109],"genre_scores_gemma":[0.1561818,0.00006282736,0.8414322,0.0001569201,0.00003315354,0.000732776,0.0002989469,0.0002084426,0.0008929871],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01356869,"threshold_uncertainty_score":0.07175893,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318187680","doi":"10.1109/bigdata55660.2022.10020691","title":"Air Pollution Surveillance System: A Big Data Approach to Monitoring Adverse Health Outcomes for Public Health Interventions","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Air Quality and Health Impacts","field":"Environmental Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Big data; Harm; Air pollution; Public health; Psychological intervention; Environmental health; Environmental planning; Business; Risk analysis (engineering); Computer science; Environmental science; Medicine; Political science; Data mining","authors":[{"name":"Shahan Salim","is_ca":true},{"name":"Irfhana Zakir Hussain","is_ca":true},{"name":"Jasleen Kaur","is_ca":true},{"name":"Plinio Pelegrini Morita","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6416240419164759,"gpt":0.4430536513876135,"spread":0.1985703905288624,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003079797,0.00144162,0.001400993,0.003367232,0.001039005,0.002180091,0.002318938,0.001589916,0.002507481],"category_scores_gemma":[0.004894571,0.0004763622,0.001052998,0.002106783,0.0005059511,0.002349727,0.00250697,0.001756324,0.001050659],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001100175,"about_ca_system_score_gemma":0.002406884,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01197529,"about_ca_topic_score_gemma":0.01575783,"domain_scores_codex":[0.99821,0.0005048711,0.0002007327,0.0004747836,0.0004983586,0.0001112687],"domain_scores_gemma":[0.9970133,0.000884708,0.0003465234,0.0005087175,0.0009107008,0.0003360269],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00142837,0.002565145,0.162631,0.001941034,0.002172753,0.0007890032,0.0007986997,0.066451,0.01688617,0.02803037,0.1962799,0.5200266],"study_design_scores_gemma":[0.0002590033,0.0007958827,0.05108413,0.0003460101,0.0006504701,0.0003826576,0.0006537117,0.7879407,0.01628886,0.04663721,0.09478208,0.0001793603],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09121646,0.005036018,0.7058788,0.02353204,0.002462893,0.005287337,0.08530173,0.04429974,0.03698498],"genre_scores_gemma":[0.6280248,0.002843535,0.3125728,0.005617256,0.001151331,0.002282972,0.03887006,0.0003939273,0.008243287],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01197529,"threshold_uncertainty_score":0.02381116,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024183","doi":"10.1109/bigdata55660.2022.10020511","title":"Dynamic Ensemble Size Adjustment for Memory Constrained Mondrian Forest","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Mondrian; Overfitting; Computer science; Machine learning; Ensemble learning; Tree (set theory); Data stream mining; Artificial intelligence; Memory model; Data mining; Shared memory; Parallel computing; Mathematics; Artificial neural network","authors":[{"name":"Martin Khannouz","is_ca":true},{"name":"Tristan Glatard","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2162287159834481,"gpt":0.3500891358264674,"spread":0.1338604198430194,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00375346,0.0007728861,0.001248586,0.001203142,0.001280219,0.001004628,0.002191028,0.001441924,0.00159002],"category_scores_gemma":[0.01286405,0.0003803779,0.0008087418,0.0007492983,0.0005959302,0.002175404,0.00138239,0.001601508,0.0005548577],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009103406,"about_ca_system_score_gemma":0.001661721,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005154857,"about_ca_topic_score_gemma":0.009597021,"domain_scores_codex":[0.9989504,0.0002617615,0.00007605456,0.0002852086,0.0002583288,0.0001682115],"domain_scores_gemma":[0.9959155,0.002009882,0.000237086,0.0005995042,0.001041912,0.0001960985],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005114623,0.0001826975,0.008136653,0.0001094505,0.0001209176,0.000154416,0.0002237228,0.5467588,0.007358931,0.01016232,0.008835535,0.4174451],"study_design_scores_gemma":[0.0000121666,0.00001998904,0.0003522657,0.0000102056,0.00000989098,0.00003781352,0.00001492252,0.9940084,0.001252187,0.003674486,0.0005997181,0.000007986261],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07279263,0.0007060293,0.9221292,0.0002889073,0.0001665382,0.0001015424,0.0001820802,0.001933154,0.001699997],"genre_scores_gemma":[0.5541571,0.0002521965,0.4421801,0.000345728,0.0001652989,0.0002664455,0.0007754272,0.0003343008,0.001523355],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005154857,"threshold_uncertainty_score":0.01985037,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024064","doi":"10.1109/bigdata55660.2022.10020783","title":"Imbalanced Multi-layer Cloud Classification with Advanced Baseline Imager (ABI) and CloudSat/CALIPSO Data","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Remote Sensing in Agriculture","field":"Environmental Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"National Research Council Canada","funders":"National Security Agency; National Science Foundation","keywords":"Cloud computing; Random forest; Lidar; Computer science; Classifier (UML); Perceptron; Baseline (sea); Deep learning; Resampling; Remote sensing; Cloud cover; Cloud fraction; Artificial intelligence; Environmental science; Artificial neural network; Geography; Geology","authors":[{"name":"Lei Ding","is_ca":false},{"name":"Roberto