{"meta":{"query_hash":"d5798dda5751","filters":{"venue":"The Annals of Applied Statistics"},"cohort_total":71,"direct_labels_cover":0,"predictions_cover":71,"exported":71,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/d5798dda5751","api":"https://metacan.xera.ac/api/v1/cohort?venue=The+Annals+of+Applied+Statistics"},"results":[{"id":"W1965857117","doi":"10.1214/08-aoas199","title":"$\\mathcal{G}$-SELC: Optimization by sequential elimination of level combinations using genetic algorithms and Gaussian processes","year":2009,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"University of Georgia Research Foundation; Acadia University; University of Georgia; Pfizer; National Science Foundation","keywords":"Gaussian process; Set (abstract data type); Process (computing); Gaussian; Data set; Genetic algorithm","score_opus":0.10767218017510702,"score_gpt":0.36841160371865145,"score_spread":0.2607394235435444,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1965857117","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010527157,0.00018701903,0.987206,0.00021189022,0.000023471572,0.00007633025,0.00005631466,0.00066750724,0.0010442934],"genre_scores_gemma":[0.13437587,0.00019200752,0.8626105,0.00032619774,0.00003799087,0.00035723904,0.00026737375,0.00023542445,0.0015973299],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991358,0.00041966527,0.000039175924,0.00010827593,0.00021852576,0.00007849234],"domain_scores_gemma":[0.99759454,0.0018126034,0.00014212316,0.00014091471,0.00024584727,0.00006399227],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002720332,0.0013253341,0.0021569734,0.001981666,0.0006288673,0.00092278223,0.0020333582,0.0018877814,0.0023686702],"category_scores_gemma":[0.005050357,0.00071220874,0.0015706373,0.0024703778,0.0014692827,0.0010102695,0.0016638237,0.0019972848,0.0006906807],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016176405,0.00016744295,0.0015184244,0.00011597617,0.00016574362,0.00009665997,0.00008066606,0.76112443,0.0014535222,0.02019557,0.0034273695,0.2114924],"study_design_scores_gemma":[0.000027012764,0.000034291414,0.00008190379,0.0000046906107,0.000013097483,0.0000144732385,0.000004889698,0.9938822,0.00044436308,0.0050436324,0.00044427795,0.0000052297905],"about_ca_topic_score_codex":0.0074077137,"about_ca_topic_score_gemma":0.0080199335,"teacher_disagreement_score":0.0074077137,"about_ca_system_score_codex":0.0010801469,"about_ca_system_score_gemma":0.0024537514,"threshold_uncertainty_score":0.014729202},"labels":[],"label_agreement":null},{"id":"W1970646576","doi":"10.1214/14-aoas727","title":"Effect of breastfeeding on gastrointestinal infection in infants: A targeted maximum likelihood approach for clustered longitudinal data","year":2014,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":52,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; Université de Montréal","funders":"National Institute of Allergy and Infectious Diseases","keywords":"Breastfeeding; Confounding; Breastfeeding promotion; Context (archaeology); Estimation; Duration (music); Intervention (counseling); Random effects model; Breast feeding","score_opus":0.10728965630770694,"score_gpt":0.3946682202134613,"score_spread":0.2873785639057544,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1970646576","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.055247303,0.0014484203,0.9405357,0.0013872463,0.00005388005,0.0003143915,0.00024634547,0.0002659401,0.00050081],"genre_scores_gemma":[0.5162977,0.0011521222,0.478884,0.0006305516,0.000113532136,0.0012053809,0.0005516768,0.000113124464,0.0010518954],"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","domain_scores_codex":[0.9626595,0.034414463,0.00047608823,0.0015069982,0.0006737583,0.00026916285],"domain_scores_gemma":[0.8911803,0.100192904,0.0037737496,0.0032758557,0.001133097,0.0004440836],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.041447703,0.0010359362,0.0020013314,0.0016753769,0.0006879593,0.00124376,0.002661648,0.0020217032,0.0017955474],"category_scores_gemma":[0.10960792,0.00086697214,0.0031458372,0.001256725,0.0014030334,0.0012024117,0.0024896385,0.002330331,0.0001699108],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0041182474,0.0009275773,0.04145105,0.0022958405,0.006426323,0.0007151958,0.001392825,0.5474221,0.0022881902,0.14923654,0.002807466,0.24091871],"study_design_scores_gemma":[0.0004477698,0.0005678952,0.0052289297,0.0001592131,0.00063130003,0.00010043245,0.000095250994,0.919386,0.0008523506,0.071371004,0.0011125376,0.000047226287],"about_ca_topic_score_codex":0.0045160996,"about_ca_topic_score_gemma":0.004033878,"teacher_disagreement_score":0.041447703,"about_ca_system_score_codex":0.0013825676,"about_ca_system_score_gemma":0.0020622548,"threshold_uncertainty_score":0.219199},"labels":[],"label_agreement":null},{"id":"W1971541813","doi":"10.1214/10-aoas442","title":"Forecasting