{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":47,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":47,"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":"2d95964af09c","filters":{"venue":"ACM Transactions on Knowledge Discovery from Data"}},"results":[{"id":"W2180566385","doi":"10.1145/2733381","title":"Hierarchical Density Estimates for Data Clustering, Visualization, and Outlier Detection","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":857,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Cluster analysis; Computer science; Visualization; Outlier; Anomaly detection; Data mining; Hierarchy; Hierarchical clustering; Pattern recognition (psychology); CURE data clustering algorithm; Mathematics; Artificial intelligence; Correlation clustering","authors":[{"name":"Ricardo J. G. B. Campello","is_ca":false},{"name":"Davoud Moulavi","is_ca":true},{"name":"Arthur Zimek","is_ca":false},{"name":"Jörg Sander","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09123092616616002,"gpt":0.3467405532963631,"spread":0.255509627130203,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004046186,0.00192391,0.001780359,0.004710677,0.001004488,0.003155233,0.00315768,0.001755723,0.003611252],"category_scores_gemma":[0.01844155,0.001238934,0.002099178,0.005171964,0.001430842,0.003033898,0.003300014,0.003781568,0.002406699],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002259806,"about_ca_system_score_gemma":0.002579156,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00835002,"about_ca_topic_score_gemma":0.005726387,"domain_scores_codex":[0.9958762,0.001731435,0.0002050647,0.0005371328,0.001494855,0.0001552614],"domain_scores_gemma":[0.9937251,0.003135016,0.0006781687,0.0009870902,0.001324417,0.0001501807],"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":[0.0001081178,0.0001071201,0.001466691,0.0006635342,0.000245612,0.0001679101,0.0006395039,0.3297228,0.008275618,0.3250408,0.01450657,0.3190558],"study_design_scores_gemma":[0.00001299644,0.00002818374,0.000495629,0.00006713119,0.00002847878,0.00009259838,0.00006836481,0.8907157,0.002891076,0.09389562,0.01165414,0.00005008999],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0002115003,0.0001250232,0.9989655,0.00004067623,0.0000106695,0.00002532772,0.00005225393,0.0003901844,0.0001788168],"genre_scores_gemma":[0.02025649,0.0004756885,0.9776962,0.00004800202,0.00007288258,0.0002803237,0.0003486545,0.0002959502,0.0005258215],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00835002,"threshold_uncertainty_score":0.02139854,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2045929671","doi":"10.1145/1376815.1376819","title":"Semantic text similarity using corpus-based word similarity and string similarity","year":2008,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Topic Modeling","field":"Computer Science","cited_by":485,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Similarity (geometry); Semantic similarity; String metric; Artificial intelligence; Word (group theory); Natural language processing; Longest common subsequence problem; Focus (optics); String searching algorithm; String (physics); Representation (politics); Information retrieval; Matching (statistics); Pattern matching; Mathematics; Algorithm","authors":[{"name":"Aminul Islam","is_ca":true},{"name":"Diana Inkpen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1204337695552436,"gpt":0.3074205502158795,"spread":0.186986780660636,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002177997,0.00092376,0.001464652,0.01774631,0.0009860167,0.002976211,0.001630763,0.001285142,0.003514882],"category_scores_gemma":[0.01814917,0.000312062,0.001163446,0.01682147,0.001240919,0.005136708,0.002024637,0.00100161,0.001902967],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009804009,"about_ca_system_score_gemma":0.001356322,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001658718,"about_ca_topic_score_gemma":0.001637855,"domain_scores_codex":[0.9943928,0.001124886,0.0007004199,0.001007105,0.002637846,0.0001368597],"domain_scores_gemma":[0.9921194,0.003030686,0.0009637382,0.00113907,0.00249476,0.0002522486],"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":[0.0005017755,0.0004054433,0.01289896,0.001379963,0.0005654665,0.0004173881,0.001131468,0.02129064,0.02876124,0.05392328,0.01116528,0.8675591],"study_design_scores_gemma":[0.0001940287,0.0009286882,0.02707705,0.0004368933,0.0005551387,0.002814416,0.001698498,0.6720217,0.04891734,0.1823989,0.06249095,0.0004662709],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05191324,0.002126402,0.930313,0.0002680698,0.0004651728,0.0006221626,0.00230973,0.002587464,0.009394729],"genre_scores_gemma":[0.2841628,0.001106184,0.7038279,0.000128194,0.0005406998,0.001079478,0.00547775,0.0004178632,0.003259144],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01774631,"threshold_uncertainty_score":0.01175845,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3120491054","doi":"10.1145/3424672","title":"Knowledge Graph Embedding for Link Prediction","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":476,"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":"Computer science; Variety (cybernetics); Embedding; Link (geometry); Graph; Scale (ratio); Knowledge graph; Data science; Machine learning; Data mining; Artificial intelligence; Theoretical computer science","authors":[{"name":"Andrea Rossi","is_ca":false},{"name":"Denilson Barbosa","is_ca":true},{"name":"Donatella Firmani","is_ca":false},{"name":"Antonio Matinata","is_ca":false},{"name":"Paolo Merialdo","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06056504407758513,"gpt":0.3279314449345395,"spread":0.2673664008569543,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001166327,0.001616835,0.001066593,0.005101556,0.0006195268,0.001677924,0.001806926,0.001522693,0.005325116],"category_scores_gemma":[0.01310774,0.0004635409,0.0008989808,0.005936936,0.0005482244,0.004817743,0.001373057,0.00197378,0.003182352],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008345653,"about_ca_system_score_gemma":0.00129753,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006439603,"about_ca_topic_score_gemma":0.009317739,"domain_scores_codex":[0.9982754,0.0003791427,0.0001117435,0.0005762422,0.0005585456,0.00009890614],"domain_scores_gemma":[0.9929838,0.004046838,0.0005855557,0.001444301,0.0008199135,0.0001195538],"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":[0.0001773007,0.0002822046,0.007204503,0.0007703234,0.0002234612,0.0001986126,0.0001673307,0.1762987,0.002172216,0.0213937,0.03528995,0.7558216],"study_design_scores_gemma":[0.00001907525,0.00007008718,0.00154271,0.0001420028,0.00007975237,0.0002312628,0.0001117621,0.9130898,0.003819365,0.06280353,0.01806149,0.00002918315],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03308751,0.006256744,0.9295486,0.001097339,0.0003032034,0.0002986421,0.0102201,0.00946766,0.009720197],"genre_scores_gemma":[0.4394885,0.003700154,0.5207769,0.0004367583,0.0002938997,0.0002922476,0.02822595,0.0005918446,0.00619373],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006439603,"threshold_uncertainty_score":0.01781428,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2257550357","doi":"10.1145/2976744","title":"Scalable and Accurate Online Feature Selection for Big Data","year":2016,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":145,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"National Natural Science Foundation of China","keywords":"Feature selection; Pairwise comparison; Big data; Computer science; Scalability; Feature (linguistics); Benchmark (surveying); Curse of dimensionality; Selection (genetic algorithm); Artificial intelligence; Data mining; Dimensionality reduction; Set (abstract data type); Machine learning; Pattern recognition (psychology); Database","authors":[{"name":"Kui Yu","is_ca":true},{"name":"Xindong Wu","is_ca":false},{"name":"Wei Ding","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1056308529088211,"gpt":0.3196476373088136,"spread":0.2140167843999925,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001923,0.001902947,0.00193513,0.002402731,0.001001444,0.001264388,0.002273472,0.00103453,0.001837085],"category_scores_gemma":[0.008684577,0.0006496051,0.001401741,0.003210819,0.0006331607,0.003054542,0.001533854,0.001808675,0.00107384],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007221173,"about_ca_system_score_gemma":0.001359197,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00394371,"about_ca_topic_score_gemma":0.005423269,"domain_scores_codex":[0.998068,0.0004296343,0.0001195535,0.0004853065,0.0007207925,0.0001767154],"domain_scores_gemma":[0.9958977,0.001999119,0.0003850236,0.0009573074,0.0005905288,0.0001704668],"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":[0.000654714,0.0004579989,0.006358625,0.0002509479,0.0002728295,0.0004185618,0.0001857175,0.2250386,0.01693994,0.004551035,0.01745765,0.7274134],"study_design_scores_gemma":[0.00005028029,0.00009853363,0.001167441,0.000008767111,0.00002547746,0.0001164455,0.00005608439,0.9821215,0.003221507,0.01083589,0.00227983,0.00001823285],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02669664,0.001049535,0.9671476,0.0003862286,0.00008806078,0.0001345761,0.0005054205,0.003355462,0.0006364111],"genre_scores_gemma":[0.4440493,0.0005337617,0.5486611,0.0003536175,0.0003011262,0.0004578608,0.003443822,0.000294214,0.00190533],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00394371,"threshold_uncertainty_score":0.01016986,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2038873345","doi":"10.1145/1857947.1857950","title":"Centralized and Distributed Anonymization for High-Dimensional Healthcare Data","year":2010,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":133,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Ontario Tech University; Concordia University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Data anonymization; Scalability; Data sharing; Information privacy; Information sharing; Health care; Data mining; Data science; Computer security; Database; World Wide Web; Medicine","authors":[{"name":"Noman Mohammed","is_ca":true},{"name":"Benjamin C. M. Fung","is_ca":true},{"name":"Patrick C. K. Hung","is_ca":true},{"name":"Cheuk‐Kwong Lee","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06175333628821846,"gpt":0.3219685442326207,"spread":0.2602152079444023,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01167583,0.0004538924,0.001031146,0.001015774,0.002284541,0.002900959,0.001769408,0.001288516,0.0005819566],"category_scores_gemma":[0.02173689,0.0004702731,0.001205825,0.002388779,0.003047314,0.005116151,0.003765543,0.001792635,0.0002378479],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002227637,"about_ca_system_score_gemma":0.003029186,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001551167,"about_ca_topic_score_gemma":0.001665782,"domain_scores_codex":[0.9844049,0.009379582,0.0007042715,0.002057423,0.002861128,0.000592785],"domain_scores_gemma":[0.9695189,0.01165719,0.002502693,0.01454605,0.001463948,0.000311212],"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":[0.0005314783,0.0001889568,0.009086363,0.0002528685,0.0002716662,0.0007134747,0.002411925,0.4419955,0.008131644,0.351915,0.006798383,0.1777028],"study_design_scores_gemma":[0.0000518512,0.00008195516,0.001908388,0.00003666709,0.00005817091,0.000579536,0.0006880186,0.7360973,0.007125081,0.2423706,0.01095743,0.00004502067],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02826652,0.0003474767,0.9682034,0.001194602,0.00004857519,0.0001006981,0.0001739526,0.0002463675,0.001418373],"genre_scores_gemma":[0.7828609,0.0004716853,0.2137386,0.0003004036,0.0001532177,0.0002269268,0.0006269678,0.00006408319,0.001557113],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01167583,"threshold_uncertainty_score":0.06174845,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2096281132","doi":"10.1145/1460797.1460799","title":"Mining frequent cross-graph quasi-cliques","year":2009,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":91,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Division of Information and Intelligent Systems","keywords":"Heuristics; Computer science; Clique; Data mining; Vertex (graph theory); Graph; Theoretical computer science; Combinatorics; Mathematics","authors":[{"name":"Daxin Jiang","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05441437221485232,"gpt":0.3372334490668411,"spread":0.2828190768519887,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002263108,0.000797415,0.001630815,0.002742835,0.001765262,0.001966204,0.002569589,0.001500838,0.002296163],"category_scores_gemma":[0.01356022,0.001116731,0.001859942,0.003231357,0.001189523,0.005258562,0.002143375,0.001442086,0.0003954698],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009719478,"about_ca_system_score_gemma":0.001221006,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003856919,"about_ca_topic_score_gemma":0.005440415,"domain_scores_codex":[0.9969391,0.0007934471,0.0001690102,0.001225605,0.0005564271,0.0003164201],"domain_scores_gemma":[0.9824293,0.01115739,0.002299686,0.001914452,0.001502833,0.0006963699],"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":[0.001332821,0.001216596,0.1023359,0.001956437,0.001693472,0.005463622,0.002171993,0.4789174,0.02022272,0.1849571,0.0280882,0.1716437],"study_design_scores_gemma":[0.00007655209,0.00008587109,0.006340689,0.00005856347,0.00009082631,0.0009850245,0.0005107828,0.8424877,0.002494572,0.1420558,0.00477494,0.00003852961],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4464625,0.001082401,0.540214,0.002123375,0.00006974472,0.0004135698,0.005588637,0.0008243404,0.00322147],"genre_scores_gemma":[0.7608451,0.0005047754,0.2262318,0.0003565323,0.0001522936,0.0002988084,0.009530523,0.0001513126,0.001928922],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003856919,"threshold_uncertainty_score":0.01196861,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2400248106","doi":"10.1145/2724720","title":"Improving