{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":26,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":26,"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":"a6c2b8c45c44","filters":{"venue":"Information Retrieval"}},"results":[{"id":"W1685426458","doi":"10.1007/s10791-011-9162-z","title":"Efficient and effective spam filtering and re-ranking for large web datasets","year":2011,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":277,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"National Institute of Standards and Technology","keywords":"Computer science; Information retrieval; Ranking (information retrieval); Relevance feedback; Relevance (law); Search engine; Learning to rank; Set (abstract data type); Honeypot; Web page; Rank (graph theory); Data mining; World Wide Web; Artificial intelligence; Image retrieval; Mathematics","authors":[{"name":"Gordon V. Cormack","is_ca":true},{"name":"Mark D. Smucker","is_ca":true},{"name":"Charles L. A. Clarke","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01549341487547832,"gpt":0.2325380502429347,"spread":0.2170446353674564,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004885714,0.001951224,0.00350499,0.008493712,0.001982605,0.00349207,0.002250577,0.002585345,0.002198454],"category_scores_gemma":[0.01644396,0.001046922,0.001901199,0.006012437,0.0007814644,0.005201903,0.001972395,0.001663329,0.003654731],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001269435,"about_ca_system_score_gemma":0.003098424,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007594257,"about_ca_topic_score_gemma":0.01963901,"domain_scores_codex":[0.9945162,0.001690384,0.0005396132,0.0008355431,0.00187982,0.0005383738],"domain_scores_gemma":[0.9875721,0.004589205,0.0008445789,0.003890014,0.002685506,0.0004186419],"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":[0.0005535228,0.0008198719,0.009108506,0.0005532436,0.0004130536,0.0002274252,0.0001915014,0.03147718,0.02803778,0.002556209,0.07151861,0.854543],"study_design_scores_gemma":[0.0001467948,0.0002221295,0.007561836,0.00003258008,0.000204024,0.0005167365,0.0002399054,0.9496571,0.01826344,0.01387241,0.009207196,0.00007591258],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1723719,0.007815761,0.7662311,0.003053631,0.0007659636,0.0009342374,0.007597427,0.03760267,0.003627311],"genre_scores_gemma":[0.2854189,0.001133407,0.6916398,0.0004526824,0.000863334,0.0003819783,0.01340214,0.0006226762,0.006084955],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008493712,"threshold_uncertainty_score":0.02583843,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2151752770","doi":"10.1023/b:inrt.0000011209.19643.e2","title":"Augmenting Naive Bayes Classifiers with Statistical Language Models","year":2004,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Topic Modeling","field":"Computer Science","cited_by":247,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Naive Bayes classifier; Artificial intelligence; Computer science; Machine learning; Bayes error rate; Bayes classifier; Bayes' theorem; Classifier (UML); Conditional independence; Bayesian programming; Natural language processing; Pattern recognition (psychology); Bayes factor; Bayesian probability; Support vector machine","authors":[{"name":"Fuchun Peng","is_ca":false},{"name":"Dale Schuurmans","is_ca":true},{"name":"Shaojun Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01324613556970664,"gpt":0.2371975164125819,"spread":0.2239513808428753,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007708231,0.001956006,0.002820781,0.004448228,0.001149722,0.00313319,0.002370568,0.00246437,0.004933987],"category_scores_gemma":[0.03147248,0.001034707,0.002005483,0.003635254,0.000610854,0.007949124,0.001710059,0.003128775,0.007152606],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008729438,"about_ca_system_score_gemma":0.001638349,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007244185,"about_ca_topic_score_gemma":0.01027938,"domain_scores_codex":[0.9940152,0.00296964,0.0004259123,0.00075319,0.001595023,0.0002410549],"domain_scores_gemma":[0.9773309,0.01723344,0.0004739081,0.001404635,0.003345979,0.0002110581],"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":[0.0005874375,0.0006005458,0.003049169,0.0005393695,0.0004737809,0.0001465533,0.0002047348,0.06320536,0.005727659,0.006089303,0.02235456,0.8970215],"study_design_scores_gemma":[0.00008858077,0.0001329543,0.000735759,0.00008684809,0.0003267741,0.0001773122,0.00007593527,0.9551525,0.003909549,0.03268756,0.006566804,0.00005937217],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02238299,0.004859613,0.9585693,0.001283084,0.0007891072,0.0002622874,0.0008154029,0.006360177,0.00467802],"genre_scores_gemma":[0.3074251,0.003535208,0.6714925,0.001190279,0.001743997,0.0005941053,0.004714453,0.0008093884,0.008495064],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007708231,"threshold_uncertainty_score":0.04076552,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2084946528","doi":"10.1007/s10791-006-9020-6","title":"Knowledge-based query expansion to support scenario-specific retrieval of medical free text","year":2007,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":79,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Institutes of Health; McMaster University","keywords":"Computer science; Query expansion; Unified Medical Language System; Information retrieval; Query language; Testbed; Query optimization; Precision and recall; Exploit; Recall; Data mining; World Wide Web","authors":[{"name":"Zhenyu Liu","is_ca":false},{"name":"Wesley W. Chu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02053831162805873,"gpt":0.3008048304354492,"spread":0.2802665188073905,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001597605,0.0006491434,0.0007664282,0.002762076,0.0004870538,0.001115879,0.001326372,0.001146939,0.00836336],"category_scores_gemma":[0.008567872,0.0003156996,0.0007312807,0.001769121,0.00032029,0.00238987,0.001527379,0.0006612243,0.00248776],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006379624,"about_ca_system_score_gemma":0.0008615773,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003199656,"about_ca_topic_score_gemma":0.004793975,"domain_scores_codex":[0.9988093,0.00043616,0.0001576911,0.0001853595,0.0003192354,0.00009224516],"domain_scores_gemma":[0.9966466,0.002170275,0.0001400595,0.0003477118,0.0005828404,0.0001124321],"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":[0.001815468,0.0009832729,0.006045635,0.001499178,0.0002964464,0.002838877,0.001911335,0.06046873,0.08260421,0.02281441,0.09860475,0.7201176],"study_design_scores_gemma":[0.0003096825,0.0002122091,0.003352615,0.0001591405,0.0002594514,0.001807261,0.0007680311,0.8807873,0.04234467,0.02346915,0.04642054,0.000109924],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09532925,0.001011649,0.856819,0.00187224,0.0001984119,0.0010242,0.005377531,0.02643069,0.01193707],"genre_scores_gemma":[0.468316,0.0006388643,0.5130407,0.0006193874,0.0001455141,0.0005836184,0.01272979,0.0006388624,0.00328711],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00836336,"threshold_uncertainty_score":0.02797824,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2032328503","doi":"10.1007/s10791-009-9108-x","title":"Document clustering of scientific texts using citation contexts","year":2009,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":77,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Cluster