emergency medical service call arrival rates","year":2011,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Forecasting Techniques and Applications","field":"Decision Sciences","cited_by":123,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Computer science; Queueing theory; Context (archaeology); Exponential smoothing; Smoothing; Covariate; Econometrics; Mathematics; Machine learning; Geography","score_opus":0.5015733237736453,"score_gpt":0.4569802608908218,"score_spread":0.0445930628828235,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1971541813","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.083385736,0.00017682252,0.91445833,0.0002686071,0.00006651613,0.000031013336,0.00022169041,0.00045272527,0.00093844556],"genre_scores_gemma":[0.7775754,0.0002603489,0.2203349,0.00005833475,0.00009176853,0.00006247038,0.0005334692,0.00004933085,0.0010340683],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991116,0.00038784667,0.000046855104,0.00013927942,0.00025520893,0.00005917321],"domain_scores_gemma":[0.9970796,0.0017890173,0.00040307423,0.00027774338,0.00036809145,0.00008245379],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001964509,0.00048793637,0.00085087336,0.0008012544,0.00018178692,0.00082396617,0.0010495236,0.0007315808,0.00061428663],"category_scores_gemma":[0.009385795,0.00035855183,0.0005127943,0.0009343312,0.0002902829,0.0010342229,0.00049783476,0.0010483004,0.00016629245],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00005524134,0.000057300582,0.005007194,0.00002168696,0.000051617095,0.00002815286,0.000029467528,0.9527652,0.0012771516,0.0057401117,0.0004424528,0.034524404],"study_design_scores_gemma":[0.0000029229818,0.0000066963516,0.00032705514,0.0000012355833,0.0000027436886,0.0000033182173,0.0000018695061,0.998588,0.00013289874,0.0008171121,0.00011298402,0.0000031906754],"about_ca_topic_score_codex":0.012757452,"about_ca_topic_score_gemma":0.0099111,"teacher_disagreement_score":0.012757452,"about_ca_system_score_codex":0.00086436525,"about_ca_system_score_gemma":0.0010412933,"threshold_uncertainty_score":0.025366366},"labels":[],"label_agreement":null},{"id":"W1987751717","doi":"10.1214/10-aoas342","title":"A nonlinear mixed effects directional model for the estimation of the rotation axes of the human ankle","year":2010,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Morphological variations and asymmetry","field":"Mathematics","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Rotation (mathematics); Orientation (vector space); Nonlinear system; Measure (data warehouse); Monte Carlo method; Linear model; Sample (material); Data set","score_opus":0.07298192094243762,"score_gpt":0.35242304500087684,"score_spread":0.27944112405843924,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1987751717","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0035962819,0.0001410688,0.9953252,0.000094623254,0.000045249446,0.00009104777,0.0002342207,0.00009164153,0.00038060744],"genre_scores_gemma":[0.1731869,0.0008730247,0.81301445,0.00028177229,0.00021180314,0.0028377327,0.0017284332,0.00014942758,0.0077165416],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9935592,0.0039349813,0.00024120734,0.0013306763,0.000697147,0.00023680605],"domain_scores_gemma":[0.99401784,0.004098355,0.0005102934,0.0007372544,0.00054896466,0.000087348926],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010937813,0.0017654544,0.0017046438,0.0015291892,0.0005720666,0.0014089778,0.0037021488,0.0017850174,0.0043646297],"category_scores_gemma":[0.01876997,0.0009219047,0.0034059517,0.0015900757,0.0013444492,0.0016804745,0.001561892,0.0023253164,0.0017891343],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000818616,0.00024182258,0.008569648,0.0005404757,0.0008718586,0.00033468596,0.00068764645,0.49895874,0.0055812015,0.31855607,0.0033303173,0.16150902],"study_design_scores_gemma":[0.0000941101,0.0004587984,0.0025821766,0.000060069327,0.00018096413,0.00018851274,0.000052591742,0.8982132,0.0010851006,0.09002818,0.0069532017,0.000103111925],"about_ca_topic_score_codex":0.005125314,"about_ca_topic_score_gemma":0.006318588,"teacher_disagreement_score":0.010937813,"about_ca_system_score_codex":0.0011516628,"about_ca_system_score_gemma":0.0014613117,"threshold_uncertainty_score":0.057845354},"labels":[],"label_agreement":null},{"id":"W2012397316","doi":"10.1214/14-aoas774","title":"Evaluating epoetin dosing strategies using observational longitudinal