Corizzo","is_ca":false},{"name":"Colin Bellinger","is_ca":true},{"name":"N. Ching","is_ca":false},{"name":"Spencer Login","is_ca":false},{"name":"Rodrigo Yepez-Lopez","is_ca":false},{"name":"Jie Gong","is_ca":false},{"name":"Dong L. Wu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2510556575753,"gpt":0.3345477456942415,"spread":0.08349208811894149,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001107328,0.0009873868,0.0005153013,0.0009360527,0.0005159648,0.0007004881,0.001064742,0.0006474014,0.0008789645],"category_scores_gemma":[0.001902729,0.0002254793,0.0009267887,0.001025555,0.0003295546,0.001381673,0.0006656392,0.0009923138,0.000382585],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008138054,"about_ca_system_score_gemma":0.0007845617,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02099788,"about_ca_topic_score_gemma":0.01855164,"domain_scores_codex":[0.9994801,0.00006869894,0.00003152246,0.000127477,0.0001654531,0.0001266619],"domain_scores_gemma":[0.9993895,0.0001192037,0.00009346495,0.0001447865,0.0001918731,0.00006099859],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","study_design_scores_codex":[0.001452019,0.001932406,0.09870572,0.0001890855,0.0003799952,0.0004675797,0.0002114756,0.5855398,0.0322347,0.001992899,0.01321477,0.2636795],"study_design_scores_gemma":[0.00002036485,0.0000604944,0.01107307,0.000004433483,0.00001535242,0.00002496315,0.00003489091,0.9829153,0.004789923,0.0004645139,0.0005847741,0.00001183175],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8987947,0.0004287933,0.09109025,0.0005656497,0.0002313694,0.0002230758,0.003009342,0.002994247,0.002662662],"genre_scores_gemma":[0.9426702,0.0000873785,0.05162381,0.0001176184,0.00005412953,0.00009422331,0.004487067,0.00005542665,0.0008100708],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02099788,"threshold_uncertainty_score":0.04175133,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318185484","doi":"10.1109/bigdata55660.2022.10020371","title":"Understanding Levels of Automation in Human-Machine Collaboration","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Human-Automation Interaction and Safety","field":"Psychology","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Automation; Variety (cybernetics); Computer science; Autonomy; Process (computing); Task (project management); Data science; Risk analysis (engineering); Knowledge management; Systems engineering; Artificial intelligence; Engineering","authors":[{"name":"Glaucia Melo","is_ca":true},{"name":"Nathalia Nascimento","is_ca":true},{"name":"Paulo Alencar","is_ca":true},{"name":"Donald Cowan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6250192258128631,"gpt":0.46840426658817,"spread":0.1566149592246931,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006928116,0.0005644304,0.0003119305,0.00303683,0.002092004,0.006474771,0.0008161165,0.001168145,0.002675985],"category_scores_gemma":[0.02488662,0.0005445731,0.0005362457,0.00123763,0.00708887,0.008912162,0.004271549,0.001345507,0.0004493788],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001812589,"about_ca_system_score_gemma":0.001793445,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002256074,"about_ca_topic_score_gemma":0.001554072,"domain_scores_codex":[0.9915212,0.004480009,0.0004620732,0.000998809,0.001796908,0.000740978],"domain_scores_gemma":[0.9672604,0.01987927,0.005719503,0.002125176,0.002951152,0.002064513],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0003784838,0.000604596,0.2148357,0.001203303,0.0003841084,0.000453762,0.1132765,0.02659161,0.00748954,0.4557799,0.002735176,0.1762674],"study_design_scores_gemma":[0.00006768142,0.000469247,0.21315,0.0005744522,0.0001388922,0.0004595064,0.04133735,0.03164074,0.002238176,0.6760964,0.03362833,0.0001992626],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6772907,0.003696894,0.2050128,0.005572604,0.0001083295,0.0002940454,0.0001369065,0.0002256936,0.107662],"genre_scores_gemma":[0.9893305,0.0003222475,0.009654126,0.00007521779,0.00001944669,0.00005617446,0.00002957314,0.00001297511,0.0004997454],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.006928116,"threshold_uncertainty_score":0.03663981,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318148398","doi":"10.1109/bigdata55660.2022.10020567","title":"Heterogeneity in feature importance and prediction performance for sales at the market and store levels: the case of branded yogurt products in Quebec","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Consumer Attitudes and Food Labeling","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Université de Montréal; McGill University","funders":"HORIZON EUROPE Health","keywords":"Gradient boosting; Random forest; Aggregate (composite); Python (programming language); Loyalty; Computer science; Loyalty business model; Product (mathematics); Aggregate data; Marketing; Machine learning; Business; Statistics","authors":[{"name":"Cameron McRae","is_ca":true},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Laurette