Top-N Recommendation for Cold-Start Users via Cross-Domain Information","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":88,"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":"Cold start (automotive); Computer science; Domain (mathematical analysis); Matrix decomposition; Cluster analysis; Precision and recall; Recommender system; Data mining; Artificial intelligence; Machine learning; Information retrieval; Mathematics; Engineering","authors":[{"name":"Nima Mirbakhsh","is_ca":true},{"name":"Charles X. Ling","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06982123719770345,"gpt":0.3200112914309484,"spread":0.2501900542332449,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003890548,0.002176764,0.003067326,0.00378822,0.00173694,0.001736884,0.002423364,0.002057101,0.002699725],"category_scores_gemma":[0.01493012,0.0008323142,0.001816801,0.003836493,0.0006673021,0.004701747,0.00142881,0.002315957,0.003809893],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007710382,"about_ca_system_score_gemma":0.001823495,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01556823,"about_ca_topic_score_gemma":0.05330112,"domain_scores_codex":[0.9969938,0.0008374988,0.0002259584,0.001180736,0.000515231,0.0002467999],"domain_scores_gemma":[0.9892943,0.004845143,0.0006473848,0.002515388,0.002227233,0.0004706206],"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":[0.001627634,0.002435092,0.09406205,0.001281783,0.001253938,0.0007392413,0.001155407,0.09770698,0.01949599,0.004815131,0.06753005,0.7078967],"study_design_scores_gemma":[0.0001512936,0.000486707,0.01200732,0.0001843449,0.0004780572,0.001104704,0.0007408861,0.9502414,0.009766207,0.01151474,0.01311419,0.0002102466],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.269363,0.01356852,0.6920493,0.001405924,0.0008338083,0.0004800938,0.004318703,0.00600376,0.0119769],"genre_scores_gemma":[0.6592122,0.001697176,0.3218853,0.0007986244,0.0004175788,0.0001782186,0.007461824,0.0003676688,0.00798136],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01556823,"threshold_uncertainty_score":0.0309552,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2069857456","doi":"10.1145/1993077.1993080","title":"Can the Utility of Anonymized Data be Used for Privacy Breaches?","year":2011,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":84,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Office of Integrative Activities; Office of International Science and Engineering; Division of Information and Intelligent Systems; Research Grants Council, University Grants Committee","keywords":"Data publishing; Computer science; Data anonymization; Adversary; Anonymity; Information privacy; k-anonymity; Closeness; Internet privacy; Data mining; Computer security; Data science; Publishing; Mathematics","authors":[{"name":"Raymond Chi-Wing Wong","is_ca":false},{"name":"Ada Wai-Chee Fu","is_ca":false},{"name":"Ke Wang","is_ca":true},{"name":"Philip S. Yu","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2143621106515954,"gpt":0.3377835758204947,"spread":0.1234214651688993,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04078216,0.001009801,0.00203924,0.003541938,0.002891614,0.01012915,0.004005242,0.007099167,0.002915004],"category_scores_gemma":[0.1596208,0.001009293,0.001942677,0.006119642,0.01262315,0.03129663,0.007270992,0.006013757,0.001910177],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003098789,"about_ca_system_score_gemma":0.002641949,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008672254,"about_ca_topic_score_gemma":0.0004869439,"domain_scores_codex":[0.9372321,0.03898875,0.002675459,0.005175985,0.01391273,0.002014912],"domain_scores_gemma":[0.7945338,0.07921044,0.01081213,0.1047579,0.00952527,0.001160375],"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":[0.0009675998,0.0001960548,0.01646381,0.0006551846,0.0006397338,0.0009428842,0.004228143,0.02973156,0.003289681,0.6696098,0.01728906,0.2559865],"study_design_scores_gemma":[0.00007523049,0.0002168803,0.002209065,0.0004855603,0.0001483319,0.001999051,0.002492058,0.04045432,0.01028277,0.8801966,0.06133516,0.0001050235],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1220613,0.009972642,0.6868676,0.1237842,0.00166768,0.0005527523,0.00219913,0.001549962,0.05134471],"genre_scores_gemma":[0.9130582,0.002738769,0.07262773,0.004907168,0.0007735603,0.0002442447,0.0005928378,0.0002940613,0.004763343],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.04078216,"threshold_uncertainty_score":0.2156793,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2602756700","doi":"10.1145/3046941","title":"An Influence Propagation View of PageRank","year":2017,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Opinion Dynamics and Social Influence","field":"Physics and Astronomy","cited_by":82,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"National Natural Science Foundation of China; National Science Fund for Distinguished Young Scholars; Youth Innovation Promotion Association; National Science Foundation","keywords":"PageRank; Computer science; Ranking (information retrieval); Computation; Constraint (computer-aided design); Node (physics); Theoretical computer science; Data mining; Information retrieval; Algorithm; Mathematics","authors":[{"name":"Qi Liu","is_ca":false},{"name":"Biao Xiang","is_ca":false},{"name":"Nicholas Jing Yuan","is_ca":false},{"name":"Enhong Chen","is_ca":false},{"name":"Hui Xiong","is_ca":false},{"name":"Zheng Yi","is_ca":false},{"name":"Yang Yu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04113487103308461,"gpt":0.345878011987555,"spread":0.3047431409544704,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001372822,0.0009984666,0.0009339707,0.003115342,0.0008998935,0.00316679,0.001761522,0.001400025,0.005173957],"category_scores_gemma":[0.007133948,0.0004701633,0.0008065512,0.002550086,0.002154153,0.006023257,0.001033648,0.002168429,0.001449826],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001666118,"about_ca_system_score_gemma":0.0009021221,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004847981,"about_ca_topic_score_gemma":0.003735316,"domain_scores_codex":[0.998477,0.0004269178,0.00005133149,0.0003170171,0.0006237673,0.0001038719],"domain_scores_gemma":[0.9968976,0.00168493,0.0003345018,0.0003427686,0.0005911713,0.0001489215],"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":[0.00002769503,0.00003273383,0.0007888108,0.0001680523,0.00005474352,0.0001150263,0.0002242625,0.06631402,0.00117399,0.8932958,0.005389822,0.03241502],"study_design_scores_gemma":[0.00001987828,0.00004197473,0.0004022416,0.00003806987,0.00003688884,0.00014613,0.0000559068,0.3878573,0.0006470433,0.5943128,0.01641701,0.00002475461],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00864299,0.001885635,0.9496561,0.002608612,0.0002054332,0.00007108715,0.0002864632,0.0003949816,0.0362487],"genre_scores_gemma":[0.7756464,0.006628706,0.1793624,0.0009647612,0.002081324,0.0003380799,0.000645915,0.0002585665,0.0340738],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005173957,"threshold_uncertainty_score":0.01730859,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3027864066","doi":"10.1145/3377939","title":"Story Forest","year":2020,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":60,"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":"Timeline; Computer science; Event (particle physics); Information retrieval; The Internet; Novelty; World Wide Web; Cluster analysis; Set (abstract data type); News aggregator; Graph; Data science; Artificial intelligence; History","authors":[{"name":"Bang Liu","is_ca":true},{"name":"Fred X. Han","is_ca":true},{"name":"Di Niu","is_ca":true},{"name":"Linglong Kong","is_ca":true},{"name":"Kunfeng Lai","is_ca":false},{"name":"Yu Xu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06664237805374586,"gpt":0.3114345391788544,"spread":0.2447921611251086,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006922385,0.001755382,0.0006985309,0.003107492,0.001331704,0.002605887,0.002297776,0.001302518,0.07456609],"category_scores_gemma":[0.003858663,0.0006255303,0.00185235,0.002570536,0.0003868566,0.004175462,0.002283463,0.001478413,0.03810922],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006778545,"about_ca_system_score_gemma":0.001383188,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004977882,"about_ca_topic_score_gemma":0.01420694,"domain_scores_codex":[0.9991633,0.0001269304,0.00005922194,0.0003169754,0.0002345139,0.00009914594],"domain_scores_gemma":[0.9990376,0.0003215243,0.00006552265,0.0002593464,0.0002348382,0.0000811186],"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":[0.0003507535,0.0001809252,0.002870011,0.0009049827,0.0001305217,0.0004553862,0.000242131,0.01235606,0.002915271,0.02594774,0.4491531,0.5044931],"study_design_scores_gemma":[0.0001353358,0.0001163247,0.00206778,0.0002357142,0.0001019398,0.001017794,0.000442149,0.1716942,0.006384603,0.08241154,0.735324,0.00006871013],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01626917,0.003983403,0.5543593,0.002554548,0.0015691,0.002101169,0.1660955,0.08918893,0.1638788],"genre_scores_gemma":[0.09396937,0.002320024,0.4876962,0.001304079,0.0003954143,0.001380324,0.3339684,0.005827719,0.07313858],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.07456609,"threshold_uncertainty_score":0.2494484,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1973522621","doi":"10.1145/2757281","title":"Refining Social Graph Connectivity via Shortcut Edge Addition","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":52,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Graph; Computation; Theoretical computer science; Algorithm; Data mining","authors":[{"name":"Manos Papagelis","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08509666336746628,"gpt":0.3278070009381637,"spread":0.2427103375706974,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001886966,0.00159817,0.001770165,0.003588905,0.001247882,0.001308111,0.002114056,0.002158473,0.004177266],"category_scores_gemma":[0.01707276,0.0008754465,0.001572851,0.001989403,0.001308685,0.0036042,0.002545883,0.001977112,0.0006758735],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009465799,"about_ca_system_score_gemma":0.001216568,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003753771,"about_ca_topic_score_gemma":0.007454382,"domain_scores_codex":[0.9983686,0.000517352,0.00006941529,0.0004163206,0.000511595,0.0001167725],"domain_scores_gemma":[0.9888445,0.007804366,0.0009782662,0.001048128,0.0009603046,0.000364498],"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":[0.0002956913,0.0002884455,0.005943553,0.0006089823,0.0001901919,0.0006607939,0.0004896233,0.6297741,0.01099347,0.05502411,0.008154308,0.2875768],"study_design_scores_gemma":[0.0000360454,0.00008316633,0.0005762115,0.00003582282,0.00003829811,0.000168695,0.00007473069,0.9575903,0.002575051,0.0361926,0.002609476,0.00001958964],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06978099,0.0005801123,0.924515,0.0003733182,0.00008555193,0.0002985526,0.000394555,0.001410531,0.002561403],"genre_scores_gemma":[0.3823691,0.0004641385,0.6111705,0.0003105763,0.0001345061,0.0004813865,0.001417369,0.0003489769,0.003303441],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004177266,"threshold_uncertainty_score":0.01397437,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1965792576","doi":"10.1145/1376815.1376816","title":"Joint cluster analysis of attribute data and relationship data","year":2008,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Data mining; Cluster analysis; Heuristic; Identification (biology); Tree (set theory); Machine learning; Mathematics; Artificial intelligence","authors":[{"name":"Rong Ge","is_ca":true},{"name":"Martin Ester","is_ca":true},{"name":"Byron J. Gao","is_ca":true},{"name":"Zengjian Hu","is_ca":true},{"name":"Binay Bhattacharya","is_ca":true},{"name":"Boaz Ben‐Moshe","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1540393998356944,"gpt":0.3374659296680593,"spread":0.1834265298323649,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00516296,0.001162493,0.002037524,0.006583357,0.001886862,0.003518092,0.003214465,0.001726726,0.002199593],"category_scores_gemma":[0.0227216,0.0007094794,0.00208849,0.009076138,0.001448626,0.004764669,0.003139258,0.002371179,0.0008813201],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002006607,"about_ca_system_score_gemma":0.002068265,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005120876,"about_ca_topic_score_gemma":0.004685152,"domain_scores_codex":[0.9936633,0.001942538,0.0003271466,0.001550275,0.002011249,0.0005054963],"domain_scores_gemma":[0.984099,0.008251806,0.001208199,0.003380702,0.00252976,0.0005305047],"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":[0.001047525,0.0006635603,0.02884979,0.0005928837,0.0006617541,0.0005288633,0.001025869,0.5278539,0.006805596,0.1317543,0.01243629,0.2877797],"study_design_scores_gemma":[0.00001945373,0.00004276196,0.002221785,0.00001821956,0.00005448686,0.000131772,0.0002464179,0.9074759,0.002907759,0.08348496,0.003364379,0.0000321173],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04467297,0.0002616396,0.9510509,0.0003806653,0.00003355224,0.0001509502,0.001242048,0.0007017161,0.001505553],"genre_scores_gemma":[0.4644668,0.0003810547,0.5261062,0.0001314788,0.0001014967,0.0004402197,0.005582307,0.0002861098,0.002504308],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006583357,"threshold_uncertainty_score":0.02730465,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2724995883","doi":"10.1145/3080574","title":"Mining