analysis; Information retrieval; Computer science; Document clustering; Citation; Vocabulary; Context (archaeology); Similarity (geometry); Representation (politics); Document retrieval; Natural language processing; Artificial intelligence; World Wide Web; Linguistics","authors":[{"name":"Bader Aljaber","is_ca":false},{"name":"Nicola Stokes","is_ca":false},{"name":"James Bailey","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0160256859248469,"gpt":0.2950436954188471,"spread":0.2790180094940002,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.001669988,0.0008086043,0.001106689,0.02648126,0.002412321,0.004088819,0.001049552,0.001127128,0.003157671],"category_scores_gemma":[0.01184732,0.0003891796,0.001247355,0.02233986,0.0006420256,0.002487009,0.00120861,0.0009756474,0.002097595],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001043688,"about_ca_system_score_gemma":0.002374801,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00395782,"about_ca_topic_score_gemma":0.008434736,"domain_scores_codex":[0.9978617,0.0005519332,0.0003084471,0.0004282437,0.0006915503,0.0001581303],"domain_scores_gemma":[0.9923053,0.003579697,0.0006003455,0.0005478903,0.002680715,0.0002859683],"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":[0.0008256128,0.0004109048,0.01502229,0.001657673,0.0004256607,0.0003710898,0.001342308,0.01013203,0.03755221,0.01368703,0.01772108,0.900852],"study_design_scores_gemma":[0.000441689,0.001151652,0.100115,0.001224836,0.003434505,0.002277089,0.003867063,0.507206,0.1036586,0.1263126,0.1497516,0.0005594292],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4352847,0.02988485,0.488127,0.002138877,0.00194183,0.001189411,0.009118278,0.005384562,0.02693051],"genre_scores_gemma":[0.6016658,0.006460908,0.3663079,0.0001524675,0.001650409,0.0005931752,0.01205742,0.0006688847,0.01044304],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9735187,"threshold_uncertainty_score":0.01056349,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2033681837","doi":"10.1007/s10791-011-9163-y","title":"Improving document clustering using Okapi BM25 feature weighting","year":2011,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Image Retrieval and Classification Techniques","field":"Computer Science","cited_by":77,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Weighting; Cluster analysis; Computer science; Artificial intelligence; Data mining; Feature (linguistics); Document clustering; Pattern recognition (psychology); Term (time)","authors":[{"name":"John S. Whissell","is_ca":true},{"name":"Charles L. A. Clarke","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02482047029488173,"gpt":0.2420373186875677,"spread":0.2172168483926859,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001286215,0.001523025,0.002147975,0.003822709,0.001504776,0.001497301,0.001797576,0.001301136,0.003830502],"category_scores_gemma":[0.003912301,0.0003931382,0.001579602,0.0052016,0.0003445635,0.00181931,0.001047734,0.001271639,0.004966393],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009573669,"about_ca_system_score_gemma":0.001502157,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01665235,"about_ca_topic_score_gemma":0.0221511,"domain_scores_codex":[0.9980074,0.0003271185,0.0001747653,0.0004182296,0.0008859279,0.0001865955],"domain_scores_gemma":[0.9984641,0.0002158647,0.0000707106,0.000296273,0.0008949515,0.00005813068],"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":[0.0003351036,0.0003208881,0.0009958755,0.0001676287,0.0001665119,0.0000499503,0.00005875853,0.01102796,0.03079706,0.001411118,0.02052903,0.9341402],"study_design_scores_gemma":[0.0001069064,0.0003264932,0.006820633,0.00005601645,0.0005100734,0.0004436292,0.0001924954,0.8821591,0.08027036,0.007953618,0.02102135,0.0001392535],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04692125,0.004228389,0.9325089,0.0002710973,0.0007061816,0.000263758,0.0009773949,0.01039429,0.003728774],"genre_scores_gemma":[0.2293731,0.001367304,0.7497625,0.0003207954,0.0003474439,0.0002854152,0.005917789,0.0009853313,0.01164025],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01665235,"threshold_uncertainty_score":0.03311086,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2141520705","doi":"10.1023/a:1026028229881","title":"Applying Machine Learning to Text Segmentation for Information Retrieval","year":2003,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Topic Modeling","field":"Computer Science","cited_by":76,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Mitacs","keywords":"Segmentation; Computer science; Text segmentation; Artificial intelligence; Pattern recognition (psychology); Natural language processing; Word (group theory); Mathematics","authors":[{"name":"Fuchun Peng","is_ca":true},{"name":"Dale Schuurmans","is_ca":true},{"name":"Nick Cercone","is_ca":true},{"name":"Stephen Robertson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01664541220876287,"gpt":0.2556849407473694,"spread":0.2390395285386065,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002340805,0.001185265,0.001679018,0.004115161,0.001110679,0.002500342,0.001268672,0.001517651,0.003127824],"category_scores_gemma":[0.00960903,0.0007180565,0.001551814,0.003868007,0.0008875153,0.003265153,0.001038711,0.00174213,0.00299296],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001234185,"about_ca_system_score_gemma":0.001421111,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00624612,"about_ca_topic_score_gemma":0.006248933,"domain_scores_codex":[0.9980595,0.0009328028,0.0001661074,0.0003937302,0.0003157029,0.0001322089],"domain_scores_gemma":[0.994518,0.003997478,0.0002636971,0.000485777,0.0006177155,0.000117413],"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":[0.0003393371,0.0003601187,0.001841152,0.000602051,0.0002307993,0.0001377912,0.0004224027,0.05574257,0.03849183,0.008367009,0.01118995,0.8822751],"study_design_scores_gemma":[0.00004460394,0.0001103723,0.001212593,0.00003842956,0.0001122746,0.0001072889,0.0001463887,0.9312439,0.02028291,0.04036074,0.006301579,0.00003882279],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0164607,0.001856822,0.9744391,0.0004295451,0.0001664203,0.0001914365,0.000233325,0.004581825,0.001640817],"genre_scores_gemma":[0.1985327,0.001685588,0.7919687,0.0002758656,0.0005119633,0.0003276343,0.001710998,0.0008884883,0.004098064],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00624612,"threshold_uncertainty_score":0.01241952,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1503824486","doi":"10.1023/a:1023936321956","title":"Query Expansion with Long-Span Collocates","year":2003,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":53,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Microsoft Research","keywords":"Collocation (remote sensing); Computer science; Window (computing); Information retrieval; Mathematics; Machine learning","authors":[{"name":"Olga Vechtomova","is_ca":true},{"name":"Stephen Robertson","is_ca":false},{"name":"Susan