Dubé","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.248189087809298,"gpt":0.3494119663719226,"spread":0.1012228785626246,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001452631,0.0007065887,0.0004265893,0.001040764,0.001067477,0.001281907,0.001218274,0.0005966685,0.002441511],"category_scores_gemma":[0.004629441,0.0002111372,0.0007554435,0.001496832,0.0005770521,0.0008285425,0.0005889322,0.001188771,0.0008230084],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007689576,"about_ca_system_score_gemma":0.00485266,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.9598614,"about_ca_topic_score_gemma":0.9728767,"domain_scores_codex":[0.9993728,0.00009424137,0.00002881494,0.0002016116,0.0001217942,0.0001807909],"domain_scores_gemma":[0.9978549,0.0006293877,0.0001155046,0.0002933342,0.0009014956,0.0002052724],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0009962459,0.0004673126,0.8026795,0.0001533116,0.0004741925,0.0003305974,0.0004329208,0.05636028,0.002726048,0.0007878559,0.04266669,0.09192511],"study_design_scores_gemma":[0.00007593433,0.0001219647,0.6335072,0.00005519743,0.0001409767,0.0001285568,0.001027115,0.3509734,0.001638064,0.0005573159,0.01169134,0.00008290732],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9753532,0.0009602467,0.002429241,0.0008159621,0.00006471211,0.00004485552,0.01633692,0.0004074241,0.003587394],"genre_scores_gemma":[0.9663442,0.0001569056,0.002568608,0.0001381237,0.00001801933,0.0000181652,0.02783091,0.0000521985,0.002872916],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0401386,"threshold_uncertainty_score":0.08074993,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024230","doi":"10.1109/bigdata55660.2022.10020333","title":"Thick Data Techniques for Identifying Abnormality in Video Frames for Wireless Capsule Endoscopy","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Gastrointestinal Bleeding Diagnosis and Treatment","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"NOSM University; Lakehead University","funders":"","keywords":"Computer science; Artificial intelligence; Capsule endoscopy; Abnormality; Modality (human–computer interaction); Filter (signal processing); Decision tree; Deep learning; Computer vision; Feature (linguistics); Feature extraction; Pattern recognition (psychology); Medicine; Radiology","authors":[{"name":"Jinan Fiaidhi","is_ca":true},{"name":"Sabah Mohammed","is_ca":true},{"name":"Petros Zezos","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4424567085698399,"gpt":0.4287686708050186,"spread":0.0136880377648213,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009395764,0.001525093,0.0008050377,0.002517628,0.0003381333,0.001259488,0.001005987,0.001135601,0.002621608],"category_scores_gemma":[0.004724199,0.0003496494,0.0008122787,0.001682247,0.0005141721,0.002073682,0.001436554,0.001535428,0.002038891],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005182775,"about_ca_system_score_gemma":0.0005083336,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002412103,"about_ca_topic_score_gemma":0.003547211,"domain_scores_codex":[0.9991678,0.000123373,0.00009514423,0.0001630951,0.0003841132,0.000066485],"domain_scores_gemma":[0.9976776,0.0006010651,0.0003460615,0.0004584233,0.0008265268,0.00009024808],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0007001752,0.0001949194,0.005645827,0.0003753756,0.0001235264,0.0005474265,0.0001486238,0.03420948,0.1056816,0.003903456,0.008865939,0.8396037],"study_design_scores_gemma":[0.00004393686,0.0006266122,0.01171526,0.0001868545,0.0001256578,0.001171204,0.0003272525,0.8455157,0.1006045,0.009470291,0.0300888,0.0001238793],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03202165,0.001925969,0.9601663,0.0005202111,0.0003786385,0.0001929946,0.001330772,0.001533403,0.001930133],"genre_scores_gemma":[0.2419928,0.00225419,0.7459046,0.0003730727,0.0003809446,0.0002692173,0.004016993,0.0002723873,0.004535842],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002621608,"threshold_uncertainty_score":0.008770168,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024146","doi":"10.1109/bigdata55660.2022.10020239","title":"A Stylized Presence Detection System in the Era of Blockchain and Big Data","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Bluetooth and Wireless Communication Technologies","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"","keywords":"Stylized fact; Big data; Computer science; Blockchain; Popularity; Beacon; Identification (biology); Internet of Things; Bluetooth; Data science; Computer security; Telecommunications; Data mining; Wireless","authors":[{"name":"Anastasios Alexandridis","is_ca":true},{"name":"Ghassan Al-Sumaidaee","is_ca":true},{"name":"Rami Alkhudary","is_ca":false},{"name":"Željko Žilić","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2687715235562875,"gpt":0.3297032818409691,"spread":0.06093175828468167,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00334486,0.0004185891,0.0009659063,0.001296477,0.001539469,0.003029916,0.001702998,0.001662271,0.00365429],"category_scores_gemma":[0.009589803,0.0004318468,0.0003513577,0.001105923,0.001816221,0.006981939,0.00508178,0.001729682,0.001842333],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001033262,"about_ca_system_score_gemma":0.001891135,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001919375,"about_ca_topic_score_gemma":0.001605429,"domain_scores_codex":[0.9972715,0.0009131682,0.0002080507,0.0005371309,0.000823263,0.0002468322],"domain_scores_gemma":[0.9931307,0.001840435,0.0006978938,0.00250145,0.00116187,0.0006675685],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002459912,0.0006303397,0.01244654,0.0004712264,0.0001509645,0.002158091,0.002828509,0.06680445,0.04706989,0.4129696,0.0269348,0.4250757],"study_design_scores_gemma":[0.0001153408,0.0003270654,0.0010796,0.00007133393,0.00004904113,0.0006801421,0.0002453186,0.8257682,0.02407894,0.1161432,0.03130398,0.0001379116],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08045241,0.0004983822,0.890103,0.002838904,0.0003899212,0.0005202151,0.000596437,0.0123251,0.01227553],"genre_scores_gemma":[0.8758192,0.0002169575,0.1138895,0.0005084017,0.0001555787,0.0002942332,0.0005382022,0.00018019,0.008397681],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00365429,"threshold_uncertainty_score":0.01768953,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024115","doi":"10.1109/bigdata55660.2022.10020387","title":"An