Community Structures in Multidimensional Networks","year":2017,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":41,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Relevance (law); Identification (biology); Set (abstract data type); Data mining; Node (physics); Function (biology); Machine learning; Feature (linguistics); Artificial intelligence; Data science","authors":[{"name":"Oualid Boutemine","is_ca":true},{"name":"Mohamed Bouguessa","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06927095770474913,"gpt":0.3432355618617533,"spread":0.2739646041570042,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002474856,0.0009803575,0.001079037,0.007375643,0.001663395,0.001937046,0.001807813,0.002006049,0.0007324074],"category_scores_gemma":[0.01504709,0.0006916371,0.001106311,0.005463107,0.001436113,0.003842318,0.002740303,0.001522614,0.0003428553],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001006001,"about_ca_system_score_gemma":0.0006864919,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002171645,"about_ca_topic_score_gemma":0.002515809,"domain_scores_codex":[0.9969614,0.001152954,0.0001613263,0.0008433151,0.0006877701,0.0001932527],"domain_scores_gemma":[0.9880393,0.007468848,0.002224336,0.0009363902,0.0009880707,0.0003430294],"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":[0.0002030073,0.0003752656,0.0496773,0.00128909,0.0005847695,0.001148663,0.001894754,0.5474613,0.01442554,0.1086907,0.00666435,0.2675851],"study_design_scores_gemma":[0.0000156501,0.00004151317,0.003936504,0.00007491137,0.00004964223,0.000275128,0.0003433081,0.8966676,0.002534346,0.09237883,0.00364706,0.0000355233],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06779605,0.001411265,0.9278592,0.000583895,0.00003929428,0.0001414576,0.0005157563,0.0002669963,0.001386076],"genre_scores_gemma":[0.5175289,0.001130371,0.4783867,0.0001737597,0.0001476992,0.0003062076,0.001244423,0.00006502021,0.001017051],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007375643,"threshold_uncertainty_score":0.01308846,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2405494619","doi":"10.1145/2983533","title":"Structural Analysis of User Choices for Mobile App Recommendation","year":2016,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":38,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Fundamental Research Funds for the Central Universities; National Natural Science Foundation of China","keywords":"Computer science; Mobile apps; App store; Variety (cybernetics); Recommender system; World Wide Web; Margin (machine learning); Focus (optics); Taxonomy (biology); Data science; Human–computer interaction; Artificial intelligence; Machine learning","authors":[{"name":"Bin Liu","is_ca":false},{"name":"Yao Wu","is_ca":true},{"name":"Neil Zhenqiang Gong","is_ca":false},{"name":"Junjie Wu","is_ca":false},{"name":"Hui Xiong","is_ca":false},{"name":"Martin Ester","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05976271325554005,"gpt":0.3313509147680883,"spread":0.2715882015125483,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004507178,0.0008160123,0.001312896,0.002259586,0.000801587,0.000975322,0.00177839,0.001598286,0.005125963],"category_scores_gemma":[0.02138644,0.0008551861,0.002093413,0.002427602,0.001008918,0.002695148,0.001039954,0.001993058,0.0009657572],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00199139,"about_ca_system_score_gemma":0.001346326,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01650641,"about_ca_topic_score_gemma":0.03453283,"domain_scores_codex":[0.9972163,0.001549976,0.0001303551,0.0005795545,0.0003355131,0.0001882523],"domain_scores_gemma":[0.9807716,0.01585925,0.001023721,0.001203682,0.0007573565,0.0003843436],"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":[0.001035654,0.0009398473,0.1917157,0.0007723205,0.001053837,0.0006155616,0.001822358,0.4648748,0.003829166,0.1345475,0.009059825,0.1897334],"study_design_scores_gemma":[0.00001557496,0.00008021106,0.007134695,0.0000231939,0.00003452053,0.00006717318,0.00007924354,0.9630389,0.0001982862,0.02845795,0.0008494901,0.00002075215],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3808426,0.001144491,0.6096746,0.001986672,0.00006141481,0.0004205467,0.002328772,0.0004809297,0.003059961],"genre_scores_gemma":[0.9377237,0.0004162717,0.05695714,0.0001588191,0.00005129493,0.0002728906,0.001768512,0.00003891182,0.002612406],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01650641,"threshold_uncertainty_score":0.03282064,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2020791430","doi":"10.1145/1552303.1552306","title":"Link spam target detection using page farms","year":2009,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"PageRank; Computer science; Backlink; Web page; Spamdexing; HITS algorithm; Information retrieval; Ranking (information retrieval); Link analysis; Search engine; Precision and recall; Data mining; World Wide Web; Web search engine; Static web page; Web search query; Web development","authors":[{"name":"Bin Zhou","is_ca":true},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05779916987046218,"gpt":0.2945533312560526,"spread":0.2367541613855904,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001416917,0.000686666,0.0009854912,0.009510489,0.0009221472,0.00151927,0.0006686596,0.001224445,0.001261436],"category_scores_gemma":[0.008350234,0.0003563109,0.0005219337,0.002613494,0.0007458146,0.003632901,0.0009666558,0.0005194026,0.001283279],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004196323,"about_ca_system_score_gemma":0.0005436183,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00104984,"about_ca_topic_score_gemma":0.001656089,"domain_scores_codex":[0.9986269,0.0003350224,0.00009824903,0.000268068,0.0005510918,0.0001207255],"domain_scores_gemma":[0.9910792,0.003562806,0.001676727,0.001302262,0.002093913,0.0002851163],"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":[0.0007850026,0.0006806268,0.2218591,0.0005149396,0.0003010226,0.0009694705,0.0006764358,0.03251593,0.04790626,0.008431553,0.02171392,0.6636458],"study_design_scores_gemma":[0.00005904761,0.000388619,0.06485593,0.00007257026,0.0001477872,0.002329091,0.0003912146,0.8528716,0.05240298,0.01615833,0.0102053,0.0001174349],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4980715,0.002115235,0.4785665,0.0004431481,0.0001569315,0.0005939534,0.001688786,0.01116487,0.007199109],"genre_scores_gemma":[0.8544123,0.0005341041,0.138685,0.00009952424,0.0002197949,0.0001889589,0.002329884,0.0001478459,0.003382475],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009510489,"threshold_uncertainty_score":0.007493496,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3039137796","doi":"10.1145/3394053","title":"Internal Evaluation of Unsupervised Outlier Detection","year":2020,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Outlier; Anomaly detection; Computer science; Cluster analysis; Data mining; Artificial intelligence; Domain (mathematical analysis); Pattern recognition (psychology); Binary number; Mathematics","authors":[{"name":"Henrique O. Marques","is_ca":false},{"name":"Ricardo J. G. B. Campello","is_ca":false},{"name":"Jörg Sander","is_ca":true},{"name":"Arthur Zimek","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1100711224031969,"gpt":0.3275591643988292,"spread":0.2174880419956323,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02575282,0.002216955,0.002269383,0.00693881,0.001342251,0.004383758,0.00228229,0.002177985,0.001320314],"category_scores_gemma":[0.09601629,0.0003817446,0.001436239,0.004000065,0.001967856,0.003794039,0.003600796,0.001694348,0.0008051068],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00151691,"about_ca_system_score_gemma":0.002161565,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001395493,"about_ca_topic_score_gemma":0.001960982,"domain_scores_codex":[0.9721696,0.008948387,0.003028749,0.003417876,0.01151485,0.000920582],"domain_scores_gemma":[0.9299034,0.03028598,0.009880595,0.008466583,0.01983213,0.001631266],"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":[0.001810462,0.0009420441,0.1164282,0.001585192,0.001530322,0.0002738165,0.001127759,0.2266285,0.02338895,0.02658168,0.01352165,0.5861814],"study_design_scores_gemma":[0.0001174579,0.001581934,0.03122169,0.0002946969,0.0002887738,0.0004777281,0.0009715053,0.8807139,0.05042863,0.02639274,0.007284181,0.0002267736],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2234891,0.001834084,0.7572403,0.0006020296,0.0004099645,0.0005633871,0.001746069,0.00428494,0.009830108],"genre_scores_gemma":[0.7263756,0.0003531367,0.2666107,0.0002104416,0.0001544421,0.0003283015,0.003687615,0.000604468,0.001675188],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02575282,"threshold_uncertainty_score":0.1361956,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2214080290","doi":"10.1145/2700409","title":"Classification with Streaming Features: An Emerging-Pattern Mining Approach","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Streaming data; Feature selection; Data mining; Benchmark (surveying); Classifier (UML); Curse of dimensionality; Feature (linguistics); Machine learning; Artificial intelligence; Feature vector; Pattern recognition (psychology)","authors":[{"name":"Kui Yu","is_ca":true},{"name":"Wei Ding","is_ca":false},{"name":"Dan A. Simovici","is_ca":false},{"name":"Hao Wang","is_ca":false},{"name":"Jian Pei","is_ca":true},{"name":"Xindong Wu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1110131822858443,"gpt":0.3239055552908267,"spread":0.2128923730049824,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00176813,0.0009650734,0.001521119,0.002994987,0.0005563234,0.00131399,0.002436844,0.0009118359,0.0008835172],"category_scores_gemma":[0.006130014,0.0003556556,0.001008936,0.003925871,0.0005004139,0.003561039,0.001093864,0.001398631,0.0006812078],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004834595,"about_ca_system_score_gemma":0.0006526641,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0016669,"about_ca_topic_score_gemma":0.001982375,"domain_scores_codex":[0.9987291,0.0002076957,0.0001490076,0.0004312988,0.0003893489,0.00009358759],"domain_scores_gemma":[0.9963097,0.001218486,0.0005455472,0.0007189023,0.001058568,0.0001488493],"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":[0.0003592313,0.0004661958,0.02531894,0.000277821,0.0002482418,0.000600466,0.0002823576,0.08947644,0.009867474,0.008495264,0.00989962,0.854708],"study_design_scores_gemma":[0.00001892565,0.000105178,0.00229843,0.00002155433,0.0000413139,0.0002590644,0.00008384064,0.978508,0.002448478,0.01287021,0.003322941,0.00002211739],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.036985,0.0007916342,0.9590371,0.0003847621,0.0001107105,0.0001713821,0.0006274522,0.001108763,0.0007832348],"genre_scores_gemma":[0.4880339,0.000791578,0.5062186,0.0001977513,0.000315269,0.0002912916,0.00220712,0.0001248509,0.00181983],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.002994987,"threshold_uncertainty_score":0.009350836,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2055028639","doi":"10.1145/1460797.1460801","title":"On domination game analysis for microeconomic data mining","year":2009,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":31,"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":"National University of Singapore","keywords":"Nash equilibrium; Game theory; Dimension (graph theory); Product (mathematics); Commodity; Computer science; Point (geometry); Best response; Space (punctuation); Mathematical economics; Economics; Mathematics","authors":[{"name":"Zhenjie Zhang","is_ca":false},{"name":"Laks V. S. Lakshmanan","is_ca":true},{"name":"Anthony K. H. Tung","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07051834428210217,"gpt":0.3199586946775136,"spread":0.2494403503954114,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008198864,0.002995659,0.003391332,0.00417315,0.001543905,0.003303134,0.002994104,0.001951651,0.00522264],"category_scores_gemma":[0.02809138,0.001211644,0.003032939,0.005536519,0.004365378,0.0049341,0.003647832,0.003690699,0.0007182587],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00431597,"about_ca_system_score_gemma":0.003318073,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01083767,"about_ca_topic_score_gemma":0.006749301,"domain_scores_codex":[0.9949528,0.003103144,0.000234184,0.0005499937,0.0007952742,0.0003646441],"domain_scores_gemma":[0.9705976,0.02643483,0.0008856081,0.000780491,0.0008934896,0.0004080156],"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":[0.00006142603,0.0001123886,0.001341758,0.0002185113,0.000172056,0.0001608364,0.0001874658,0.6798005,0.0005257613,0.2949516,0.001799232,0.02066856],"study_design_scores_gemma":[0.00001871201,0.00001510531,0.00008055411,0.00001685578,0.00001282782,0.00002060951,0.00001673577,0.8109585,0.00009537335,0.1876908,0.001061216,0.00001279906],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005846548,0.0007105552,0.9889575,0.0005452654,0.00005830654,0.0001789973,0.0001432661,0.00009317749,0.003466409],"genre_scores_gemma":[0.380694,0.002950071,0.6056885,0.0008449686,0.0004467708,0.001764754,0.0006342878,0.0002273936,0.006749227],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01083767,"threshold_uncertainty_score":0.04336023,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3212326024","doi":"10.1145/3487046","title":"Scalable