Jones","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.009880289279499396,"gpt":0.2239308871336277,"spread":0.2140505978541283,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00156122,0.0007202252,0.00101458,0.002013361,0.0008203689,0.0008628581,0.001007551,0.0009309647,0.008164558],"category_scores_gemma":[0.01240056,0.0004386032,0.0005324911,0.003224327,0.0005407943,0.004916726,0.001688348,0.000855206,0.00231876],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000580765,"about_ca_system_score_gemma":0.0009312957,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00505371,"about_ca_topic_score_gemma":0.007730771,"domain_scores_codex":[0.9982452,0.0006771127,0.0001270224,0.0003484123,0.0004503863,0.0001519191],"domain_scores_gemma":[0.9938959,0.003472231,0.0002398094,0.001261187,0.0009695214,0.0001612446],"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":[0.003545396,0.001580514,0.009916513,0.001009787,0.0002884629,0.001171203,0.001534068,0.08860492,0.09957104,0.02613287,0.04911566,0.7175297],"study_design_scores_gemma":[0.0002158767,0.0005539703,0.004741539,0.00004920645,0.0001997677,0.0009767666,0.0005398252,0.9217857,0.0285642,0.02473149,0.01754844,0.00009317265],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.3972259,0.004419899,0.5721728,0.001287634,0.0003545566,0.0005837611,0.002123321,0.006577665,0.01525451],"genre_scores_gemma":[0.8546096,0.000704416,0.1340604,0.0002469746,0.0001758083,0.0002239484,0.002663836,0.0003816099,0.006933384],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008164558,"threshold_uncertainty_score":0.02731317,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3196754070","doi":"10.1007/s10791-022-09411-0","title":"Shallow pooling for sparse labels","year":2022,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":42,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Ranking (information retrieval); Mean reciprocal rank; Information retrieval; Pooling; Computer science; Relevance (law); Set (abstract data type); Rank (graph theory); Learning to rank; Preference; Artificial intelligence; Statistics; Mathematics; Combinatorics","authors":[{"name":"Negar Arabzadeh","is_ca":true},{"name":"Alexandra Vtyurina","is_ca":true},{"name":"Xinyi Yan","is_ca":true},{"name":"Charles L. A. Clarke","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02332481057054104,"gpt":0.2593679142488165,"spread":0.2360431036782755,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001719401,0.001432507,0.002121566,0.001166741,0.0007461854,0.001815102,0.002124665,0.002058574,0.008960826],"category_scores_gemma":[0.006325822,0.0008750011,0.00138168,0.001576713,0.001129561,0.004686048,0.00298583,0.002314791,0.002275381],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001455712,"about_ca_system_score_gemma":0.001562808,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009584728,"about_ca_topic_score_gemma":0.01471811,"domain_scores_codex":[0.9992208,0.000184027,0.0000481915,0.0002270214,0.0001525008,0.0001673862],"domain_scores_gemma":[0.99806,0.0009865154,0.0001309907,0.0005402643,0.000179737,0.0001024108],"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":[0.0007602893,0.0003662741,0.001364476,0.0005191336,0.0002626531,0.0002362693,0.000255457,0.1290361,0.02177069,0.1074845,0.03245687,0.7054873],"study_design_scores_gemma":[0.00002883352,0.00006922768,0.0004070293,0.0000312474,0.00005796073,0.00004178239,0.00002636457,0.8792862,0.004380703,0.1134881,0.002160218,0.00002219116],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01883929,0.0008545947,0.973186,0.0006998106,0.00009826892,0.00007427199,0.0008081374,0.002788971,0.002650663],"genre_scores_gemma":[0.5975841,0.001203764,0.3711297,0.0007501825,0.0004374877,0.0003437189,0.005846831,0.0008018462,0.02190244],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009584728,"threshold_uncertainty_score":0.0299769,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2967813257","doi":"10.1007/s10791-019-09361-0","title":"Evaluating sentence-level relevance feedback for high-recall information retrieval","year":2019,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Topic Modeling","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Relevance feedback; Relevance (law); Recall; Computer science; Sentence; Baseline (sea); Information retrieval; Precision and recall; Natural language processing; Artificial intelligence; Cognitive psychology; Psychology; Image retrieval","authors":[{"name":"Haotian Zhang","is_ca":true},{"name":"Gordon V. Cormack","is_ca":true},{"name":"Maura R. Grossman","is_ca":true},{"name":"Mark D. Smucker","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04688903707632153,"gpt":0.292025163440755,"spread":0.2451361263644335,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005965294,0.001196349,0.001583896,0.001913242,0.0005973083,0.001288273,0.0008333756,0.001868367,0.003316588],"category_scores_gemma":[0.02758873,0.0003225321,0.0006906472,0.0007258973,0.0003188745,0.001566517,0.0007450907,0.0008624804,0.001549442],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007001727,"about_ca_system_score_gemma":0.001136423,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003736874,"about_ca_topic_score_gemma":0.004950128,"domain_scores_codex":[0.9964738,0.00177839,0.0002779076,0.0003779434,0.0009265538,0.0001653156],"domain_scores_gemma":[0.9842606,0.01257447,0.0004635515,0.0003900566,0.002015383,0.0002958674],"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":[0.01054371,0.002388965,0.009490937,0.002628943,0.0009281894,0.0004327959,0.0005421565,0.04405293,0.1439818,0.001094848,0.01943057,0.7644842],"study_design_scores_gemma":[0.00087746,0.004431417,0.01835079,0.0001107224,0.001201183,0.0005673236,0.0002498139,0.9054238,0.06366134,0.00205599,0.00294272,0.0001273288],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7510055,0.0154103,0.2187283,0.0008884288,0.0008114804,0.0008568126,0.001380443,0.005765983,0.005152695],"genre_scores_gemma":[0.9170604,0.001111057,0.07484335,0.0002256332,0.0004723941,0.0001908092,0.002711774,0.0001971688,0.003187427],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005965294,"threshold_uncertainty_score":0.03154784,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2556771250","doi":"10.1007/s10791-016-9290-6","title":"Efficient distributed selective search","year":2016,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Directorate for Computer and Information Science and Engineering; Australian Research Council; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Ranking (information retrieval); Resource (disambiguation); Index (typography); Mirroring; Data mining; Selection (genetic algorithm); Web search query; Query expansion; Search engine; Information retrieval; Distributed computing; Machine learning; Computer network; World Wide Web","authors":[{"name":"Yubin Kim","is_ca":false},{"name":"Jamie Callan","is_ca":false},{"name":"J. Shane Culpepper","is_ca":false},{"name":"Alistair Moffat","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01353276958545393,"gpt":0.2529916586149543,"spread":0.2394588890295003,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009818925,0.0006345296,0.001392447,0.001145976,0.0009627509,0.001400744,0.001720053,0.0008989691,0.009528766],"category_scores_gemma":[0.004595172,0.000353717,0.0004040439,0.001752265,0.0006371703,0.002353519,0.001781361,0.000632657,0.002903014],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009296191,"about_ca_system_score_gemma":0.001959297,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002521502,"about_ca_topic_score_gemma":0.004615814,"domain_scores_codex":[0.9987881,0.0002739257,0.00004945858,0.0002047647,0.0004768264,0.0002069874],"domain_scores_gemma":[0.9970944,0.001058843,0.0001276006,0.001100201,0.0004926331,0.0001262594],"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":[0.002494303,0.001006585,0.003473057,0.0003973843,0.0001432748,0.0002643205,0.0002643452,0.1507904,0.04555923,0.07728012,0.03810048,0.6802265],"study_design_scores_gemma":[0.0001360585,0.0001056069,0.0007408044,0.000007698695,0.00005070572,0.0001860867,0.00007037853,0.9465881,0.007496448,0.0404508,0.004153468,0.00001372164],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1554367,0.001362346,0.8065796,0.0007676325,0.0001580933,0.0002428078,0.0006189715,0.004950657,0.02988324],"genre_scores_gemma":[0.8382453,0.0003547582,0.140838,0.0001399484,0.0001405254,0.0001811067,0.0006968207,0.0002506225,0.01915284],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009528766,"threshold_uncertainty_score":0.03187686,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2082055394","doi":"10.1007/s10791-012-9218-8","title":"Increasing