Intelligent Class: The Development Of A Novel Context Capturing Method For The Functional Auto Classification Of Records","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Classifier (UML); Artificial intelligence; Context (archaeology); Machine learning; Process (computing); Statistical classification; Set (abstract data type); One-class classification; Task (project management); Data mining; Engineering","authors":[{"name":"Nathaniel Payne","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5787873684648848,"gpt":0.3602816516620805,"spread":0.2185057168028043,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001139545,0.0005886219,0.0007802515,0.001959912,0.0009808326,0.001872744,0.001624822,0.001009064,0.002369023],"category_scores_gemma":[0.003895996,0.0003212892,0.0007920287,0.001585445,0.0005383082,0.002637029,0.001464588,0.00166772,0.001350311],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007486302,"about_ca_system_score_gemma":0.001378255,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004697255,"about_ca_topic_score_gemma":0.007860785,"domain_scores_codex":[0.9985492,0.0001723529,0.00009733324,0.0005372779,0.0005309103,0.0001129264],"domain_scores_gemma":[0.9983652,0.0004634205,0.0001985503,0.0003297174,0.0005455139,0.00009760083],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002517816,0.0002477051,0.009230618,0.0001402563,0.00006019934,0.0001680775,0.0008191464,0.005475278,0.02205749,0.008737153,0.008674349,0.9441379],"study_design_scores_gemma":[0.00007620944,0.000361556,0.01621988,0.0001319217,0.0001344214,0.000897837,0.0009912517,0.8634073,0.04604265,0.01836937,0.05323703,0.000130618],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03810064,0.0004352102,0.9494147,0.0006675543,0.0002906028,0.000341936,0.0005620894,0.004984863,0.005202341],"genre_scores_gemma":[0.2275788,0.0003090232,0.7614347,0.0005278261,0.0002372326,0.0003492797,0.0009717743,0.0003543165,0.008236966],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004697255,"threshold_uncertainty_score":0.009339869,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4317781640","doi":"10.1109/bigdata55660.2022.10020852","title":"The Lannion report on Big Data and Security Monitoring Research","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Network Security and Intrusion Detection","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Bell (Canada)","funders":"Engineering and Physical Sciences Research Council; Institut national de recherche en informatique et en automatique (INRIA)","keywords":"Big data; Computer science; Data science; Analytics; Data management; Computer security; Computation; Data mining","authors":[{"name":"Laurent d’Orazio","is_ca":false},{"name":"Jalil Boukhobza","is_ca":false},{"name":"Omer Rana","is_ca":false},{"name":"Juba Agoun","is_ca":false},{"name":"Le Gruenwald","is_ca":false},{"name":"Herve Rannou","is_ca":false},{"name":"Elisa Bertino","is_ca":false},{"name":"Mohand-Saïd Hacid","is_ca":false},{"name":"Taofik Saidi","is_ca":false},{"name":"Georges Bossert","is_ca":false},{"name":"Van Long Nguyen Huu","is_ca":true},{"name":"Dimitri Tombroff","is_ca":false},{"name":"Makoto Onizuka","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4622280971601727,"gpt":0.414755484398714,"spread":0.0474726127614587,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02042904,0.003071628,0.001194774,0.01144335,0.001913456,0.009701525,0.002177225,0.00519304,0.01925339],"category_scores_gemma":[0.03141782,0.000784034,0.001683935,0.01180735,0.0014769,0.01123216,0.005483309,0.006553189,0.01657471],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005979531,"about_ca_system_score_gemma":0.01186232,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01846299,"about_ca_topic_score_gemma":0.01978466,"domain_scores_codex":[0.9807779,0.00258835,0.001014226,0.001361258,0.0129599,0.001298387],"domain_scores_gemma":[0.9501753,0.01562389,0.00269892,0.003129947,0.0234815,0.00489043],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005124936,0.00002875942,0.000581548,0.0002403353,0.00002534145,0.00004621841,0.00003889892,0.0001729481,0.0002292928,0.003412974,0.9642434,0.03092918],"study_design_scores_gemma":[0.00001095185,0.00002297472,0.001073586,0.0004339238,0.00002711745,0.0000347879,0.00007852734,0.0002882284,0.0005852003,0.001789473,0.9956332,0.00002202199],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"other","genre_scores_codex":[0.00293765,0.1569454,0.01903385,0.4100804,0.2065157,0.0008573867,0.0317159,0.00512211,0.1667915],"genre_scores_gemma":[0.0302805,0.2713797,0.03607196,0.133557,0.121027,0.001709845,0.09679031,0.003589948,0.3055937],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.02042904,"threshold_uncertainty_score":0.1080404,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024117","doi":"10.1109/bigdata55660.2022.10021046","title":"A novel self-adaptive method for improving patient monitoring with composite early-warning scores","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Healthcare Technology and Patient Monitoring","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Wearable computer; Remote patient monitoring; Real-time computing; Pruning; Wearable technology; Transmission (telecommunications); Adaptive sampling; Machine learning; Artificial intelligence; Data mining; Embedded