Mining of High-Utility Sequential Patterns With Three-Tier MapReduce Model","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Brandon University","funders":"Narodowe Centrum Badań i Rozwoju; Høgskulen på Vestlandet","keywords":"Computer science; Scalability; Correctness; Data mining; Completeness (order theory); Computation; Scale (ratio); Database; Algorithm","authors":[{"name":"Jerry Chun‐Wei Lin","is_ca":false},{"name":"Youcef Djenouri","is_ca":false},{"name":"Gautam Srivastava","is_ca":true},{"name":"Yuanfa Li","is_ca":false},{"name":"Philip S. Yu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05944832735846505,"gpt":0.2862248184868035,"spread":0.2267764911283385,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001134107,0.0006174631,0.0009325662,0.0007874378,0.0007444914,0.00122444,0.002463921,0.0004940712,0.00113602],"category_scores_gemma":[0.00214518,0.000448815,0.001413014,0.001452695,0.0004829036,0.002052542,0.001265268,0.0006497084,0.0003359456],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007087222,"about_ca_system_score_gemma":0.002430061,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008288474,"about_ca_topic_score_gemma":0.01088543,"domain_scores_codex":[0.9989473,0.0001765561,0.00007841155,0.000285604,0.0003446543,0.0001675414],"domain_scores_gemma":[0.9991201,0.0002224569,0.00006500541,0.0002340859,0.0002671217,0.00009122945],"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":[0.0005822737,0.0004269541,0.01107233,0.0004835542,0.0003617129,0.0009621267,0.0005152492,0.7235528,0.01425434,0.05944934,0.01137528,0.176964],"study_design_scores_gemma":[0.00001571453,0.00003363323,0.0003596914,0.000003205113,0.0000207153,0.0001029372,0.00004794797,0.9810283,0.001166464,0.01609148,0.001121061,0.000008944599],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04069964,0.0002629322,0.954759,0.0003608438,0.00004706087,0.0001691549,0.0004628312,0.001503985,0.001734654],"genre_scores_gemma":[0.6603029,0.0004857481,0.3342477,0.0001590569,0.00006949661,0.0003183775,0.001139479,0.0001226553,0.003154552],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008288474,"threshold_uncertainty_score":0.01648045,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4205969642","doi":"10.1145/3495011","title":"sGrapp: Butterfly Approximation in Streaming Graphs","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":23,"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; Bipartite graph; Theoretical computer science; Streaming algorithm; Graph; Mathematics; Upper and lower bounds","authors":[{"name":"Aida Sheshbolouki","is_ca":true},{"name":"M. TAMER ÖZSU","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03476832363944742,"gpt":0.2913620502884173,"spread":0.2565937266489698,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001150138,0.001096666,0.001182132,0.001648893,0.0008120152,0.001559667,0.002893224,0.001221448,0.003907339],"category_scores_gemma":[0.00900408,0.0005149547,0.0007662442,0.002197029,0.0007535898,0.00351893,0.001827725,0.00169215,0.001236444],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001331246,"about_ca_system_score_gemma":0.002249854,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01048898,"about_ca_topic_score_gemma":0.01342499,"domain_scores_codex":[0.9992397,0.0001108999,0.00004479069,0.0002510375,0.0002318436,0.000121757],"domain_scores_gemma":[0.9970964,0.001470005,0.0002607719,0.000677752,0.0003178523,0.0001771645],"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":[0.000641951,0.0002435628,0.005858849,0.0003997457,0.0001537249,0.0002497283,0.0003869247,0.5152089,0.01097521,0.05078537,0.02584288,0.3892531],"study_design_scores_gemma":[0.0000235341,0.00003133776,0.0002449355,0.000008209091,0.000008652174,0.00005206506,0.00003741479,0.9801456,0.001330857,0.01659227,0.001519117,0.000005933778],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05321199,0.0005491522,0.9364994,0.0005823655,0.0001292896,0.000156027,0.0007024371,0.00570373,0.002465544],"genre_scores_gemma":[0.3043662,0.0004062916,0.6886351,0.0003029993,0.0001300852,0.0002568729,0.002013742,0.0006824942,0.003206308],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01048898,"threshold_uncertainty_score":0.02085584,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2786524829","doi":"10.1145/3047010","title":"ATR-Vis","year":2018,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Video Analysis and Summarization","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Dalhousie University","funders":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação de Amparo à Pesquisa do Estado de São Paulo; Natural Sciences and Engineering Research Council of Canada; International Development Research Centre","keywords":"Computer science; Variety (cybernetics); Terminology; Task (project management); Information retrieval; Set (abstract data type); Visualization; Data science; World Wide Web; Data mining; Artificial intelligence","authors":[{"name":"Raheleh Makki","is_ca":true},{"name":"Eder José de Carvalho","is_ca":false},{"name":"Axel J. Soto","is_ca":true},{"name":"Stephen Brooks","is_ca":true},{"name":"Maria Cristina Ferreira de Oliveira","is_ca":false},{"name":"Evangelos Milios","is_ca":true},{"name":"Rosane Minghim","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04897813893720501,"gpt":0.2997655827834099,"spread":0.2507874438462049,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00159393,0.001548579,0.0006746308,0.002560374,0.000855409,0.00348702,0.001553962,0.001591172,0.06988068],"category_scores_gemma":[0.00647447,0.0006004754,0.001204812,0.002037151,0.0004106753,0.003505655,0.003086129,0.001379713,0.07513471],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007120278,"about_ca_system_score_gemma":0.001074368,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00560389,"about_ca_topic_score_gemma":0.009078515,"domain_scores_codex":[0.9986048,0.000285358,0.00009289219,0.0003611349,0.0005100364,0.000145837],"domain_scores_gemma":[0.9975546,0.0005458175,0.0001695173,0.0008798582,0.0006339551,0.0002163795],"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":[0.001065383,0.0002135992,0.003542906,0.001479408,0.0001148779,0.0002624067,0.001315887,0.002880107,0.01833561,0.01207516,0.6799309,0.2787838],"study_design_scores_gemma":[0.0001660401,0.0002174758,0.004158979,0.0002174874,0.00004364649,0.0004758156,0.0006853417,0.02543051,0.01275875,0.01501215,0.9407105,0.0001233073],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"software","genre_gemma":"other","genre_scores_codex":[0.02410632,0.003118783,0.1837181,0.002259042,0.001305529,0.001511288,0.2123531,0.4092123,0.1624155],"genre_scores_gemma":[0.1266793,0.002023633,0.2846733,0.001799798,0.0005843976,0.002472554,0.4176916,0.03861568,0.1254597],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.9301193,"threshold_uncertainty_score":0,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2895492593","doi":"10.1145/3237191","title":"Enumerating Trillion Subgraphs On Distributed Systems","year":2018,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Kootenay Association for Science & Technology","funders":"Chongqing Research Program of Basic Research and Frontier Technology","keywords":"Enumeration; Computer science; Scalability; Theoretical computer science; Graph; Algorithm; Combinatorics; Mathematics; Database","authors":[{"name":"Ha-Myung Park","is_ca":true},{"name":"Francesco Silvestri","is_ca":false},{"name":"Rasmus Pagh","is_ca":false},{"name":"Chin‐Wan Chung","is_ca":false},{"name":"Sung-Hyon Myaeng","is_ca":true},{"name":"U Kang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04566446021945211,"gpt":0.2890463675566758,"spread":0.2433819073372237,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006562368,0.0006812023,0.0008503117,0.001318931,0.001009317,0.001191674,0.001761834,0.0006518395,0.002588041],"category_scores_gemma":[0.005064799,0.0004441322,0.0007236362,0.001999208,0.0008612851,0.004051041,0.00202189,0.0008265829,0.000791176],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001003861,"about_ca_system_score_gemma":0.001088495,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00465644,"about_ca_topic_score_gemma":0.00631414,"domain_scores_codex":[0.9987999,0.0002481948,0.00008482573,0.0003436655,0.0003242422,0.0001992725],"domain_scores_gemma":[0.9967264,0.001294233,0.0003004962,0.0009715089,0.0005003418,0.0002070033],"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":[0.0005595796,0.0002502761,0.008487561,0.0006907436,0.0001601752,0.0005616485,0.001170765,0.5629785,0.02795366,0.06943382,0.01769184,0.3100615],"study_design_scores_gemma":[0.00005327486,0.00006901436,0.0007964118,0.00001985569,0.00002682904,0.0001793086,0.0003105316,0.8937838,0.007976204,0.09074271,0.006024022,0.00001806494],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1599276,0.0005850857,0.8274297,0.0008560457,0.0000841735,0.0002364841,0.0009616364,0.00542403,0.004495275],"genre_scores_gemma":[0.5527877,0.0004165529,0.4387371,0.0002528307,0.00005218238,0.00032071,0.002627997,0.0004465739,0.004358317],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00465644,"threshold_uncertainty_score":0.009258687,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2504431993","doi":"10.1145/2948068","title":"The Convergence Behavior of Naive Bayes on Large Sparse Datasets","year":2016,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Naive Bayes classifier; Computer science; Machine learning; Missing data; Bayes' theorem; Artificial intelligence; Convergence (economics); Data mining; Bayesian probability; Support vector machine","authors":[{"name":"Xiang Li","is_ca":true},{"name":"Charles X. Ling","is_ca":true},{"name":"Huaimin Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05267241250637728,"gpt":0.3180910509676954,"spread":0.2654186384613181,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03662071,0.001379853,0.002446742,0.003578983,0.00194114,0.002782564,0.002111165,0.00234012,0.001748873],"category_scores_gemma":[0.1798772,0.0008991663,0.001350531,0.002164804,0.003616876,0.006263993,0.00252599,0.003621093,0.0008480089],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00243485,"about_ca_system_score_gemma":0.001975697,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004712497,"about_ca_topic_score_gemma":0.003403174,"domain_scores_codex":[0.9835578,0.009208647,0.0007935746,0.002027664,0.00361214,0.0008002378],"domain_scores_gemma":[0.8319974,0.1399123,0.004686417,0.009698394,0.01208273,0.001622687],"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":[0.00163393,0.0004143298,0.03283789,0.00106157,0.0005448213,0.0005932128,0.001679861,0.5049282,0.004500951,0.171064,0.01414255,0.2665986],"study_design_scores_gemma":[0.00006888292,0.0001527159,0.001955651,0.0001453485,0.00003623563,0.000284537,0.0001762059,0.8470767,0.001434693,0.1472828,0.001341053,0.00004518577],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09432787,0.002971777,0.8919097,0.003434048,0.0001921118,0.000345948,0.0004199694,0.001186371,0.005212131],"genre_scores_gemma":[0.6786148,0.001512285,0.3140092,0.001321331,0.0003244362,0.0005516678,0.001200377,0.0003377174,0.002128234],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.03662071,"threshold_uncertainty_score":0.1936711,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1993150471","doi":"10.1145/1870096.1870099","title":"Discovering Knowledge-Sharing Communities in Question-Answering Forums","year":2010,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Expert finding and Q&A systems","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec en Outaouais","funders":"","keywords":"Computer science; Cluster analysis; Transaction data; Question answering; Set (abstract data type); Information retrieval; Identification (biology); Focus (optics); Function (biology); Database transaction; Data mining; Artificial intelligence; Database","authors":[{"name":"Mohamed Bouguessa","is_ca":true},{"name":"Shengrui Wang","is_ca":true},{"name":"Benoît Dumoulin","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05505658305164102,"gpt":0.3200791919554526,"spread":0.2650226089038116,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01091546,0.0009805014,0.001300923,0.008974935,0.002487013,0.003028107,0.002362711,0.002584624,0.001291093],"category_scores_gemma":[0.04403192,0.0007049059,0.001410804,0.00503656,0.001820465,0.00865994,0.003765093,0.001341411,0.0006060864],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001190373,"about_ca_system_score_gemma":0.000998556,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003167849,"about_ca_topic_score_gemma":0.003875063,"domain_scores_codex":[0.9891086,0.005772627,0.0005099854,0.002354976,0.001666512,0.0005872613],"domain_scores_gemma":[0.9492251,0.03596724,0.004773722,0.004502148,0.003878017,0.001653831],"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":[0.003048549,0.002647969,0.3429502,0.002155127,0.0007812362,0.004212623,0.04684595,0.09399186,0.04387166,0.0557952,0.008198698,0.3955009],"study_design_scores_gemma":[0.000178044,0.0004802692,0.07268889,0.0002417076,0.0002176017,0.001664163,0.01488411,0.7654831,0.01636647,0.1116125,0.01598546,0.0001977168],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6938417,0.001190601,0.2983999,0.0007057365,0.00004767989,0.0006892813,0.001654585,0.0008700353,0.002600487],"genre_scores_gemma":[0.8668677,0.0001573891,0.1289277,0.00007456115,0.00005760634,0.0002840301,0.002724605,0.00006526893,0.0008412034],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01091546,"threshold_uncertainty_score":0.05772716,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4386497976","doi":"10.1145/3622940","title":"Discovering