evaluation sensitivity to diversity","year":2013,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"National Science Foundation","keywords":"Computer science; Diversity (politics); Context (archaeology); Information retrieval; Discriminative model; Measure (data warehouse); Data science; Data mining; Artificial intelligence; Geography","authors":[{"name":"Peter B. Golbus","is_ca":false},{"name":"Javed A. Aslam","is_ca":false},{"name":"Charles L. A. Clarke","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02380544427014461,"gpt":0.2593833951838307,"spread":0.2355779509136861,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03186589,0.0007833574,0.001394857,0.002833553,0.0009257513,0.004663874,0.001071175,0.002570878,0.003536755],"category_scores_gemma":[0.2705443,0.0007639112,0.0007706834,0.001766243,0.001700003,0.007153212,0.004030129,0.00277407,0.0006886284],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002269134,"about_ca_system_score_gemma":0.001143727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00133173,"about_ca_topic_score_gemma":0.0008784998,"domain_scores_codex":[0.9497899,0.03108793,0.00242184,0.00580089,0.00940295,0.001496581],"domain_scores_gemma":[0.5014905,0.427412,0.01602565,0.03529934,0.01712867,0.002643789],"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":[0.006039691,0.001363112,0.1730998,0.001315174,0.001667957,0.0007646561,0.004143725,0.1022283,0.1067958,0.04872802,0.006792291,0.5470616],"study_design_scores_gemma":[0.0004234088,0.003205668,0.1506961,0.0004257455,0.001517262,0.00289762,0.001958851,0.5333737,0.09894913,0.1924105,0.01372273,0.000419242],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8302637,0.002683844,0.1340203,0.003195961,0.0002006663,0.0002245028,0.0002306398,0.001389529,0.02779097],"genre_scores_gemma":[0.9887512,0.0001527054,0.009589763,0.000322578,0.0000856514,0.00002820682,0.00006204025,0.0001001476,0.0009077214],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03186589,"threshold_uncertainty_score":0.1685249,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2850470857","doi":"10.1007/s10791-018-9337-y","title":"User interest prediction over future unobserved topics on social networks","year":2018,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Computer science; Focus (optics); Work (physics); User modeling; Data science; Social network (sociolinguistics); World Wide Web; Social media; User interface; Engineering","authors":[{"name":"Fattane Zarrinkalam","is_ca":true},{"name":"Mohsen Kahani","is_ca":false},{"name":"Ebrahim Bagheri","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02993305481423278,"gpt":0.2657125739616944,"spread":0.2357795191474616,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001338331,0.0006047754,0.000764186,0.002261897,0.0003445786,0.001098142,0.0008394047,0.0009796218,0.001153669],"category_scores_gemma":[0.005838231,0.0003540031,0.0007518898,0.001645691,0.0002711148,0.002414616,0.0005002736,0.001150596,0.0007857906],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005315804,"about_ca_system_score_gemma":0.0003148853,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005740302,"about_ca_topic_score_gemma":0.01085682,"domain_scores_codex":[0.9993617,0.0002100858,0.00004201749,0.0001882621,0.000136373,0.00006155601],"domain_scores_gemma":[0.9951997,0.003594409,0.0003026162,0.0003471036,0.0004253571,0.0001307987],"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":[0.002137733,0.001551974,0.2772848,0.000577836,0.001091577,0.0006389449,0.0009321478,0.3037501,0.02047477,0.01348303,0.01139962,0.3666774],"study_design_scores_gemma":[0.00001489191,0.00009470178,0.01412945,0.00001472782,0.0001112858,0.00009010141,0.000059874,0.9793241,0.00159872,0.003781282,0.0007675905,0.00001332056],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6385431,0.002383777,0.3510517,0.001396621,0.0001280134,0.0001169356,0.002349428,0.0005703279,0.00346016],"genre_scores_gemma":[0.9807387,0.0006331287,0.014616,0.00005073599,0.0001812506,0.0000366144,0.00156243,0.0000199126,0.002161236],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005740302,"threshold_uncertainty_score":0.01141381,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3044191208","doi":"10.1007/s10791-020-09379-9","title":"Robust keyword search in large attributed graphs","year":2020,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":20,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"IBM (Canada); Toronto Metropolitan University; University of Waterloo; Ontario Tech University","funders":"","keywords":"Scalability; Computer science; SPARK (programming language); Theoretical computer science; Set (abstract data type); Graph; Keyword search; Data mining; Information retrieval; Database","authors":[{"name":"Spencer Bryson","is_ca":true},{"name":"Heidar Davoudi","is_ca":true},{"name":"Lukasz Golab","is_ca":true},{"name":"Mehdi Kargar","is_ca":true},{"name":"Yuliya Lytvyn","is_ca":true},{"name":"Piotr Mierzejewski","is_ca":true},{"name":"Jaroslaw Szlichta","is_ca":true},{"name":"Morteza Zihayat","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0316519411098354,"gpt":0.2600099557485858,"spread":0.2283580146387504,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002763451,0.001073884,0.002522538,0.006952778,0.001112939,0.002799647,0.002750269,0.001637917,0.002488917],"category_scores_gemma":[0.02370071,0.001058736,0.001241565,0.00818918,0.001159131,0.006580669,0.002241346,0.001455188,0.00107456],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001981282,"about_ca_system_score_gemma":0.001966516,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01269522,"about_ca_topic_score_gemma":0.01477791,"domain_scores_codex":[0.9972365,0.0008453807,0.0002265697,0.0006501672,0.0007953383,0.0002460728],"domain_scores_gemma":[0.9788225,0.01590914,0.001738935,0.002050733,0.001108231,0.0003703826],"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":[0.0007120758,0.0004602531,0.00491163,0.001083645,0.000355848,0.0004036022,0.0003176144,0.6645991,0.009484603,0.04619632,0.01294967,0.2585257],"study_design_scores_gemma":[0.00001898809,0.00002332018,0.0003583156,0.00001347202,0.00003850962,0.00005002448,0.00005216168,0.9507022,0.001492051,0.04642257,0.0008177201,0.00001070116],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06803502,0.001873433,0.9231253,0.0008518716,0.00006954399,0.0001602971,0.002129813,0.002033773,0.001720968],"genre_scores_gemma":[0.7359213,0.001839236,0.247955,0.0002980134,0.0002865305,0.0001776137,0.005533916,0.0004597362,0.007528734],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01269522,"threshold_uncertainty_score":0.02524269,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2056081579","doi":"10.1007/s10791-007-9042-8","title":"Hybrid