system; Medicine; Telecommunications","authors":[{"name":"Antonio Iyda Paganelli","is_ca":false},{"name":"Pedro Elkind Velmovitsky","is_ca":true},{"name":"Adriano Branco","is_ca":false},{"name":"Markus Endler","is_ca":false},{"name":"Plinio Pelegrini Morita","is_ca":true},{"name":"Paulo Alencar","is_ca":true},{"name":"Donald Cowan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3307912242688666,"gpt":0.395012270814312,"spread":0.06422104654544541,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008445518,0.000909698,0.0009496451,0.001443181,0.0002552893,0.0007314538,0.000972738,0.0007078245,0.001229351],"category_scores_gemma":[0.003861478,0.0003154649,0.0008179776,0.001077768,0.0002384337,0.001097442,0.0007936604,0.0009402243,0.0006569615],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000247795,"about_ca_system_score_gemma":0.0005265867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001650918,"about_ca_topic_score_gemma":0.001969423,"domain_scores_codex":[0.9991937,0.0001165509,0.00007951752,0.0002687817,0.0002775532,0.00006380535],"domain_scores_gemma":[0.9987805,0.0004388799,0.0001440988,0.0001147353,0.0004465853,0.00007511853],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005653279,0.0005445454,0.01063723,0.0001505727,0.0001314358,0.0002157056,0.0002201974,0.02741588,0.04472598,0.00131331,0.004342791,0.9097371],"study_design_scores_gemma":[0.00007120831,0.0004265525,0.01177297,0.00003044475,0.0001306528,0.0005160868,0.00006283009,0.9583992,0.02193056,0.001877981,0.004719302,0.00006217632],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04525417,0.0004214891,0.9505029,0.0002000573,0.0001547566,0.0001517693,0.0002374235,0.002004031,0.001073536],"genre_scores_gemma":[0.4457489,0.0004597574,0.5482261,0.0004270768,0.0002734835,0.0002944826,0.0007251191,0.0002251729,0.00361995],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.001650918,"threshold_uncertainty_score":0.004466474,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318186095","doi":"10.1109/bigdata55660.2022.10020676","title":"DelphAI: A human-centered approach to time-series forecasting","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Forecasting Techniques and Applications","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Guelph; Vector Institute","funders":"","keywords":"Series (stratigraphy); Time series; Computer science; Machine learning; Geology","authors":[{"name":"Kristina L. Kupferschmidt","is_ca":true},{"name":"Joshua G. Skorburg","is_ca":true},{"name":"Graham W. Taylor","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.762746653460341,"gpt":0.4577294211491272,"spread":0.3050172323112137,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03098363,0.001733542,0.001078629,0.002619781,0.002057609,0.007480372,0.005349159,0.002816634,0.00880087],"category_scores_gemma":[0.04941436,0.001253461,0.001871035,0.001989471,0.004426493,0.005040559,0.008106864,0.005500954,0.003361788],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00251004,"about_ca_system_score_gemma":0.005266347,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002070817,"about_ca_topic_score_gemma":0.002540345,"domain_scores_codex":[0.9732924,0.01978577,0.0008361845,0.002463599,0.003199352,0.000422678],"domain_scores_gemma":[0.9656336,0.02360318,0.001933028,0.004437908,0.003509086,0.0008831054],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002741196,0.0005641765,0.003683204,0.001108861,0.000257822,0.0005738488,0.01029591,0.08584151,0.005907151,0.6437961,0.0138612,0.2338361],"study_design_scores_gemma":[0.00008399101,0.0002177201,0.0007545276,0.0003838415,0.00007586538,0.000242436,0.00162532,0.3706576,0.005063135,0.5429016,0.07785869,0.0001353933],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0009555132,0.00005835968,0.9923248,0.001057167,0.00004451931,0.0002254476,0.0001034029,0.000630876,0.004599962],"genre_scores_gemma":[0.06039764,0.0001959276,0.9342922,0.0005049738,0.00009578666,0.001376533,0.0002908648,0.0003654411,0.002480543],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.03098363,"threshold_uncertainty_score":0.163859,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147664","doi":"10.1109/bigdata55660.2022.10021035","title":"Analysis of Airfare during Pandemic: A Multi-Agent Based Modeling Approach","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Aviation Industry Analysis and Trends","field":"Economics, Econometrics and Finance","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"","keywords":"Pandemic; Coronavirus disease 2019 (COVID-19); Air travel; Computer science; 2019-20 coronavirus outbreak; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Aviation; Business; Risk analysis (engineering); Operations research; Engineering; Aerospace engineering; Infectious disease (medical specialty)","authors":[{"name":"Abdul Mutakabbir","is_ca":true},{"name":"Chung–Horng Lung","is_ca":true},{"name":"Samuel