Interesting Patterns from Hypergraphs","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":14,"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":"Natural Sciences and Engineering Research Council of Canada; University of Manitoba","keywords":"Hypergraph; Data mining; Computer science; Theoretical computer science; Mathematics; Combinatorics","authors":[{"name":"Md. Tanvir Alam","is_ca":false},{"name":"Chowdhury Farhan Ahmed","is_ca":false},{"name":"Md. Samiullah","is_ca":false},{"name":"Carson K. Leung","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07976721731493004,"gpt":0.3152483108612908,"spread":0.2354810935463608,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001025094,0.0007097881,0.0006533889,0.00623309,0.0009177239,0.002275039,0.001038389,0.0008503133,0.001454786],"category_scores_gemma":[0.007772056,0.00060046,0.001198922,0.005455595,0.0008695932,0.003730134,0.001616578,0.001109795,0.0004656785],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007589065,"about_ca_system_score_gemma":0.001293607,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002253375,"about_ca_topic_score_gemma":0.004695178,"domain_scores_codex":[0.9981,0.0004964424,0.000148362,0.0005737024,0.0005540704,0.0001274482],"domain_scores_gemma":[0.9946021,0.003154426,0.0005899136,0.0007431779,0.0007184056,0.0001920978],"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":[0.0003478696,0.0003431453,0.03773529,0.001200729,0.0006253048,0.003206151,0.001751629,0.09249736,0.02259602,0.1665487,0.0154213,0.6577266],"study_design_scores_gemma":[0.00004659355,0.00008376808,0.005189066,0.0001604437,0.0002102283,0.001713714,0.0009051104,0.5177238,0.008780741,0.4417001,0.02342821,0.00005820341],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06337195,0.001016367,0.9281379,0.0008289096,0.00004022895,0.0002887736,0.001864577,0.001312645,0.00313857],"genre_scores_gemma":[0.3182794,0.001339423,0.6726004,0.0003100804,0.00008197915,0.0003944276,0.004742163,0.0001557847,0.002096257],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00623309,"threshold_uncertainty_score":0.005506217,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1993625478","doi":"10.1145/1409620.1409623","title":"On disclosure risk analysis of anonymized itemsets in the presence of prior knowledge","year":2008,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Hacker; Computer science; Heuristics; Context (archaeology); Set (abstract data type); Function (biology); Data mining; Benchmark (surveying); Information retrieval; Computer security","authors":[{"name":"Laks V. S. Lakshmanan","is_ca":true},{"name":"Raymond T. Ng","is_ca":true},{"name":"Ganesh Ramesh","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05564791742840004,"gpt":0.3095288978769892,"spread":0.2538809804485891,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02787268,0.001962691,0.003217729,0.004885206,0.001532687,0.004289513,0.003065348,0.003676991,0.002107014],"category_scores_gemma":[0.1670895,0.001152028,0.002363153,0.004193527,0.004377269,0.009897526,0.003753643,0.004901734,0.0003130187],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003128707,"about_ca_system_score_gemma":0.001915636,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003009452,"about_ca_topic_score_gemma":0.001532457,"domain_scores_codex":[0.9860148,0.007254244,0.001032981,0.001990435,0.002880194,0.0008272452],"domain_scores_gemma":[0.6065529,0.3590526,0.01692611,0.01031936,0.005523658,0.001625378],"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":[0.0004142344,0.0001364718,0.01936088,0.0004256357,0.0002821795,0.0008421764,0.0007393445,0.8536233,0.0008570471,0.08845264,0.001706144,0.03316],"study_design_scores_gemma":[0.00001296409,0.00007621016,0.001511193,0.0001060203,0.00004824139,0.0003529197,0.0001385493,0.9268185,0.0005380995,0.069902,0.0004549331,0.00004031576],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1545204,0.003942685,0.8345001,0.003075713,0.00009231806,0.0002654153,0.0008333632,0.0003900247,0.002380042],"genre_scores_gemma":[0.9049052,0.00279791,0.08811141,0.0004358682,0.0003844567,0.0003457514,0.001201893,0.0001067852,0.001710636],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02787268,"threshold_uncertainty_score":0.1474066,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4388421529","doi":"10.1145/3631609","title":"Laplacian Change Point Detection for Single and Multi-view Dynamic Graphs","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Mila - Quebec Artificial Intelligence Institute; Canadian Institute for Advanced Research; McGill University","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Computer science; Anomaly detection; Laplacian matrix; Snapshot (computer storage); Laplace operator; Embedding; Theoretical computer science; Graph; Algorithm; Data mining; Artificial intelligence; Mathematics","authors":[{"name":"Shenyang Huang","is_ca":true},{"name":"Samy Coulombe","is_ca":true},{"name":"Yasmeen Hitti","is_ca":true},{"name":"Reihaneh Rabbany","is_ca":true},{"name":"Guillaume Rabusseau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09321852727549758,"gpt":0.335970799174395,"spread":0.2427522718988974,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000818288,0.0009655742,0.00091316,0.00294637,0.0005664857,0.001153283,0.001085563,0.0009044359,0.001084674],"category_scores_gemma":[0.005999636,0.0003351344,0.0007900964,0.001869783,0.000702141,0.001941992,0.001176037,0.001363375,0.0005973061],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008444819,"about_ca_system_score_gemma":0.0004574746,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006202641,"about_ca_topic_score_gemma":0.008459907,"domain_scores_codex":[0.9992094,0.0001261431,0.00003139327,0.0002978352,0.000237987,0.00009720829],"domain_scores_gemma":[0.997911,0.0008959554,0.0002738768,0.0003451315,0.0004184591,0.0001555561],"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":[0.0004789415,0.0002408896,0.02462607,0.0002770113,0.000268332,0.0009825889,0.0005321047,0.3702402,0.03117643,0.01847031,0.0154628,0.5372444],"study_design_scores_gemma":[0.000006271336,0.00002840889,0.002576447,0.00000647777,0.00001339464,0.000195365,0.0001052134,0.983192,0.002751525,0.009818552,0.001286707,0.00001963296],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1255008,0.0007737945,0.8680271,0.0004456663,0.0001354472,0.00006566054,0.0006518402,0.002446767,0.00195294],"genre_scores_gemma":[0.844093,0.0004435389,0.1515268,0.0001985334,0.00009895614,0.0000542551,0.001562702,0.0002175561,0.001804685],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006202641,"threshold_uncertainty_score":0.0123331,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3163912335","doi":"10.1145/3448021","title":"Similarity Embedding Networks for Robust Human Activity Recognition","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Context-Aware Activity Recognition Systems","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 Alberta","funders":"","keywords":"Artificial intelligence; Computer science; Embedding; Pattern recognition (psychology); Similarity (geometry); Pairwise comparison; Activity recognition; Cluster analysis; Convolutional neural network; Deep learning; Machine learning; Generalization; Data mining; Mathematics","authors":[{"name":"Chenglin Li","is_ca":true},{"name":"Carrie Lu Tong","is_ca":true},{"name":"Di Niu","is_ca":true},{"name":"Bei Jiang","is_ca":true},{"name":"Xiao Ming Zuo","is_ca":false},{"name":"Lei Cheng","is_ca":false},{"name":"Jian Xiong","is_ca":false},{"name":"Jianming Yang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1595380997247741,"gpt":0.343239936842662,"spread":0.1837018371178879,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009492827,0.0009024217,0.0009630568,0.0008023462,0.0002109474,0.0007712764,0.001225635,0.0008800828,0.001791654],"category_scores_gemma":[0.004010358,0.0004057517,0.0006170257,0.001015916,0.0006643957,0.002010965,0.001427469,0.001584969,0.0008189767],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007221204,"about_ca_system_score_gemma":0.000474578,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002245747,"about_ca_topic_score_gemma":0.002685249,"domain_scores_codex":[0.9991352,0.0002266792,0.00005789469,0.0003254824,0.0001817008,0.00007298773],"domain_scores_gemma":[0.9991407,0.0003126073,0.0001573647,0.000188999,0.0001595331,0.00004084373],"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":[0.0002435842,0.0001774074,0.001635229,0.0001490973,0.0001197323,0.00008309972,0.0001033502,0.422533,0.01033064,0.01506634,0.004227622,0.5453308],"study_design_scores_gemma":[0.000004088632,0.0000307113,0.0003367043,0.000006128843,0.00000587696,0.00002323574,0.000009236749,0.988605,0.00201917,0.008323506,0.0006284455,0.000007926209],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01856285,0.0006486047,0.9782406,0.000163357,0.00005656257,0.00003702268,0.0001871744,0.001140155,0.0009637485],"genre_scores_gemma":[0.8094994,0.0008149525,0.1828446,0.0002396224,0.0001131586,0.0001692256,0.001364513,0.0002002118,0.004754179],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002245747,"threshold_uncertainty_score":0.005993664,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3011191765","doi":"10.1145/3380928","title":"Continuous Influence Maximization","year":2020,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Maximization; Computer science; Benchmark (surveying); Mathematical optimization; Product (mathematics); Function (biology); Mathematics","authors":[{"name":"Yang Yu","is_ca":false},{"name":"Xiangbo Mao","is_ca":true},{"name":"Jian Pei","is_ca":true},{"name":"Xiaofei He","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04107850160594025,"gpt":0.291683042845804,"spread":0.2506045412398638,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002586059,0.001463045,0.001651378,0.0009892838,0.0006375557,0.001856213,0.001849225,0.001391133,0.005189287],"category_scores_gemma":[0.01171755,0.0006758786,0.001264889,0.001411959,0.001684666,0.002274976,0.001441041,0.001601709,0.001067386],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001661474,"about_ca_system_score_gemma":0.001273557,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003750336,"about_ca_topic_score_gemma":0.003299517,"domain_scores_codex":[0.9980754,0.000865783,0.00007791579,0.0004836986,0.000347292,0.0001498844],"domain_scores_gemma":[0.9943722,0.004555423,0.0002474408,0.0003298759,0.0003433122,0.0001517984],"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":[0.0001100574,0.00007610811,0.001531701,0.0003685885,0.0001283237,0.0001968679,0.0002162615,0.6878033,0.001981445,0.2226782,0.005266139,0.07964313],"study_design_scores_gemma":[0.0000154815,0.00002730936,0.0002193479,0.00001721973,0.00002033425,0.00005392382,0.00001838657,0.9270104,0.0007059812,0.06926904,0.002631487,0.00001116521],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007721723,0.0004294049,0.9849676,0.0002918538,0.00003940901,0.00007586874,0.000147845,0.0001653184,0.00616089],"genre_scores_gemma":[0.6202383,0.001579429,0.3647035,0.0003776563,0.0003317239,0.0005556305,0.000688521,0.0002786536,0.01124651],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005189287,"threshold_uncertainty_score":0.01735991,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3209720970","doi":"10.1145/3473910","title":"Semi-Supervised Ensemble Learning for Dealing with Inaccurate and Incomplete Supervision","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"IBM (Canada); University of Alberta","funders":"","keywords":"Computer science; Machine learning; Artificial intelligence; Classifier (UML); Ensemble learning; Supervised learning; Noise (video); Semi-supervised learning; Labeled data; Data mining; Artificial neural network","authors":[{"name":"Mona Nashaat","is_ca":true},{"name":"Aindrila Ghosh","is_ca":true},{"name":"James A. Miller","is_ca":true},{"name":"Shaikh Quader","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05977660722155591,"gpt":0.3027910564366063,"spread":0.2430144492150504,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006365681,0.00152307,0.00255694,0.001902694,0.001019347,0.001205699,0.00304222,0.001869999,0.0008660409],"category_scores_gemma":[0.01210102,0.0006573084,0.001236245,0.00159379,0.001086938,0.002953575,0.002196195,0.002915104,0.0006696826],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006138321,"about_ca_system_score_gemma":0.001248289,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001828173,"about_ca_topic_score_gemma":0.003292588,"domain_scores_codex":[0.996216,0.001434374,0.0002134348,0.0009356693,0.001020147,0.0001804414],"domain_scores_gemma":[0.991839,0.003455444,0.000776606,0.002154808,0.00156041,0.0002137802],"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":[0.0003321688,0.0003417772,0.006472449,0.0002892499,0.0003826716,0.0002280589,0.0005591214,0.389867,0.009666024,0.01126886,0.005197058,0.5753955],"study_design_scores_gemma":[0.000006389278,0.00004588888,0.00036829,0.00001414725,0.00002220925,0.00004719408,0.00002563344,0.990003,0.002059148,0.006590726,0.0008043593,0.00001299399],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01296276,0.0003550395,0.9853951,0.00009248387,0.00003100363,0.00004045676,0.00005083894,0.0006173758,0.0004549841],"genre_scores_gemma":[0.4929367,0.0004638242,0.5031748,0.0002189276,0.0002352911,0.0002251711,0.0009319268,0.0001647708,0.00164873],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006365681,"threshold_uncertainty_score":0.03366536,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3119526997","doi":"10.1145/3428077","title":"Context-Based