index maintenance for contiguous inverted lists","year":2008,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"University of Waterloo","keywords":"Merge (version control); Computer science; Inverted index; Information retrieval; Data mining; Search engine indexing","authors":[{"name":"Stefan Büttcher","is_ca":false},{"name":"Charles L. A. Clarke","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01558255272543664,"gpt":0.2299373068822662,"spread":0.2143547541568295,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001445964,0.0005462507,0.001244085,0.003879826,0.001329549,0.00229771,0.003012614,0.0009342733,0.004715884],"category_scores_gemma":[0.007664511,0.0006331683,0.0005964817,0.005781825,0.0006857652,0.005239277,0.002228166,0.0008422362,0.001847503],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007723093,"about_ca_system_score_gemma":0.001501297,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003435718,"about_ca_topic_score_gemma":0.007006031,"domain_scores_codex":[0.9984607,0.0002131002,0.000204544,0.0002452738,0.0007285208,0.0001479929],"domain_scores_gemma":[0.9921969,0.001839405,0.0004786143,0.003597821,0.00171803,0.0001693323],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0009543591,0.0003333837,0.0029622,0.000394938,0.0001456444,0.0002097547,0.0002762909,0.01835036,0.03390261,0.01476685,0.02139299,0.9063106],"study_design_scores_gemma":[0.0004643917,0.001104393,0.004955783,0.0001570614,0.0004976747,0.00191794,0.0004460004,0.7295048,0.1318735,0.08967099,0.03919877,0.0002086401],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1154493,0.003366579,0.8596303,0.0004731377,0.0003775763,0.0003572772,0.001912582,0.01170938,0.006723809],"genre_scores_gemma":[0.3398567,0.000687284,0.6459939,0.0002603591,0.0003099726,0.0003014631,0.004051459,0.0008863189,0.007652565],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004715884,"threshold_uncertainty_score":0.01577622,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2576646839","doi":"10.1007/s10791-016-9292-4","title":"Enhancing click models with mouse movement information","year":2017,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Relevance (law); Computer science; Information retrieval; Search engine; Process (computing); Movement (music); Artificial intelligence; World Wide Web; Human–computer interaction; Programming language","authors":[{"name":"Zeyang Liu","is_ca":false},{"name":"Jiaxin Mao","is_ca":false},{"name":"Chao Wang","is_ca":false},{"name":"Qingyao Ai","is_ca":false},{"name":"Yiqun Liu","is_ca":false},{"name":"Jian‐Yun Nie","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01890283138193856,"gpt":0.2518378594100482,"spread":0.2329350280281096,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000855926,0.001022135,0.0007125706,0.0007762575,0.0001994842,0.001112557,0.0008591985,0.001190338,0.005631525],"category_scores_gemma":[0.008251334,0.0003115177,0.000524937,0.0006059538,0.0002804932,0.002572003,0.0007720094,0.0007977426,0.001663969],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005704832,"about_ca_system_score_gemma":0.0006215954,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003123349,"about_ca_topic_score_gemma":0.004271904,"domain_scores_codex":[0.9995083,0.0001708075,0.00002482256,0.00007829734,0.0001633861,0.00005449936],"domain_scores_gemma":[0.9963174,0.002428164,0.000247007,0.0003950546,0.0004395333,0.0001727367],"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":[0.003927821,0.002944244,0.01136482,0.0006284298,0.0001892537,0.0002996876,0.0003047657,0.4080191,0.1916747,0.02074739,0.009900971,0.3499988],"study_design_scores_gemma":[0.00005356316,0.0002740346,0.0009497042,0.00001425204,0.00004767306,0.00005467052,0.00001610003,0.9833791,0.01148063,0.002881631,0.0008281659,0.00002060089],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2923671,0.0009142594,0.689815,0.000570634,0.0001595352,0.0001479165,0.0004351464,0.00666689,0.008923553],"genre_scores_gemma":[0.9427782,0.0003607825,0.05044261,0.0001128342,0.00004123841,0.00007968053,0.0003081202,0.0003885909,0.005487863],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005631525,"threshold_uncertainty_score":0.01883936,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4387453959","doi":"10.1007/s10791-023-09421-6","title":"Learning heterogeneous subgraph representations for team discovery","year":2023,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Guelph; York University; Toronto Metropolitan University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Overfitting; Ranking (information retrieval); Machine learning; Set (abstract data type); Graph; Task (project management); Artificial intelligence; Baseline (sea); Representation (politics); Data science; Artificial neural network; Theoretical computer science","authors":[{"name":"Radin Hamidi Rad","is_ca":true},{"name":"Hoang Nguyen","is_ca":true},{"name":"Feras Al‐Obeidat","is_ca":false},{"name":"Ebrahim Bagheri","is_ca":true},{"name":"Mehdi Kargar","is_ca":true},{"name":"Divesh Srivastava","is_ca":false},{"name":"Jaroslaw Szlichta","is_ca":true},{"name":"Fattane Zarrinkalam","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01446712919078797,"gpt":0.2738418785845982,"spread":0.2593747493938103,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000920961,0.0009869824,0.001369966,0.004785508,0.001030824,0.001564969,0.002209328,0.002163976,0.003712635],"category_scores_gemma":[0.006646629,0.0005319881,0.001502816,0.005129785,0.0007665528,0.003887212,0.001990564,0.001816073,0.001311781],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001267835,"about_ca_system_score_gemma":0.001338967,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007990863,"about_ca_topic_score_gemma":0.01630181,"domain_scores_codex":[0.9991055,0.000259794,0.00004908142,0.0003065098,0.0001591652,0.0001200107],"domain_scores_gemma":[0.9977761,0.001068959,0.000244702,0.000508157,0.0002300791,0.000172108],"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":[0.0008524562,0.001054961,0.01047769,0.0006929187,0.0006412416,0.0004404584,0.0006172446,0.2771743,0.007795522,0.06788401,0.05477018,0.577599],"study_design_scores_gemma":[0.00004078356,0.00005533857,0.000611565,0.0000406615,0.00008859437,0.00006318796,0.0001306556,0.8740997,0.001163358,0.1210531,0.002637853,0.00001518969],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1024408,0.001513328,0.8831118,0.001216534,0.0001565698,0.0002349013,0.003352443,0.003290899,0.004682787],"genre_scores_gemma":[0.7187168,0.0008339609,0.2631214,0.0004138146,0.0002112718,0.0002890989,0.01056242,0.0003681094,0.005483144],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007990863,"threshold_uncertainty_score":0.01588869,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2121315872","doi":"10.1007/s10791-013-9230-7","title":"Discover hidden web properties by random walk on bipartite graph","year":2013,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada; State Key Laboratory of Novel Software Technology; Nanjing University","keywords":"Zipf's law; Crawling; Simple random sample; Random walk; Sampling (signal processing); Bipartite graph; Statistics; Computer science; Sample size determination; Sample (material); Bernoulli's principle; Population; Mathematics; Graph; Theoretical computer science","authors":[{"name":"Yan Wang","is_ca":false},{"name":"Jie Liang","is_ca":false},{"name":"Jianguo