data","year":2014,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Medicine; Observational study; Dosing; Regimen; Kidney disease; Intensive care medicine; Proportional hazards model; Dialysis; Anemia; Marginal structural model; Randomized controlled trial; Hemodialysis; Epoetin alfa; Clinical trial; Survival analysis; Internal medicine","score_opus":0.9581370777790906,"score_gpt":0.6916552922079258,"score_spread":0.26648178557116486,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2012397316","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.88609374,0.001471413,0.10748344,0.0013614804,0.000115742994,0.0007818246,0.0016175135,0.0001785185,0.0008963159],"genre_scores_gemma":[0.9690387,0.00040867593,0.02817684,0.00022431748,0.0000657598,0.0006416365,0.0011206401,0.000017906446,0.00030559412],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.933827,0.05932428,0.0019045097,0.0026054175,0.0016868516,0.00065200607],"domain_scores_gemma":[0.58508366,0.36640406,0.02612159,0.016653135,0.004075227,0.001662293],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.12217002,0.0010225582,0.0014020443,0.0013706823,0.0005965897,0.0014516244,0.0019504916,0.002007581,0.0014248699],"category_scores_gemma":[0.25398064,0.0006246307,0.0027710842,0.001528064,0.0013422641,0.0019150414,0.0017715422,0.0020067377,0.00023121196],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.012445979,0.0029350955,0.47380507,0.00078580866,0.0090482095,0.00037173336,0.00093723735,0.3588532,0.0010250309,0.014906509,0.00217621,0.12270991],"study_design_scores_gemma":[0.0018165585,0.010742045,0.07082165,0.00020298685,0.002395799,0.00015921274,0.00041135502,0.88783926,0.0015173991,0.021933794,0.0020272017,0.00013272335],"about_ca_topic_score_codex":0.0056675076,"about_ca_topic_score_gemma":0.0045482074,"teacher_disagreement_score":0.12217002,"about_ca_system_score_codex":0.0016634847,"about_ca_system_score_gemma":0.0024535088,"threshold_uncertainty_score":0.6461046},"labels":[],"label_agreement":null},{"id":"W2028703934","doi":"10.1214/14-aoas747","title":"Imputation of truncated p-values for meta-analysis methods and its genomic application","year":2014,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; University of Toronto","funders":"National Institute of Mental Health","keywords":"Inference; Imputation (statistics); False discovery rate; Raw data; Genomics; Statistical power; Multiple comparisons problem","score_opus":0.10692720296288404,"score_gpt":0.41959529675988433,"score_spread":0.3126680937970003,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2028703934","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.001822385,0.0037250577,0.9925212,0.00056423637,0.00015151511,0.00012357696,0.0003431188,0.0003570046,0.00039187304],"genre_scores_gemma":[0.10724651,0.002769875,0.8864388,0.00067880197,0.00025854207,0.0009650266,0.00078325166,0.00021361632,0.00064547407],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.89496726,0.092913315,0.003319055,0.003828209,0.0045708967,0.00040124753],"domain_scores_gemma":[0.7368174,0.23172693,0.0066727945,0.018530961,0.005822635,0.0004293321],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.099262334,0.0012109454,0.0027815467,0.0020518564,0.0006154513,0.0034624906,0.0035742517,0.0022947153,0.0033671325],"category_scores_gemma":[0.30501172,0.001015109,0.00433594,0.0045049763,0.0012871583,0.0024178512,0.0021595077,0.00441591,0.0008586056],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010680937,0.0001320457,0.022091458,0.005302738,0.01542736,0.0011534891,0.00059296365,0.19866659,0.002264348,0.15892161,0.015664533,0.57871485],"study_design_scores_gemma":[0.0004939796,0.00059682963,0.007731185,0.001466893,0.003677885,0.0014427633,0.00010486075,0.56473655,0.0038325447,0.38296804,0.032750424,0.000198009],"about_ca_topic_score_codex":0.0013085266,"about_ca_topic_score_gemma":0.0014119846,"teacher_disagreement_score":0.90073764,"about_ca_system_score_codex":0.0012555862,"about_ca_system_score_gemma":0.0025862579,"threshold_uncertainty_score":0.52495563},"labels":[],"label_agreement":null},{"id":"W2042759248","doi":"10.1214/14-aoas804","title":"A two-step approach to model precipitation extremes in California based on max-stable and marginal point processes","year":2015,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Hydrology and Drought Analysis","field":"Environmental Science","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Environment and Climate Change Canada","funders":"","keywords":"Maxima; Extreme value theory; Pairwise comparison; Statistics; Univariate; Mathematics; Block (permutation group theory); Maxima and minima; Generalized extreme value