Evaluation of Dimensionality Reduction Algorithms—Experiments and Statistical Significance Analysis","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"IBM (Canada); University of Alberta","funders":"","keywords":"Dimensionality reduction; Computer science; Data mining; Curse of dimensionality; Algorithm; Redundancy (engineering); Context (archaeology); Generalizability theory; Parametric statistics; Machine learning; Reduction (mathematics); Artificial intelligence; Mathematics; Statistics","authors":[{"name":"Aindrila Ghosh","is_ca":true},{"name":"Mona Nashaat","is_ca":true},{"name":"James Miller","is_ca":true},{"name":"Shaikh Quader","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1002620059718009,"gpt":0.3783655596835932,"spread":0.2781035537117924,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02975358,0.001966676,0.00120873,0.00342695,0.00120524,0.001885955,0.001580348,0.001726404,0.001135425],"category_scores_gemma":[0.1269986,0.0003713992,0.001784838,0.003211533,0.002067736,0.002753998,0.002322182,0.0017453,0.0004276221],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001007082,"about_ca_system_score_gemma":0.001496211,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001263761,"about_ca_topic_score_gemma":0.001224465,"domain_scores_codex":[0.9557993,0.0289112,0.004028373,0.00392074,0.006690117,0.0006502416],"domain_scores_gemma":[0.880228,0.08487849,0.006319347,0.0155427,0.01207087,0.0009605265],"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":[0.005469771,0.005147805,0.06422506,0.0073135,0.004243555,0.0005568605,0.002103827,0.2787105,0.03547942,0.03219227,0.01655585,0.5480015],"study_design_scores_gemma":[0.001113057,0.01180022,0.04674504,0.001150222,0.001571067,0.001137523,0.001801934,0.7949512,0.0742805,0.04378216,0.02111079,0.0005562126],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4810531,0.01056673,0.4917964,0.00102803,0.0009239874,0.003142475,0.003242603,0.002262943,0.005983756],"genre_scores_gemma":[0.6253403,0.00148576,0.3651791,0.0003961243,0.0002948105,0.003095797,0.00335065,0.0002840807,0.0005732671],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02975358,"threshold_uncertainty_score":0.1573539,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4317209824","doi":"10.1145/3580367","title":"Differentially Private Release of Heterogeneous Network for Managing Healthcare Data","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Concordia University","funders":"","keywords":"Flexibility (engineering); Computer science; Health care; Scalability; Variety (cybernetics); Data sharing; Data science; Enhanced Data Rates for GSM Evolution; Big data; Information exchange; Private information retrieval; Information sharing; Data mining; Computer security; World Wide Web; Database; Artificial intelligence","authors":[{"name":"Rashid Hussain Khokhar","is_ca":true},{"name":"Benjamin C. M. Fung","is_ca":true},{"name":"Farkhund Iqbal","is_ca":false},{"name":"Khalil Al-Hussaeni","is_ca":false},{"name":"Mohammed Hussain","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09979903213826967,"gpt":0.3332541246626596,"spread":0.23345509252439,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003653309,0.0005139298,0.0008069944,0.0007425875,0.001253022,0.001532416,0.001999087,0.001124597,0.00139497],"category_scores_gemma":[0.009331152,0.0002969851,0.0007010311,0.001247158,0.001235641,0.005118817,0.003743903,0.001253355,0.0003669042],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001455601,"about_ca_system_score_gemma":0.001162384,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009887405,"about_ca_topic_score_gemma":0.001088299,"domain_scores_codex":[0.996935,0.001279456,0.0001716311,0.0006484503,0.000713003,0.0002525245],"domain_scores_gemma":[0.9947119,0.001710701,0.0005585894,0.002522683,0.0003049737,0.0001912083],"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":[0.001572121,0.0002667675,0.005512164,0.0002587158,0.0001837162,0.001088419,0.001063185,0.4906022,0.02363507,0.193213,0.013932,0.2686726],"study_design_scores_gemma":[0.00004769675,0.0001061689,0.0005050456,0.00001306494,0.00003160691,0.0003578349,0.0001333829,0.9315695,0.008223752,0.0530332,0.005955028,0.00002376994],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04900011,0.0003512402,0.9465158,0.000813713,0.00009285005,0.0002090414,0.0004293523,0.0006376378,0.001950362],"genre_scores_gemma":[0.8304092,0.0003183693,0.1646127,0.0003551002,0.00008484987,0.0001986053,0.0009307917,0.00007416061,0.003016237],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003653309,"threshold_uncertainty_score":0.01932073,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4225986943","doi":"10.1145/3502726","title":"On the Robustness of Metric Learning: An Adversarial Perspective","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"National Science Foundation","keywords":"Robustness (evolution); Pairwise comparison; Computer science; Metric (unit); Adversarial system; Artificial intelligence; Machine learning","authors":[{"name":"Mengdi Huai","is_ca":false},{"name":"Tianhang Zheng","is_ca":true},{"name":"Chenglin Miao","is_ca":false},{"name":"Liuyi Yao","is_ca":false},{"name":"Aidong Zhang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04692128029731266,"gpt":0.309908607807081,"spread":0.2629873275097683,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008724856,0.002867329,0.001839927,0.002522672,0.001017471,0.002540281,0.002394927,0.003051115,0.001505144],"category_scores_gemma":[0.06184284,0.0008873204,0.001885651,0.00169526,0.005880516,0.007431112,0.005489551,0.006412732,0.0004715812],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002261467,"about_ca_system_score_gemma":0.00121863,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001732851,"about_ca_topic_score_gemma":0.0007233074,"domain_scores_codex":[0.9910679,0.003880535,0.0005443894,0.001660833,0.002289719,0.0005565733],"domain_scores_gemma":[0.9481629,0.03964544,0.00364277,0.005218845,0.002455746,0.000874379],"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":[0.0001659623,0.00006000792,0.002101506,0.0002023731,0.0001911554,0.0001921633,0.0001684388,0.8275742,0.00338991,0.12822,0.001254928,0.03647931],"study_design_scores_gemma":[0.000007899924,0.0001277084,0.00031284,0.00004748107,0.00002444312,0.0001214788,0.00002629316,0.9165372,0.001944743,0.07993735,0.0008845998,0.00002796576],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01619739,0.001809816,0.9777396,0.001089835,0.00009837188,0.00004830044,0.00007548917,0.0002429527,0.00269819],"genre_scores_gemma":[0.8461897,0.003681102,0.145357,0.0007511621,0.0005203259,0.0002490075,0.0003793454,0.0003290675,0.002543369],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008724856,"threshold_uncertainty_score":0.04614198,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399815117","doi":"10.1145/3673761","title":"Learning Individual Treatment Effects under Heterogeneous Interference in Networks","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Causal Inference Techniques","field":"Mathematics","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 Waterloo","funders":"National Natural Science Foundation of China","keywords":"Weighting; Benchmark (surveying); Interference (communication); Computer science; Generalization; Regression; Spillover effect; Observational study; Range (aeronautics); Confounding; Sample (material); Process (computing); Machine learning; Artificial intelligence; Mathematics; Statistics; Medicine","authors":[{"name":"Ziyu Zhao","is_ca":false},{"name":"Yuqi Bai","is_ca":true},{"name":"Ruoxuan Xiong","is_ca":false},{"name":"Qingyu Cao","is_ca":false},{"name":"Chao Ma","is_ca":false},{"name":"Ning Jiang","is_ca":false},{"name":"Fei Wu","is_ca":false},{"name":"Kun Kuang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.169365479198976,"gpt":0.4090027000751428,"spread":0.2396372208761667,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02238196,0.001759375,0.003431172,0.001872823,0.001030723,0.001929435,0.004115478,0.003208547,0.001976611],"category_scores_gemma":[0.06644186,0.001208616,0.002252161,0.002220202,0.002514911,0.004079605,0.003049316,0.004821913,0.0004722553],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002539011,"about_ca_system_score_gemma":0.002051357,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007067861,"about_ca_topic_score_gemma":0.00652333,"domain_scores_codex":[0.9858081,0.009180245,0.0004805138,0.003321516,0.0008216018,0.0003879888],"domain_scores_gemma":[0.9393761,0.05354351,0.002958707,0.002643198,0.001054514,0.0004239984],"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":[0.0006417006,0.00036883,0.01724811,0.000732196,0.0008809552,0.0002537027,0.0007145982,0.6464299,0.001019206,0.05056709,0.003741027,0.2774026],"study_design_scores_gemma":[0.0001136613,0.0000831752,0.001772994,0.00007393608,0.0001981453,0.00005098489,0.00005771133,0.8908637,0.0005343805,0.104644,0.00158455,0.00002283737],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02953356,0.001349153,0.9661303,0.001207571,0.00006241685,0.0002036539,0.0003003421,0.0003373097,0.0008756701],"genre_scores_gemma":[0.6409831,0.001833051,0.3485887,0.001561308,0.0005474329,0.001355509,0.001602105,0.0001621562,0.003366699],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02238196,"threshold_uncertainty_score":0.1183686,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4306407742","doi":"10.1145/3563947","title":"Lifelong Online Learning from Accumulated Knowledge","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Domain Adaptation and Few-Shot Learning","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":"Western University; Université Laval","funders":"Universidade de Macau","keywords":"Computer science; Lifelong learning; Task (project management); Artificial intelligence; Machine learning; Process (computing); Transfer of learning; Knowledge base; Online learning; Multi-task learning; Knowledge transfer; Knowledge management","authors":[{"name":"Changjian Shui","is_ca":true},{"name":"William Wang","is_ca":true},{"name":"Ihsen Hedhli","is_ca":true},{"name":"Chi Man Wong","is_ca":false},{"name":"Feng Wan","is_ca":false},{"name":"Boyu Wang","is_ca":true},{"name":"Christian Gagné","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09332883032205538,"gpt":0.3227769710322922,"spread":0.2294481407102368,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003019385,0.00150666,0.002116941,0.0009355801,0.0008289207,0.001701199,0.004165362,0.002084608,0.003430638],"category_scores_gemma":[0.01294441,0.0008364461,0.0007551648,0.001114632,0.001879949,0.006091597,0.00402462,0.003170045,0.001068883],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001284593,"about_ca_system_score_gemma":0.001860166,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002285255,"about_ca_topic_score_gemma":0.003478659,"domain_scores_codex":[0.9984699,0.0004339525,0.00009279959,0.0005567474,0.0002809043,0.0001656015],"domain_scores_gemma":[0.9913865,0.005918662,0.0004128315,0.00130527,0.0006007385,0.0003760037],"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":[0.0003920033,0.0006548032,0.002929919,0.0004611516,0.0001470349,0.0003126846,0.0003725291,0.6109626,0.002869973,0.04257281,0.006976719,0.3313478],"study_design_scores_gemma":[0.00001925992,0.00007267649,0.000183142,0.00002071944,0.00001347987,0.00004522245,0.00002708948,0.9478279,0.0009966219,0.04989397,0.0008867288,0.00001312704],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02023055,0.0006855964,0.9758618,0.0005259564,0.00006025943,0.0001028624,0.0001418064,0.0008167634,0.001574278],"genre_scores_gemma":[0.7585849,0.0006591877,0.2323412,0.0006914213,0.0002426798,0.0005127579,0.001159871,0.0002304976,0.005577468],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004165362,"threshold_uncertainty_score":0.0159682,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4220966911","doi":"10.1145/3502727","title":"Stochastic Variational Optimization of a Hierarchical Dirichlet Process Latent Beta-Liouville Topic Model","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Bayesian Methods and Mixture Models","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":"Concordia University","funders":"","keywords":"Latent Dirichlet allocation; Dirichlet process; Hierarchical Dirichlet process; Topic model; Computer science; Measure (data warehouse); Dirichlet distribution; Inference; Prior probability; Parametric statistics; Bayesian inference; Posterior probability; Mathematics; Bayesian probability; Applied mathematics; Artificial intelligence; Data mining; Statistics","authors":[{"name":"Koffi Eddy Ihou","is_ca":true},{"name":"Manar Amayri","is_ca":false},{"name":"Nizar