Lü","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.008604713258223986,"gpt":0.189222311925318,"spread":0.180617598667094,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006936954,0.0006015337,0.001191527,0.00500939,0.0006803951,0.001262102,0.0009171172,0.001476467,0.001332494],"category_scores_gemma":[0.006454105,0.0007158424,0.001027045,0.002416113,0.000810214,0.003526289,0.000926211,0.001109065,0.0005484079],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006540375,"about_ca_system_score_gemma":0.0005645808,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003279721,"about_ca_topic_score_gemma":0.005123858,"domain_scores_codex":[0.9993382,0.0002151849,0.00003314131,0.0001819096,0.0001361052,0.00009545631],"domain_scores_gemma":[0.993468,0.004428194,0.0007872588,0.000676261,0.0003815564,0.000258807],"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":[0.002032819,0.001571557,0.113534,0.001432239,0.001104718,0.00183215,0.0009285028,0.4128532,0.05285919,0.1325716,0.01901474,0.2602652],"study_design_scores_gemma":[0.00002657625,0.00004502894,0.002637314,0.00001446041,0.00004700684,0.0001109718,0.00004997443,0.9417186,0.001270377,0.05362458,0.0004420239,0.00001308435],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5718945,0.0008809852,0.41945,0.0008579242,0.00004232299,0.0001183634,0.001983277,0.002010462,0.002762224],"genre_scores_gemma":[0.9553607,0.0003024278,0.04074347,0.0001050237,0.00006371925,0.00004821331,0.001919499,0.0001121971,0.001344797],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.00500939,"threshold_uncertainty_score":0.006521225,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2074280490","doi":"10.1007/s10791-011-9168-6","title":"A study of the integration of passage-, document-, and cluster-based information for re-ranking search results","year":2011,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Israel Science Foundation; McGill University","keywords":"Ranking (information retrieval); Computer science; Information retrieval; Context (archaeology); Representation (politics); Query expansion; Task (project management); Cluster (spacecraft); Document retrieval; Data mining; Information integration; Geography","authors":[{"name":"Eyal Krikon","is_ca":false},{"name":"Oren Kurland","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05817615587411926,"gpt":0.295348948589657,"spread":0.2371727927155378,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006305924,0.0005698484,0.001215488,0.005196036,0.0008662579,0.002647455,0.001329419,0.0009063322,0.001726457],"category_scores_gemma":[0.05731126,0.0004386837,0.0008618647,0.007775557,0.0009791002,0.005612029,0.0006633564,0.00117376,0.0003250126],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001678835,"about_ca_system_score_gemma":0.001574199,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01719508,"about_ca_topic_score_gemma":0.0158172,"domain_scores_codex":[0.9950857,0.002605007,0.0002217867,0.0004329934,0.001428387,0.0002261803],"domain_scores_gemma":[0.9294398,0.05667843,0.002283234,0.003627575,0.006995222,0.0009757788],"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":[0.004022839,0.00254097,0.06229819,0.001270402,0.001333333,0.0003079995,0.003954783,0.04607515,0.03499655,0.04167482,0.006326998,0.7951979],"study_design_scores_gemma":[0.000161238,0.001762359,0.05884199,0.0000960419,0.001091208,0.0004044502,0.0009706925,0.906197,0.01370009,0.01172397,0.004868258,0.0001826247],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.765665,0.00486557,0.2151009,0.0008940519,0.0001067169,0.0007048118,0.0005224843,0.001188644,0.01095192],"genre_scores_gemma":[0.8924393,0.0005293049,0.1046824,0.000055436,0.00007401819,0.00009423339,0.000355192,0.0001613888,0.001608752],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01719508,"threshold_uncertainty_score":0.03419,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2976855657","doi":"10.1007/s10791-019-09364-x","title":"ReBoost: a retrieval-boosted sequence-to-sequence model for neural response generation","year":2019,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Topic Modeling","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"National Key Research and Development Program of China; National Natural Science Foundation of China","keywords":"Computer science; Benchmark (surveying); Conversation; Sequence (biology); Process (computing); Artificial intelligence; Natural language generation; Artificial neural network; Language model; Natural language processing; Recurrent neural network; Machine learning; Natural language; Programming language; Linguistics","authors":[{"name":"Yutao Zhu","is_ca":false},{"name":"Zhicheng Dou","is_ca":false},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Ji-Rong Wen","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06926171490965688,"gpt":0.2995985784680832,"spread":0.2303368635584264,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002331715,0.00156461,0.002680978,0.001497456,0.0006504215,0.0009734162,0.004262301,0.002923491,0.004797721],"category_scores_gemma":[0.004123311,0.0007834593,0.001474213,0.001730412,0.0005806806,0.001667576,0.001369275,0.003364769,0.002956673],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001098784,"about_ca_system_score_gemma":0.001641656,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01168498,"about_ca_topic_score_gemma":0.01393623,"domain_scores_codex":[0.9989858,0.0003609039,0.00005319845,0.0002519102,0.000206764,0.0001413754],"domain_scores_gemma":[0.9986563,0.0007063322,0.00007757792,0.000123855,0.000354867,0.00008095574],"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":[0.0009091464,0.0006341359,0.0007158202,0.0003633931,0.0003357559,0.0001176212,0.00008754877,0.2835102,0.01213226,0.005943159,0.02459648,0.6706545],"study_design_scores_gemma":[0.00002182311,0.00005285735,0.00008452019,0.000008001718,0.00001748038,0.0000187151,0.000004084248,0.9949279,0.001774216,0.00224053,0.0008398411,0.00001005415],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01248478,0.001670672,0.9764766,0.0004027088,0.0002771945,0.0001841589,0.0004784975,0.007150414,0.000874906],"genre_scores_gemma":[0.3098129,0.001253056,0.6651555,0.001105819,0.0005669449,0.0009375329,0.003264374,0.001276681,0.01662724],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01168498,"threshold_uncertainty_score":0.02323395,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2008283248","doi":"10.1007/s10791-013-9220-9","title":"Latent word context model for information retrieval","year":2013,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Topic Modeling","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"Ministry of Education, Culture, Sports, Science and Technology","keywords":"Computer science; Latent Dirichlet allocation; Word (group theory); Context (archaeology); Information retrieval; Natural language processing; Latent semantic analysis; Search engine indexing; Topic model; Relevance (law); Artificial intelligence; Probabilistic latent semantic analysis; Linguistics","authors":[{"name":"Bernard Brosseau-Villeneuve","is_ca":true},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Noriko