A. Ajila","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5808910784080009,"gpt":0.3499821377569418,"spread":0.2309089406510591,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006742024,0.0008308241,0.0008777208,0.0008611348,0.0006457219,0.001598024,0.001623731,0.002115801,0.002836693],"category_scores_gemma":[0.001355999,0.0005206235,0.001304559,0.0006714755,0.000490061,0.001363373,0.001044675,0.001244758,0.0002787787],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001311995,"about_ca_system_score_gemma":0.001370858,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02847286,"about_ca_topic_score_gemma":0.01438336,"domain_scores_codex":[0.9996204,0.0001654932,0.00002182608,0.00007356124,0.00005644973,0.00006219599],"domain_scores_gemma":[0.9993733,0.0003427964,0.0001227886,0.00001857795,0.00008793166,0.00005464969],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001403491,0.00002691685,0.001170051,0.00002928455,0.0000426655,0.00009187254,0.00004424328,0.9897367,0.0002665119,0.00673196,0.0003971574,0.00144874],"study_design_scores_gemma":[0.000003997069,0.00001177209,0.0002326949,0.000004834554,0.000009353374,0.000009606134,0.00002558362,0.9976906,0.00002719614,0.001591217,0.0003879873,0.000005159083],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2414849,0.002592124,0.7102187,0.004931782,0.0003631054,0.0003539975,0.001897995,0.0004223382,0.03773503],"genre_scores_gemma":[0.9484584,0.001383653,0.03903445,0.0002131936,0.0001523559,0.0002958421,0.0005548195,0.00005290829,0.009854268],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02847286,"threshold_uncertainty_score":0.05661422,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4226052253","doi":"10.1109/bigdata55660.2022.10020281","title":"PMFL: Partial Meta-Federated Learning for heterogeneous tasks and its applications on real-world medical records","year":2022,"lang":"en","type":"preprint","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"Fundamental Research Funds for the Central Universities","keywords":"Computer science; Federated learning; Transfer of learning; Meta learning (computer science); Task (project management); Machine learning; Generalization; Artificial intelligence; Homogeneous; Exploit","authors":[{"name":"Tianyi Zhang","is_ca":false},{"name":"Shirui Zhang","is_ca":false},{"name":"Ziwei Chen","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4329335683524456,"gpt":0.4114458854522562,"spread":0.02148768290018938,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005617348,0.001002542,0.001139513,0.001496944,0.0009552889,0.00141394,0.002614115,0.00226035,0.001647251],"category_scores_gemma":[0.01102421,0.0004120609,0.001236042,0.001379372,0.0008547495,0.003124344,0.002620481,0.001456701,0.0005248367],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001227489,"about_ca_system_score_gemma":0.002009205,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003680521,"about_ca_topic_score_gemma":0.004059264,"domain_scores_codex":[0.9982469,0.0007234514,0.0001345198,0.0004379111,0.0002683064,0.000188905],"domain_scores_gemma":[0.9965102,0.001469671,0.0001902596,0.001144063,0.0004914401,0.0001943087],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006933011,0.0004834754,0.006185242,0.0001751386,0.0002277633,0.0002419269,0.0002527941,0.5033939,0.002866019,0.005703421,0.008904213,0.4708728],"study_design_scores_gemma":[0.00003779728,0.00005926521,0.0003762026,0.00001257617,0.0000168204,0.00008482819,0.000032468,0.9886401,0.001768169,0.008003837,0.0009586975,0.00000931082],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04903661,0.0005332097,0.9430329,0.0009310658,0.0000863206,0.000158156,0.0003489488,0.004984394,0.0008883804],"genre_scores_gemma":[0.5784202,0.0001728245,0.4171147,0.0005365901,0.00007178167,0.0002657374,0.001326612,0.0002207224,0.001870812],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005617348,"threshold_uncertainty_score":0.02970773,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147681","doi":"10.1109/bigdata55660.2022.10020382","title":"A Framework for the Design, Development, Testing and Deployment of Reliable Big Data Platforms","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Distributed systems and fault tolerance","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ontario Tech University","funders":"","keywords":"Software deployment; Computer science; Scalability; Big data; Distributed computing; Reliability (semiconductor); Fault tolerance; Process (computing); Pipeline (software); Data science; Systems engineering; Software engineering; Database; Engineering; Data mining","authors":[{"name":"Carolyn McGregor","is_ca":true},{"name":"Catherine Inibhunu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5304000793599303,"gpt":0.3686517544763004,"spread":0.1617483248836298,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01148915,0.001631175,0.001015096,0.001780759,0.001970628,0.006735869,0.005956908,0.003560612,0.002529058],"category_scores_gemma":[0.01263356,0.001466992,0.001946866,0.001009358,0.005299695,0.005598854,0.005599977,0.004431737,0.001224309],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002779607,"about_ca_system_score_gemma":0.008199648,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005929203,"about_ca_topic_score_gemma":0.005731648,"domain_scores_codex":[0.9932532,0.002259456,0.0006739167,0.0007474697,0.002508733,0.000557156],"domain_scores_gemma":[0.9928603,0.002138383,0.0007730268,0.001728577,0.001648444,0.0008512865],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00004258617,0.0001440316,0.0009847308,0.0005475085,0.00006967544,0.0005291897,0.0009571418,0.1611515,0.005667023,0.7894374,0.004581101,0.03588817],"study_design_scores_gemma":[0.00009846102,0.0003599758,0.0004303649,0.0008004114,0.0000956666,0.0004501413,0.0006494134,0.5183591,0.0076939,0.350369,0.1205662,0.0001273943],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001612364,0.0002504588,0.9930333,0.0008742964,0.00006939592,0.0004124715,0.00005991506,0.0007243783,0.0029634],"genre_scores_gemma":[0.05570433,0.0006088589,0.9397196,0.0002549409,0.00007829457,0.001006164,0.0001972275,0.0002430444,0.002187521],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01148915,"threshold_uncertainty_score":0.06076115,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318148079","doi":"10.1109/bigdata55660.2022.10020681","title":"Don’t