Bouguila","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04890874881870901,"gpt":0.3044790594634345,"spread":0.2555703106447255,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003003508,0.001084221,0.00188884,0.000955915,0.0005818132,0.001995345,0.002457764,0.002408809,0.004363791],"category_scores_gemma":[0.006316235,0.00106298,0.001511776,0.001371748,0.001741178,0.002190324,0.002179776,0.002525476,0.0007825039],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002122641,"about_ca_system_score_gemma":0.002415505,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009584678,"about_ca_topic_score_gemma":0.01055693,"domain_scores_codex":[0.9988552,0.0005477867,0.00003966273,0.0002479093,0.0001902646,0.0001191603],"domain_scores_gemma":[0.99804,0.001472193,0.000107111,0.00007957705,0.0002042008,0.00009703454],"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":[0.000036371,0.00002785297,0.0004757103,0.00009379716,0.00003816555,0.00008849273,0.00009638261,0.8604539,0.0006214658,0.1261932,0.002169467,0.009705183],"study_design_scores_gemma":[0.000005873683,0.000004224346,0.00004242123,0.000005852576,0.000003264368,0.000007241191,0.000008244187,0.978869,0.00005936613,0.02044772,0.0005417023,0.000005115843],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007446159,0.000346289,0.9876891,0.0005571695,0.00004737103,0.00004750785,0.0001976751,0.00012457,0.003543992],"genre_scores_gemma":[0.4961058,0.001277062,0.4688516,0.0007188032,0.000316067,0.001006607,0.001578246,0.000601186,0.02954459],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009584678,"threshold_uncertainty_score":0.01905775,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4391026313","doi":"10.1145/3641108","title":"Interpreting Deep Forest through Feature Contribution and MDI Feature Importance","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Explainable Artificial Intelligence (XAI)","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":"Artificial Intelligence in Medicine (Canada)","funders":"China Postdoctoral Science Foundation; National Natural Science Foundation of China","keywords":"Interpretability; Categorical variable; Feature (linguistics); Random forest; Computer science; Tree (set theory); Artificial intelligence; Variety (cybernetics); Feature model; Data mining; Layer (electronics); Machine learning; Pattern recognition (psychology); Mathematics; Software","authors":[{"name":"Yi-Xiao He","is_ca":true},{"name":"Shen-Huan Lyu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03162727012661239,"gpt":0.3094050055243785,"spread":0.2777777353977661,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00291983,0.0008387495,0.0006601466,0.002015786,0.0005450215,0.001322347,0.001459462,0.0009185467,0.001363619],"category_scores_gemma":[0.01267661,0.0004030963,0.000605987,0.001176991,0.001235656,0.00313847,0.001875747,0.001915853,0.0002765679],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001021225,"about_ca_system_score_gemma":0.0009711304,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003596465,"about_ca_topic_score_gemma":0.005124371,"domain_scores_codex":[0.99921,0.0002474354,0.0000457223,0.0001967903,0.0002105772,0.000089357],"domain_scores_gemma":[0.9953935,0.002869468,0.0004247228,0.0005264149,0.0006342753,0.0001516439],"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":[0.00023408,0.0001565961,0.02214518,0.0002518524,0.0001128451,0.0003884978,0.0004257333,0.4774155,0.00865249,0.1148135,0.004986024,0.3704177],"study_design_scores_gemma":[0.000007325226,0.00001581593,0.001080181,0.00001605203,0.00001309525,0.00005050493,0.00002356094,0.9366219,0.001598621,0.05986723,0.0006939699,0.00001177496],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02730687,0.0002639818,0.970561,0.000287015,0.00003116623,0.00003579173,0.0001126013,0.0004712312,0.0009303527],"genre_scores_gemma":[0.6904657,0.0003182907,0.3067762,0.0002045295,0.0000904221,0.000113755,0.0004979827,0.0001829947,0.001350084],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003596465,"threshold_uncertainty_score":0.01544166,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2791982474","doi":"10.1145/3023363","title":"Visual Analysis of Brain Networks Using Sparse Regression Models","year":2018,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Army Research Office; National Institute of Biomedical Imaging and Bioengineering; DoD Alzheimer's Disease Neuroimaging Initiative; Defense Threat Reduction Agency; Canadian Institutes of Health Research; National Institute on Aging; National Institutes of Health; National Natural Science Foundation of China","keywords":"Visual analytics; Computer science; Discriminative model; Visualization; Artificial intelligence; Machine learning; Intersection (aeronautics); Human–computer interaction","authors":[{"name":"Lei Shi","is_ca":false},{"name":"Hanghang Tong","is_ca":false},{"name":"Madelaine Daianu","is_ca":false},{"name":"Feng Tian","is_ca":false},{"name":"Paul M. Thompson","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09753420314028476,"gpt":0.3740269390566549,"spread":0.2764927359163701,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00160969,0.001095852,0.0005614434,0.001945667,0.0003134688,0.001676226,0.0008897752,0.000683139,0.003677198],"category_scores_gemma":[0.007184749,0.0004391433,0.001010382,0.0008117484,0.0005600204,0.001717113,0.001937957,0.001229253,0.000584753],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005750563,"about_ca_system_score_gemma":0.0005131713,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002128297,"about_ca_topic_score_gemma":0.003166077,"domain_scores_codex":[0.9994254,0.0002985742,0.0000193083,0.0001008039,0.0001225694,0.00003334184],"domain_scores_gemma":[0.9980384,0.001316296,0.0001916121,0.0001978261,0.0001808301,0.00007499838],"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":[0.000479549,0.0001406491,0.002830996,0.0003866765,0.0001625855,0.0003870222,0.0009692682,0.6431458,0.05022204,0.02906737,0.007708958,0.264499],"study_design_scores_gemma":[0.00001593737,0.0000347639,0.0005166578,0.00001874386,0.00001168929,0.00008115799,0.00007380589,0.9744108,0.004492682,0.01890995,0.0014191,0.00001466288],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02342485,0.0001482157,0.9728295,0.0004385345,0.00001682534,0.00004137179,0.0002189718,0.00195294,0.000928808],"genre_scores_gemma":[0.4277291,0.0006135806,0.568591,0.0001679461,0.0000659884,0.00017219,0.0006661568,0.000639444,0.001354668],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003677198,"threshold_uncertainty_score":0.01230145,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408221131","doi":"10.1145/3721983","title":"SED2AM: Solving Multi-Trip Time-Dependent Vehicle Routing Problem Using Deep Reinforcement Learning","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Smart Parking Systems Research","field":"Engineering","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"National Research Council Canada; University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Reinforcement learning; Vehicle routing problem; Computer science; Routing (electronic design automation); Artificial intelligence; Reinforcement; Mathematical optimization; Machine learning; Mathematics; Engineering; Computer network","authors":[{"name":"Arash Mozhdehi","is_ca":true},{"name":"Yunli Wang","is_ca":true},{"name":"Sun Sun","is_ca":true},{"name":"Xin Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04577542143170239,"gpt":0.3068384094486623,"spread":0.2610629880169599,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008849364,0.001096896,0.001037064,0.0004963949,0.0002919098,0.0007217231,0.002034654,0.001204723,0.002500915],"category_scores_gemma":[0.002493943,0.0005586143,0.0007831141,0.0005563481,0.000545259,0.001154022,0.001159731,0.002027834,0.0005109726],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001116907,"about_ca_system_score_gemma":0.002046055,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01278268,"about_ca_topic_score_gemma":0.01883083,"domain_scores_codex":[0.9996766,0.00009991512,0.00001650323,0.0001003901,0.00005257871,0.00005413895],"domain_scores_gemma":[0.9991961,0.0005151466,0.00006754343,0.00006887693,0.00009734188,0.00005500005],"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":[0.00006165716,0.00012227,0.001398371,0.00009550885,0.00006473283,0.00006967853,0.00004072507,0.8982621,0.0009258761,0.00421136,0.00390756,0.0908401],"study_design_scores_gemma":[0.00000636504,0.00001180781,0.0000381491,0.000002623151,0.000003345278,0.000004432881,0.000003278129,0.9981508,0.0001743179,0.001384349,0.0002189953,0.000001541459],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05463678,0.0008796024,0.9349949,0.0008279784,0.0001510018,0.0001240444,0.0004527343,0.003675127,0.004257772],"genre_scores_gemma":[0.7367107,0.0003608075,0.2554043,0.0006337798,0.00009176603,0.0002560044,0.001293591,0.0002671391,0.004981884],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01278268,"threshold_uncertainty_score":0.02541655,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4313437244","doi":"10.1145/3578586","title":"Summarizing User-item Matrix By Group Utility Maximization","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and 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":"Simon Fraser University","funders":"","keywords":"Greedy algorithm; Computer science; Selection (genetic algorithm); Population; Similarity (geometry); Group (periodic table); Maximization; Data mining; Preference; Cluster analysis; Information retrieval; Theoretical computer science; Artificial intelligence; Algorithm; Mathematical optimization; Mathematics; Statistics","authors":[{"name":"Yongjie Wang","is_ca":false},{"name":"Ke Wang","is_ca":true},{"name":"Cheng Long","is_ca":false},{"name":"Chunyan Miao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06137206547769925,"gpt":0.3151175045849592,"spread":0.2537454391072599,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005186846,0.002130528,0.002749487,0.001817507,0.0007964453,0.001982058,0.002702496,0.001566115,0.002708581],"category_scores_gemma":[0.01300673,0.0009199542,0.001387492,0.004304815,0.0009159183,0.004154531,0.002128079,0.001668152,0.001162516],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001618165,"about_ca_system_score_gemma":0.001847778,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004764843,"about_ca_topic_score_gemma":0.006427701,"domain_scores_codex":[0.9957985,0.002196511,0.000206781,0.0008570653,0.0006508101,0.000290311],"domain_scores_gemma":[0.9946431,0.003335195,0.000382829,0.0006433123,0.0007567223,0.000238864],"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":[0.000371552,0.0003498326,0.004762113,0.0005122391,0.0004132624,0.0002564578,0.0006003747,0.7154962,0.0038186,0.04401039,0.01512704,0.214282],"study_design_scores_gemma":[0.00002540214,0.00008724707,0.0003798124,0.00001371148,0.00004970345,0.0000717142,0.00007450442,0.9625072,0.0007816516,0.03449769,0.001491454,0.00001977498],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01090069,0.0003810503,0.9865541,0.000314354,0.00002570116,0.0002057061,0.0003864466,0.0003629279,0.0008689581],"genre_scores_gemma":[0.34218,0.0007814487,0.6479152,0.0004235131,0.0001707692,0.0008971153,0.002666049,0.0002670297,0.004698923],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005186846,"threshold_uncertainty_score":0.02743095,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4319349118","doi":"10.1145/3563040","title":"DeltaShield: Information Theory for Human- Trafficking Detection","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Spam and Phishing Detection","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":"McGill University","funders":"National Science Foundation","keywords":"Computer science; Interpretability; Scalability; Spotting; Domain (mathematical analysis); Law enforcement; Laptop; Information retrieval; Data mining; Matching (statistics); Artificial intelligence; Database","authors":[{"name":"Catalina Vajiac","is_ca":false},{"name":"Meng-Chieh Lee","is_ca":false},{"name":"Aayushi Kulshrestha","is_ca":true},{"name":"Sacha Lévy","is_ca":true},{"name":"Namyong Park","is_ca":false},{"name":"Andreas Olligschlaeger","is_ca":false},{"name":"Cara Jones","is_ca":false},{"name":"Reihaneh Rabbany","is_ca":true},{"name":"Christos Faloutsos","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05985050683720564,"gpt":0.3110097719310146,"spread":0.2511592650938089,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006366846,0.00133591,0.002069376,0.007617709,0.001205812,0.003743444,0.00281259,0.00214894,0.006852128],"category_scores_gemma":[0.03694038,0.001087188,0.001707792,0.005263496,0.002374706,0.007302968,0.002858155,0.003863789,0.002458626],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003461247,"about_ca_system_score_gemma":0.003025845,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007198271,"about_ca_topic_score_gemma":0.005846748,"domain_scores_codex":[0.9957352,0.001661923,0.0002714169,0.0007042718,0.001430161,0.0001969779],"domain_scores_gemma":[0.9734607,0.02137815,0.0007634778,0.002365985,0.001744455,0.0002872043],"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":[0.0004091464,0.0002634937,0.00489338,0.00107468,0.0003354607,0.0003338617,0.0003993181,0.2182462,0.002054524,0.2529109,0.06048967,0.4585893],"study_design_scores_gemma":[0.00003286169,0.00004377522,0.0005168497,0.00008309272,0.00003284299,0.0001080712,0.00005528268,0.7159135,0.001225067,0.2732481,0.008703804,0.00003665801],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005061099,0.002201581,0.9803991,0.00132092,0.0001705548,0.0001679774,0.001920664,0.003991156,0.004766905],"genre_scores_gemma":[0.3255334,0.003430679,0.6563412,0.001486105,0.0009824674,0.0009423281,0.004571973,0.0008396819,0.005872179],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007617709,"threshold_uncertainty_score":0.0336715,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4366087542","doi":"10.1145/3592857","title":"Modeling Regime Shifts in Multiple Time Series","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Time Series Analysis and Forecasting","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; Université de Sherbrooke","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Series (stratigraphy); Computer science; Time series; Dependency (UML); Representation (politics); Grid; Machine learning; Artificial intelligence; Data mining; Mathematics","authors":[{"name":"Etienne Gael Tajeuna","is_ca":true},{"name":"Mohamed Bouguessa","is_ca":true},{"name":"Shengrui Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06498033503273945,"gpt":0.2730715737043471,"spread":0.2080912386716076,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002155861,0.0005353346,0.0006789092,0.001261362,0.0003612713,0.0009547403,0.001466627,0.0011224,0.001132561],"category_scores_gemma":[0.007990829,0.0003377039,0.001002479,0.001174262,0.0007979368,0.002082985,0.001248466,0.001608995,0.0001635718],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007642749,"about_ca_system_score_gemma":0.0004837663,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004167433,"about_ca_topic_score_gemma":0.002946609,"domain_scores_codex":[0.9993482,0.0001623685,0.00004027778,0.000247012,0.000122147,0.00008002389],"domain_scores_gemma":[0.9967709,0.002070523,0.0006998152,0.0002261122,0.0001222226,0.0001104051],"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":[0.00005410289,0.00007740819,0.01915495,0.00009176744,0.000169271,0.0002887562,0.0002473379,0.9032897,0.002350923,0.04331004,0.0004946891,0.03047098],"study_design_scores_gemma":[0.000002183392,0.00001440245,0.001301214,0.000006531256,0.00001154816,0.00002037761,0.00001680062,0.9816559,0.0001730279,0.01645178,0.0003405492,0.000005666931],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.293159,0.0007485248,0.7016759,0.001045418,0.00009444123,0.00005781512,0.0003629408,0.0003681704,0.002487859],"genre_scores_gemma":[0.9594027,0.0004564543,0.03826475,0.00008101848,0.00008177425,0.00006972969,0.0002032865,0.00003225814,0.001407949],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004167433,"threshold_uncertainty_score":0.01140141,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2257410811","doi":"10.1145/3365672","title":"A New Smooth Approximation to the Zero One Loss with a Probabilistic Interpretation","year":2019,"lang":"en","type":"preprint","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Polytechnique Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Hinge loss; Algorithm; Computer science; Probabilistic logic; Maxima and minima; Artificial intelligence; Mathematics; Benchmark (surveying); Mathematical optimization; Machine learning; Pattern recognition (psychology); Support vector machine","authors":[{"name":"K. M. Azharul Hasan","is_ca":true},{"name":"Christopher Pal","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04480686181851787,"gpt":0.2922726434183354,"spread":0.2474657815998175,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004260919,0.001613886,0.001286351,0.001785732,0.0004972948,0.002421061,0.002937926,0.0018072,0.002255877],"category_scores_gemma":[0.01443675,0.0006885165,0.001174595,0.001698346,0.002117438,0.004091696,0.002964606,0.003757273,0.001284666],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001360316,"about_ca_system_score_gemma":0.001830917,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001942482,"about_ca_topic_score_gemma":0.001481538,"domain_scores_codex":[0.9974664,0.0008675138,0.0001131039,0.0004897101,0.0008796952,0.0001835314],"domain_scores_gemma":[0.9955639,0.002358324,0.00051476,0.0007130739,0.0006796796,0.0001703255],"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":[0.000269928,0.0002099284,0.00314695,0.0002596287,0.0001724282,0.0003012864,0.000190106,0.643355,0.006884985,0.1347636,0.01103256,0.1994137],"study_design_scores_gemma":[0.00001335167,0.0000634009,0.0003336038,0.00002051893,0.00001312533,0.0000777551,0.00001063757,0.9612232,0.0009905138,0.03541619,0.001819495,0.00001831446],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005807767,0.0002710307,0.9920567,0.0003460587,0.00003371428,0.00002852043,0.0000786449,0.0003064206,0.001071225],"genre_scores_gemma":[0.4355817,0.001194834,0.5498866,0.0008931443,0.000414704,0.0003932716,0.001046643,0.0006626853,0.009926406],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004260919,"threshold_uncertainty_score":0.02253413,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3044604502","doi":"10.1145/3480243","title":"Time-Aware Graph Embedding: A Temporal Smoothness and Task-Oriented Approach","year":2021,"lang":"en","type":"preprint","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Graph Neural Networks","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":"Simon Fraser University","funders":"","keywords":"Embedding; Computer science; Granularity; Theoretical computer science; Graph embedding; Graph; Smoothing; Artificial intelligence; Machine learning","authors":[{"name":"Yonghui Xu","is_ca":false},{"name":"Shengjie Sun","is_ca":false},{"name":"Yuan Miao","is_ca":false},{"name":"Dong Yang","is_ca":false},{"name":"Xiaonan Meng","is_ca":false},{"name":"Yi Hu","is_ca":false},{"name":"Ke Wang","is_ca":true},{"name":"Hengjie Song","is_ca":false},{"name":"Chuanyan Miao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03824167191490641,"gpt":0.2946700430901839,"spread":0.2564283711752775,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001279386,0.001256676,0.001149689,0.001908962,0.0004874457,0.00100857,0.002225307,0.001458835,0.001375469],"category_scores_gemma":[0.006142728,0.0005106864,0.00110763,0.002621531,0.000873297,0.00396236,0.002074467,0.002094727,0.0004696858],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007094188,"about_ca_system_score_gemma":0.00106988,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004202028,"about_ca_topic_score_gemma":0.006217894,"domain_scores_codex":[0.998956,0.0002544256,0.00007038875,0.0004070536,0.0002372725,0.00007498383],"domain_scores_gemma":[0.9970838,0.001292926,0.0003649219,0.0006855264,0.0004201513,0.0001525885],"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":[0.0003223035,0.0003286632,0.003512353,0.0003024962,0.0001752117,0.0001948134,0.0004167361,0.4350893,0.01371025,0.03551051,0.005647555,0.5047897],"study_design_scores_gemma":[0.00001205699,0.00005588982,0.0004408972,0.000009914334,0.0000296135,0.0000557201,0.00003693056,0.9750009,0.001453538,0.02133898,0.001548658,0.00001698523],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01227287,0.0002994153,0.9859934,0.0001884553,0.00003240304,0.00004306302,0.000130022,0.0005526141,0.0004877389],"genre_scores_gemma":[0.4635803,0.000868818,0.5289403,0.0002614194,0.0001971656,0.0002121404,0.001610961,0.0003725995,0.003956297],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004202028,"threshold_uncertainty_score":0.008355141,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4281975883","doi":"10.1145/3805034","title":"Streaming Hierarchical Clustering Based on Point-Set Kernel","year":2022,"lang":"en","type":"preprint","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"State Key Laboratory of Novel Software Technology; Nanjing University; National Natural Science Foundation of China","keywords":"Hierarchical clustering; Cluster analysis; Computer science; Data mining; Single-linkage clustering; Brown clustering; Tree (set theory); Consensus clustering; Correlation clustering; Hierarchical clustering of networks; CURE data clustering algorithm; Scalability; Kernel (algebra); Tree structure; Constrained clustering; Artificial intelligence; Mathematics; Algorithm; Database; Binary tree","authors":[{"name":"Xin Han","is_ca":false},{"name":"Ye Zhu","is_ca":false},{"name":"Kai Ming Ting","is_ca":true},{"name":"De‐Chuan Zhan","is_ca":true},{"name":"Gang Li","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07734526681048454,"gpt":0.3535982704819712,"spread":0.2762530036714866,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001455392,0.0005267787,0.001315876,0.002759677,0.0008972445,0.001314365,0.002190265,0.0007687507,0.001316321],"category_scores_gemma":[0.006048637,0.0003819089,0.0008336937,0.003397587,0.0007032523,0.001922261,0.001731883,0.0009941638,0.0006816812],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00188961,"about_ca_system_score_gemma":0.001559779,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0141627,"about_ca_topic_score_gemma":0.01035128,"domain_scores_codex":[0.9982589,0.0003404208,0.0001097651,0.0003704893,0.0007787325,0.0001417946],"domain_scores_gemma":[0.995756,0.0009611777,0.0004191301,0.0009189238,0.001705312,0.0002395108],"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":[0.0003153714,0.0001759995,0.005042243,0.0001634539,0.0001364812,0.0001674546,0.0003541198,0.7325006,0.01219733,0.02312427,0.004528582,0.2212942],"study_design_scores_gemma":[0.000004346829,0.00001298823,0.0003058817,0.000002004946,0.000003995419,0.00001798025,0.00001496327,0.9952342,0.00105241,0.003011401,0.0003328602,0.000007031773],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05053889,0.0002171329,0.9464798,0.0001070243,0.00004018405,0.0001092222,0.0001956824,0.001368327,0.0009438687],"genre_scores_gemma":[0.5859494,0.0001438988,0.4111653,0.00005346222,0.00003888406,0.00009514342,0.0008726843,0.0001427719,0.001538428],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0141627,"threshold_uncertainty_score":0.02816051,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2325563383","doi":"10.1145/2297456.2297460","title":"The Latent Maximum Entropy Principle","year":2012,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Statistical Mechanics and Entropy","field":"Physics and Astronomy","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 Alberta","funders":"","keywords":"Principle of maximum entropy; Mathematics; Statistics","authors":[{"name":"Dale Schuurmans","is_ca":true},{"name":"Yunxin Zhao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04418102761629817,"gpt":0.3089701903070841,"spread":0.2647891626907859,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005099506,0.001044728,0.001387322,0.001804665,0.000966015,0.002896989,0.002165378,0.001644857,0.004605115],"category_scores_gemma":[0.01962971,0.0009034575,0.001786218,0.001721728,0.003760275,0.006023741,0.004261637,0.003110921,0.001316681],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001161209,"about_ca_system_score_gemma":0.001913323,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009102251,"about_ca_topic_score_gemma":0.0007961736,"domain_scores_codex":[0.9956778,0.002160888,0.0001774828,0.0005776612,0.001156919,0.0002493636],"domain_scores_gemma":[0.9902418,0.007107961,0.0005502017,0.001324836,0.0005597707,0.0002154249],"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":[0.00005552924,0.00006434369,0.001225199,0.0002583129,0.0001173074,0.0001510116,0.000299987,0.1122988,0.002478492,0.8155758,0.002874766,0.06460045],"study_design_scores_gemma":[0.00001685685,0.00004228239,0.0003898884,0.00004842678,0.00001773101,0.00009627628,0.0000222902,0.2721435,0.001131174,0.7200282,0.006030182,0.00003316287],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003635573,0.0003020289,0.9911523,0.0004972667,0.00004188533,0.0000375137,0.0001152608,0.0001123575,0.004105724],"genre_scores_gemma":[0.3969727,0.001673875,0.5913596,0.0008821021,0.0005597147,0.000737852,0.0006482271,0.0003245152,0.006841457],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005099506,"threshold_uncertainty_score":0.02696908,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413592351","doi":"10.1145/3762667","title":"ORIC V2: Improved Feature Interaction Detection Model through Online Random Interaction Chains for Click-Through Rate Prediction","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","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":"Petro-Canada","funders":"National Natural Science Foundation of China","keywords":"Feature (linguistics); Random forest; Computer science; Pattern recognition (psychology); Artificial intelligence; Data mining","authors":[{"name":"Yannian Kou","is_ca":false},{"name":"Qiuqiang Lin","is_ca":false},{"name":"Di Fan","is_ca":false},{"name":"Chuanhou Gao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0650397683175733,"gpt":0.3459675967337768,"spread":0.2809278284162035,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00297911,0.001476603,0.002170297,0.001963459,0.0005947117,0.001515014,0.00409313,0.001660962,0.002796754],"category_scores_gemma":[0.008065027,0.0008684609,0.001390595,0.001962289,0.0006493547,0.00205092,0.001407188,0.002665124,0.001183508],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00126342,"about_ca_system_score_gemma":0.002136372,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02733926,"about_ca_topic_score_gemma":0.02915155,"domain_scores_codex":[0.9985366,0.0003128565,0.0000912771,0.0004837599,0.0003926852,0.000182817],"domain_scores_gemma":[0.9953372,0.002566422,0.0003384814,0.0005896179,0.0009517053,0.0002164588],"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":[0.0005871164,0.0004611477,0.01482711,0.0001948535,0.0002654794,0.0002452363,0.0001747102,0.7666888,0.003504061,0.010844,0.008750154,0.1934573],"study_design_scores_gemma":[0.00000874421,0.00002117412,0.0002147218,0.000003374712,0.000008839348,0.00001489577,0.000002513836,0.9983448,0.0001767664,0.0009481709,0.0002487463,0.000007299348],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0456921,0.001021462,0.9451379,0.0004412373,0.0001356321,0.000287497,0.001261694,0.003607277,0.002415356],"genre_scores_gemma":[0.7348792,0.0007705444,0.2498304,0.0005475893,0.0002578443,0.0006070639,0.004353062,0.0002802071,0.008474082],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02733926,"threshold_uncertainty_score":0.05436021,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}