Kando","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02483661079381079,"gpt":0.2376980274101974,"spread":0.2128614166163866,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002898684,0.001055283,0.002214256,0.002520422,0.0007558711,0.002495525,0.002451991,0.002219052,0.004332417],"category_scores_gemma":[0.01140179,0.0007727565,0.001673555,0.004354769,0.0009203355,0.005871985,0.001292991,0.002937991,0.00314076],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001409254,"about_ca_system_score_gemma":0.001689481,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007517935,"about_ca_topic_score_gemma":0.007526561,"domain_scores_codex":[0.997404,0.001400588,0.0001720108,0.0004671043,0.0003747321,0.0001814073],"domain_scores_gemma":[0.9961691,0.002651772,0.0002380995,0.0004703472,0.0003890096,0.00008161241],"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":[0.0008004137,0.00050865,0.003240429,0.001270995,0.0006560082,0.0003771543,0.0007813254,0.2073346,0.007605848,0.3930781,0.02753184,0.3568146],"study_design_scores_gemma":[0.00004969746,0.00005722413,0.0006815299,0.00005740674,0.0001357843,0.0001027398,0.00005152856,0.800754,0.0008455367,0.1910451,0.006172834,0.00004664924],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01018353,0.0063549,0.9784642,0.0009568263,0.0002023252,0.00008465172,0.0009955954,0.001105146,0.001652738],"genre_scores_gemma":[0.5755696,0.009994214,0.3852363,0.0007698655,0.001287237,0.001059999,0.006459353,0.0006990386,0.01892441],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007517935,"threshold_uncertainty_score":0.01532984,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2062626555","doi":"10.1007/s10791-009-9105-0","title":"Swapping documents and terms","year":2009,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Agricultural Research Development Agency","keywords":"Relevance feedback; Computer science; Information retrieval; Relevance (law); Query expansion; Data mining; Artificial intelligence; Image retrieval","authors":[{"name":"Charles L. A. Clarke","is_ca":true},{"name":"Gordon V. Cormack","is_ca":true},{"name":"Thomas R. Lynam","is_ca":true},{"name":"Chris Buckley","is_ca":false},{"name":"Donna Harman","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.009581800700385145,"gpt":0.2530781925278839,"spread":0.2434963918274988,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002680653,0.00066347,0.0009477152,0.002735638,0.000890645,0.003296962,0.001478037,0.001267794,0.02137091],"category_scores_gemma":[0.02593761,0.000512286,0.0008229727,0.004271405,0.001002904,0.006873904,0.002299457,0.001267288,0.008488594],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000513286,"about_ca_system_score_gemma":0.001160639,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009419817,"about_ca_topic_score_gemma":0.0009291171,"domain_scores_codex":[0.9963199,0.001280088,0.0002848349,0.0006523844,0.001145825,0.0003170098],"domain_scores_gemma":[0.9886584,0.004642736,0.0003853435,0.004230336,0.001660075,0.0004231975],"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":[0.002706379,0.0009855072,0.008662329,0.0006480611,0.0001794524,0.0003377571,0.001124324,0.003946127,0.1050975,0.03756563,0.01612337,0.8226236],"study_design_scores_gemma":[0.0007014805,0.002488877,0.02795899,0.0002971346,0.001100969,0.005070894,0.003291498,0.135879,0.3973177,0.2002103,0.22529,0.0003931457],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6505876,0.00506926,0.2637179,0.00303477,0.002099838,0.001109968,0.004336834,0.009796022,0.06024775],"genre_scores_gemma":[0.8401654,0.001432168,0.1236607,0.0006602156,0.000320081,0.0002839027,0.003221756,0.00109768,0.02915794],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02137091,"threshold_uncertainty_score":0.07149285,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2069744915","doi":"10.1007/s10791-012-9192-1","title":"Extended structural relevance framework: a framework for evaluating structured document retrieval","year":2012,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Relevance (law); Information retrieval; Redundancy (engineering); Probabilistic logic; Task (project management); Tree (set theory); Data mining; Range (aeronautics); Artificial intelligence; Mathematics","authors":[{"name":"M. S. Ali","is_ca":true},{"name":"Mariano P. Consens","is_ca":true},{"name":"Mounia Lalmas","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03216634289591799,"gpt":0.3490464434591703,"spread":0.3168801005632523,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01427795,0.001337994,0.002301575,0.008555544,0.001251196,0.00348423,0.002966496,0.002008161,0.005250033],"category_scores_gemma":[0.04348088,0.0005958507,0.001605282,0.004239445,0.001523536,0.006541034,0.002080584,0.001404876,0.001459181],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001906303,"about_ca_system_score_gemma":0.00327874,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005485985,"about_ca_topic_score_gemma":0.005884741,"domain_scores_codex":[0.9888593,0.004572549,0.0006744446,0.0008858722,0.004588476,0.0004193799],"domain_scores_gemma":[0.9842103,0.007205093,0.00132166,0.002326783,0.004501326,0.0004347631],"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":[0.001219705,0.0007704065,0.01014133,0.00144734,0.0006222494,0.0002600847,0.0007208939,0.08017334,0.02580173,0.1463499,0.0116886,0.7208043],"study_design_scores_gemma":[0.0001632136,0.001178319,0.005173422,0.0001562805,0.0003874185,0.0004457876,0.0001821104,0.8297148,0.01071645,0.1429529,0.008770803,0.0001585047],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0257211,0.002349278,0.9655446,0.0003368587,0.00007214854,0.0005911113,0.0007030567,0.001626539,0.003055297],"genre_scores_gemma":[0.4272656,0.0009449161,0.5663996,0.000131677,0.0002631151,0.0007178814,0.001233058,0.0002882862,0.002755886],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01427795,"threshold_uncertainty_score":0.07550991,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3133365413","doi":"10.1007/s10791-021-09398-0","title":"Neural ranking models for document retrieval","year":2021,"lang":"en","type":"preprint","venue":"Information Retrieval","topic":"Multimodal Machine Learning Applications","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":"Instituto de Ciencias del Mar y Limnología, Universidad Nacional Autónoma de México; National Institute of Standards and Technology; Institute for Catastrophic Loss Reduction; National Science Foundation","keywords":"Ranking (information retrieval); Computer science; Artificial intelligence; Machine learning; Set (abstract data type); Variety (cybernetics); Information retrieval; Artificial neural network; Deep learning; Learning to rank; Data mining","authors":[{"name":"Mohamed Trabelsi","is_ca":false},{"name":"Zhiyu Chen","is_ca":false},{"name":"Brian D. Davison","is_ca":false},{"name":"Jeff Heflin","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02256006568427007,"gpt":0.2916584035119547,"spread":0.2690983378276846,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00205581,0.00080945,0.001360283,0.00191371,0.0004486415,0.002023888,0.001612561,0.001546375,0.008365574],"category_scores_gemma":[0.006747213,0.0003773889,0.0008614809,0.002097969,0.0006518019,0.002760341,0.0006714023,0.001469502,0.002788474],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001986553,"about_ca_system_score_gemma":0.0007170533,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007438149,"about_ca_topic_score_gemma":0.007762405,"domain_scores_codex":[0.9992176,0.0002971012,0.00005252947,0.0001421329,0.0001782882,0.0001123221],"domain_scores_gemma":[0.9978795,0.001118168,0.0002430325,0.0001667478,0.0005074739,0.00008515419],"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":[0.0002744056,0.0002551485,0.00171883,0.0003829457,0.0001528416,0.0001454743,0.0001144791,0.7136781,0.002390051,0.1080023,0.0185633,0.1543222],"study_design_scores_gemma":[0.00001014955,0.0000213349,0.00019925,0.00001106528,0.00001108525,0.00001740789,0.000006081404,0.9778211,0.0001553983,0.02091975,0.0008187542,0.000008682146],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1059069,0.01228139,0.8390996,0.004325443,0.000546144,0.0002531957,0.001567553,0.00343275,0.03258703],"genre_scores_gemma":[0.8747833,0.003506575,0.07991028,0.0005968413,0.0005817625,0.000323511,0.001483064,0.0002630899,0.03855155],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008365574,"threshold_uncertainty_score":0.02798563,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2584295907","doi":"10.1007/s10791-017-9294-x","title":"Constructing