Blindly Use Data: Towards a Data Statement for Computational Financial Research","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Statement (logic); Data science; Data quality; Context (archaeology); Problem statement; Focus (optics); Quality (philosophy); Work (physics); Management science","authors":[{"name":"Stacey Taylor","is_ca":true},{"name":"Vlado Kešelj","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7268565196499841,"gpt":0.4798004812067899,"spread":0.2470560384431942,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.4443332,0.002405703,0.004070159,0.01594495,0.008088636,0.03768529,0.009397742,0.01716395,0.005826397],"category_scores_gemma":[0.6688346,0.002919219,0.003267862,0.02407745,0.03342992,0.07325903,0.02152779,0.03277919,0.004272322],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.008956787,"about_ca_system_score_gemma":0.03215597,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00399683,"about_ca_topic_score_gemma":0.002960807,"domain_scores_codex":[0.604546,0.282538,0.04243176,0.01332964,0.05395398,0.003200635],"domain_scores_gemma":[0.1819879,0.6259692,0.03656094,0.07859792,0.07081633,0.006067723],"domain_codex":null,"domain_gemma":"reporting","domain_candidate":"reporting","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002292661,0.0001200236,0.003176385,0.00133792,0.00007694487,0.0001477227,0.003967052,0.002378141,0.0003914054,0.8617699,0.05552477,0.07088062],"study_design_scores_gemma":[0.0001088385,0.00009125259,0.0005730687,0.003085067,0.00003877255,0.0001375603,0.00239775,0.009284213,0.0008814851,0.8529316,0.1303267,0.0001437949],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003764737,0.003295165,0.8015593,0.1730644,0.003013591,0.001497477,0.003434782,0.0009451883,0.009425396],"genre_scores_gemma":[0.06475701,0.003105255,0.8833661,0.03233742,0.003521182,0.005506236,0.004717309,0.0007203915,0.001969124],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.5556668,"threshold_uncertainty_score":0.6852361,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147813","doi":"10.1109/bigdata55660.2022.10021073","title":"Systematic Analysis of Public Transit Data Availability in Canada","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Traffic Prediction and Management Techniques","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Carleton University","funders":"","keywords":"Public transport; Computer science; Transit (satellite); Transport engineering; Data science; Engineering","authors":[{"name":"Kevin Dick","is_ca":true},{"name":"Azizul Hasan","is_ca":true},{"name":"Jamil Dergham","is_ca":true},{"name":"A. J. Clarke","is_ca":true},{"name":"Hoda Khalil","is_ca":true},{"name":"Gabriel Wainer","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2616339857047169,"gpt":0.3015119091661445,"spread":0.0398779234614276,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02141101,0.0009038795,0.001176173,0.0196394,0.005032693,0.004196256,0.00305268,0.0006851759,0.002207376],"category_scores_gemma":[0.1030733,0.0009335437,0.001340445,0.05423275,0.001996239,0.00130056,0.003752584,0.001289694,0.0004651565],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.07039544,"about_ca_system_score_gemma":0.2171152,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.994211,"about_ca_topic_score_gemma":0.995789,"domain_scores_codex":[0.9609146,0.004096286,0.004018577,0.004524271,0.02342962,0.003016848],"domain_scores_gemma":[0.8352763,0.02502377,0.0134646,0.009590514,0.1131212,0.003523588],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000383519,0.00009560536,0.7346997,0.006077922,0.001568215,0.0006033641,0.007399533,0.003689566,0.001140749,0.008620101,0.1499504,0.08577131],"study_design_scores_gemma":[0.00006553488,0.00004990445,0.7944499,0.004139308,0.0007057446,0.0002063171,0.005833281,0.005428701,0.001723728,0.0009410449,0.1861723,0.0002842493],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.2736846,0.01551597,0.01005129,0.004478617,0.0003418226,0.0016984,0.6749789,0.000739411,0.01851103],"genre_scores_gemma":[0.6276519,0.008442282,0.01598907,0.001322008,0.00006796268,0.001816734,0.3389808,0.0003664286,0.005362959],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.07039544,"threshold_uncertainty_score":0.5107571,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4320024084","doi":"10.1109/bigdata55660.2022.10020739","title":"The Dimensional Analysis of Data Flow Programs That Include Multidimensional and User-Defined Functions","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Victoria","funders":"University of Victoria","keywords":"Computer science; Dimension (graph theory); Curse of dimensionality; Multidimensional analysis; Data mining; Flow (mathematics); Rank (graph theory); Theoretical computer science; Data flow diagram; Raw data; Artificial intelligence; Database; Programming language; Mathematics; Statistics","authors":[{"name":"Abdulmonem I. Shennat","is_ca":true},{"name":"William W. Wadge","is_ca":true},{"name":"Alex