click models for search users","year":2017,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Image Retrieval and Classification Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; World Wide Web; Information retrieval","authors":[{"name":"Yiqun Liu","is_ca":false},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Yi Chang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05359408056046548,"gpt":0.3126871301685992,"spread":0.2590930496081337,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004057871,0.001397154,0.001734864,0.00457175,0.0007562934,0.003058692,0.002657227,0.003724395,0.006803075],"category_scores_gemma":[0.01944192,0.001149121,0.002714127,0.003568227,0.001214644,0.005983329,0.001404062,0.002852242,0.002964544],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00247909,"about_ca_system_score_gemma":0.001777798,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02195994,"about_ca_topic_score_gemma":0.02335894,"domain_scores_codex":[0.9983981,0.0006034033,0.00008284069,0.0003205653,0.0002652533,0.0003297559],"domain_scores_gemma":[0.9870195,0.00967071,0.0005737756,0.001177975,0.001056179,0.0005018943],"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":[0.003247175,0.002053069,0.08730829,0.0005271015,0.000658115,0.0007000459,0.001173352,0.5435538,0.007648187,0.116407,0.02659718,0.2101267],"study_design_scores_gemma":[0.00001905314,0.00004204756,0.001407067,0.00001343998,0.00004405824,0.00004024008,0.00004048965,0.986545,0.0005166799,0.01096444,0.0003514713,0.00001610168],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3632093,0.000887703,0.618085,0.002099141,0.0001485521,0.0003801711,0.005074852,0.003850567,0.006264576],"genre_scores_gemma":[0.9502027,0.000511642,0.03484334,0.0001737454,0.0001618748,0.0003527456,0.004239736,0.000343134,0.009171203],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02195994,"threshold_uncertainty_score":0.04366422,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2803782109","doi":"10.1007/s10791-018-9332-3","title":"(CF)2 architecture: contextual collaborative filtering","year":2018,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Recommender Systems and Techniques","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":"Oakville-Trafalgar Memorial Hospital; Western University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Collaborative filtering; Computer science; Recommender system; Context (archaeology); Architecture; The Internet; Selection (genetic algorithm); Fraction (chemistry); Machine learning; Scale (ratio); Artificial intelligence; Contextual design; Filter (signal processing); Information retrieval; Data science; World Wide Web; Human–computer interaction","authors":[{"name":"Dennis Bachmann","is_ca":true},{"name":"Katarina Grolinger","is_ca":true},{"name":"Hany F. ElYamany","is_ca":true},{"name":"Wilson A. Higashino","is_ca":true},{"name":"Miriam A. M. Capretz","is_ca":true},{"name":"Majid Fekri","is_ca":true},{"name":"B. Gopalakrishnan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01103614124488447,"gpt":0.2479555254115812,"spread":0.2369193841666967,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00123727,0.0008122571,0.0009036288,0.001122952,0.0008895458,0.001286911,0.002890309,0.002061972,0.009769381],"category_scores_gemma":[0.00328225,0.0004207328,0.0009970374,0.001225588,0.0005210144,0.001970321,0.001452053,0.001082415,0.004081005],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007906774,"about_ca_system_score_gemma":0.001604489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02847836,"about_ca_topic_score_gemma":0.03935464,"domain_scores_codex":[0.9992668,0.0001295599,0.00005049318,0.0003005393,0.0001448563,0.0001077123],"domain_scores_gemma":[0.9989371,0.0001903949,0.00004685094,0.0004046408,0.0003449329,0.00007606196],"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":[0.0008329091,0.0005125988,0.008012363,0.0002939363,0.0003183238,0.0002870731,0.0003367763,0.1162094,0.01647261,0.03064419,0.04630604,0.7797738],"study_design_scores_gemma":[0.00007077405,0.0002543027,0.001775165,0.00003490194,0.0001142708,0.0003189746,0.00005136945,0.9468513,0.009331486,0.01307649,0.0280397,0.00008118883],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01654332,0.0005290221,0.9586164,0.0004864862,0.000189076,0.0003161496,0.0007659702,0.01542887,0.007124633],"genre_scores_gemma":[0.3636149,0.0004129648,0.6181422,0.0006135899,0.0002306766,0.000524228,0.002043529,0.0002925986,0.01412523],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02847836,"threshold_uncertainty_score":0.05662519,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2000952055","doi":"10.1007/s10791-006-9397-2","title":"Introduction to the special issue on the 27th European Conference on Information Retrieval Research","year":2006,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"University of Waterloo","keywords":"Information retrieval; Political science; Library science; Data science; Computer science","authors":[{"name":"David E. Losada","is_ca":false},{"name":"Juan M. Fernández‐Luna","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03850458048473922,"gpt":0.2941051982642033,"spread":0.2556006177794641,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005975216,0.002858652,0.008110515,0.0099935,0.00274383,0.01268555,0.003116624,0.005433497,0.1946928],"category_scores_gemma":[0.01097606,0.001002239,0.002479225,0.005816402,0.001345844,0.009754769,0.004287725,0.006676254,0.1566085],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002583321,"about_ca_system_score_gemma":0.003249821,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001503992,"about_ca_topic_score_gemma":0.003562294,"domain_scores_codex":[0.9956245,0.0006161695,0.0004463394,0.0006036431,0.002205217,0.0005041853],"domain_scores_gemma":[0.9838923,0.003624508,0.0009606994,0.00136077,0.006895949,0.003265813],"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":[0.00004572573,0.00003445265,0.000082781,0.0001970312,0.0000139183,0.00002846647,0.00001244356,0.00003241692,0.0002183784,0.0005784182,0.9652282,0.03352786],"study_design_scores_gemma":[0.00001887601,0.0000593735,0.0006429012,0.0002267587,0.00003554677,0.0001376466,0.00003389785,0.0002041353,0.0001606419,0.001323429,0.9971329,0.00002393123],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.0006136336,0.08643523,0.008732948,0.03801353,0.8117113,0.0002604088,0.001213488,0.0009673828,0.0520521],"genre_scores_gemma":[0.003374342,0.06166267,0.004218692,0.01936284,0.5341825,0.0004097291,0.003383526,0.001911535,0.3714942],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.1946928,"threshold_uncertainty_score":0.6513122,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}