Kuo","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6122670306410279,"gpt":0.4310872843677545,"spread":0.1811797462732734,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01366761,0.002004721,0.001517503,0.003373741,0.001884715,0.006292379,0.003386808,0.001605832,0.005202002],"category_scores_gemma":[0.0419398,0.001446214,0.004388218,0.003394512,0.004661188,0.01256151,0.007018921,0.006038729,0.001728293],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003124471,"about_ca_system_score_gemma":0.004133917,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002839199,"about_ca_topic_score_gemma":0.002566825,"domain_scores_codex":[0.9879171,0.004846397,0.001342913,0.001722649,0.003520182,0.0006507139],"domain_scores_gemma":[0.972822,0.01711966,0.001907226,0.004471659,0.003319425,0.0003601137],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001553949,0.0001449554,0.004179864,0.0007076961,0.0001249561,0.0001611818,0.001004652,0.06679002,0.002882116,0.7203133,0.007593222,0.1959427],"study_design_scores_gemma":[0.00004154935,0.00009548967,0.0005128339,0.0002653432,0.000052946,0.0002080678,0.0002170134,0.5095927,0.007961671,0.4522032,0.02876033,0.00008896195],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002025247,0.000169175,0.9953617,0.0003361941,0.00003236456,0.00009266515,0.00009693218,0.0009850571,0.0009006074],"genre_scores_gemma":[0.03765425,0.0002877321,0.9587665,0.0003484639,0.0001014173,0.0004325172,0.0003867281,0.0006483166,0.001374013],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01366761,"threshold_uncertainty_score":0.07228214,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318147483","doi":"10.1109/bigdata55660.2022.10020802","title":"A Graph-based Analysis Approach to Cluster Lifetime Dynamics","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Cluster analysis; Cluster (spacecraft); Graph; Data mining; Outlier; Power graph analysis; Theoretical computer science; Visualization; Data visualization; Artificial intelligence","authors":[{"name":"Ivens Portugal","is_ca":true},{"name":"Paulo Alencar","is_ca":true},{"name":"Donald Cowan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2103282896521805,"gpt":0.3246473363864822,"spread":0.1143190467343017,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001165529,0.00115312,0.0007435842,0.006532225,0.001211625,0.00272646,0.001493524,0.0009755865,0.002712717],"category_scores_gemma":[0.005771438,0.0005449838,0.00131515,0.006080729,0.001104724,0.00389019,0.001173108,0.001763294,0.0006572746],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00224258,"about_ca_system_score_gemma":0.001240463,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01275642,"about_ca_topic_score_gemma":0.009784157,"domain_scores_codex":[0.999109,0.0002629965,0.00005390065,0.0002708457,0.00024261,0.00006056931],"domain_scores_gemma":[0.9966828,0.00163069,0.000412068,0.0004428435,0.000713564,0.0001179743],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001281186,0.0001775587,0.008137002,0.0004987851,0.0003410507,0.0005536452,0.001678903,0.415041,0.01046348,0.3437357,0.01324797,0.2059969],"study_design_scores_gemma":[0.00001007027,0.0000391394,0.001748223,0.00005193561,0.00005980774,0.0002483507,0.0002989972,0.8276446,0.001779465,0.1519266,0.01614624,0.00004658267],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00620072,0.0004198831,0.9896271,0.0005050937,0.00004472729,0.00008770117,0.0005651669,0.0005745456,0.00197499],"genre_scores_gemma":[0.2275273,0.00174499,0.7622585,0.0003407202,0.0001818756,0.0003023804,0.002170008,0.0004257492,0.005048508],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01275642,"threshold_uncertainty_score":0.0253644,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4318187617","doi":"10.1109/bigdata55660.2022.10020838","title":"A Semisupervised Approach to Predicting a Twitch Streamer’s Growth based on the Game Streamed","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Big Data (Big Data)","topic":"Consumer Market Behavior and Pricing","field":"Business, Management and Accounting","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"","keywords":"Entertainment; Computer science; Meaning (existential); Set (abstract data type); Order (exchange); Approximation error; Work (physics); Artificial intelligence; Simulation; Algorithm; Economics; Engineering","authors":[{"name":"Andrew Dybka","is_ca":true},{"name":"Dominic Kocjan","is_ca":true},{"name":"Samuel A. Ajila","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2948102546880593,"gpt":0.3133429269993256,"spread":0.01853267231126632,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001637205,0.0007463216,0.0007132775,0.0014695,0.0003535495,0.001001891,0.001444098,0.0009381885,0.001127402],"category_scores_gemma":[0.0064337,0.0003809756,0.0005527143,0.0007737343,0.0004758436,0.001272666,0.0005309081,0.00107146,0.0005200967],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001062446,"about_ca_system_score_gemma":0.0007765017,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009389636,"about_ca_topic_score_gemma":0.01410846,"domain_scores_codex":[0.9992132,0.0003006066,0.0000500582,0.0002462058,0.0001145518,0.00007533125],"domain_scores_gemma":[0.9954017,0.002741508,0.0006631749,0.0002277258,0.0007574411,0.0002083664],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004491725,0.000760645,0.05473229,0.0001286983,0.0002227753,0.0002613897,0.0002925811,0.7776881,0.003452183,0.005961948,0.00618054,0.1498697],"study_design_scores_gemma":[0.000002213966,0.0000146494,0.0009938309,0.000002395339,0.000003181496,0.00001186197,0.00001075164,0.9978169,0.0001779425,0.0008410031,0.0001219893,0.000003360499],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4005104,0.0004430587,0.5903531,0.001439149,0.0001019576,0.0002656833,0.001132804,0.001473019,0.004280906],"genre_scores_gemma":[0.9202433,0.0001093417,0.07526793,0.0001460676,0.00007375338,0.0001254207,0.001093277,0.00006205679,0.002879023],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009389636,"threshold_uncertainty_score":0.01866996,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}