distribution; Confidence interval; Point estimation; Precipitation; Econometrics; Multivariate statistics; Meteorology; Geography; Combinatorics; Mathematical analysis","score_opus":0.05744828277066105,"score_gpt":0.28995323345361307,"score_spread":0.23250495068295202,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2042759248","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.08942405,0.00013959166,0.90749705,0.00019695456,0.000019650512,0.000057055866,0.00014480186,0.00017017384,0.0023506219],"genre_scores_gemma":[0.8494677,0.00016870817,0.14530605,0.000063269785,0.000033288474,0.00028336167,0.0002501524,0.000053486412,0.004374061],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995958,0.0001980623,0.00001738515,0.000083329105,0.000067778914,0.000037567148],"domain_scores_gemma":[0.999358,0.00041503864,0.0000636482,0.000042610074,0.000076440156,0.000044258777],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014991296,0.00045970717,0.00046192826,0.0005351242,0.0004849935,0.00067288615,0.0015323989,0.0006444431,0.0018182955],"category_scores_gemma":[0.0025413202,0.00041581987,0.0010351366,0.00048227474,0.000587634,0.0009084611,0.001075949,0.0011142573,0.00015405165],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00008339152,0.000045523582,0.004247198,0.000023503539,0.00006108692,0.00009042004,0.00011339894,0.9470827,0.0008632771,0.034596708,0.00037871415,0.012413979],"study_design_scores_gemma":[0.0000058373316,0.000019924197,0.00043932814,0.000001768147,0.000008505677,0.000010988787,0.000005989453,0.99210477,0.00008472157,0.0071056453,0.00020571455,0.0000068522486],"about_ca_topic_score_codex":0.008346222,"about_ca_topic_score_gemma":0.011654834,"teacher_disagreement_score":0.008346222,"about_ca_system_score_codex":0.0006324719,"about_ca_system_score_gemma":0.00092810136,"threshold_uncertainty_score":0.016595304},"labels":[],"label_agreement":null},{"id":"W2043089632","doi":"10.1214/12-aoas537","title":"A model for sequential evolution of ligands by exponential enrichment (SELEX) data","year":2012,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced biosensing and bioanalysis techniques","field":"Biochemistry, Genetics and Molecular Biology","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Systematic evolution of ligands by exponential enrichment; Oligonucleotide; Computational biology; Biology; DNA; Computer science; Genetics; RNA","score_opus":0.08406956993041677,"score_gpt":0.3643286796780921,"score_spread":0.2802591097476754,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2043089632","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.14497553,0.0007826888,0.83711684,0.0035919964,0.00015804522,0.00045536144,0.003674537,0.0013279582,0.007917084],"genre_scores_gemma":[0.8410969,0.0012918839,0.10680577,0.0013225042,0.00015319459,0.0025371166,0.004482165,0.00042024287,0.041890312],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99791175,0.0006902599,0.0001238589,0.0006535137,0.00030253385,0.00031812373],"domain_scores_gemma":[0.9904061,0.006922055,0.0009178024,0.0005066629,0.0008039379,0.00044342547],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004474633,0.001830582,0.002777398,0.0013622566,0.0008011494,0.0025253622,0.0042934725,0.0039290236,0.007541339],"category_scores_gemma":[0.0159659,0.0012509092,0.0018801141,0.0013998483,0.0023508375,0.0030279816,0.001956016,0.0033386077,0.0018419237],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00020774204,0.000077802884,0.002298427,0.00011818038,0.000066738045,0.0002497535,0.0001677844,0.942601,0.0023392632,0.04512472,0.0015378785,0.005210665],"study_design_scores_gemma":[0.0000459711,0.000038479662,0.0003292924,0.000010963054,0.000016599783,0.000043416076,0.000014220486,0.9864775,0.00035554354,0.011894786,0.0007513403,0.000021827975],"about_ca_topic_score_codex":0.010487275,"about_ca_topic_score_gemma":0.0051015317,"teacher_disagreement_score":0.010487275,"about_ca_system_score_codex":0.0021365099,"about_ca_system_score_gemma":0.0018521609,"threshold_uncertainty_score":0.025228262},"labels":[],"label_agreement":null},{"id":"W2057931901","doi":"10.1214/13-aoas684","title":"Beta regression for time series analysis of bounded data, with application to Canada Google® Flu Trends","year":2014,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced Statistical Process Monitoring","field":"Decision Sciences","cited_by":64,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Bounded function; Autoregressive model; Mathematics; Econometrics; Statistics; Autoregressive integrated moving average; Regression analysis; Time series; Computer science","score_opus":0.11027838743097933,"score_gpt":0.42605134573024833,"score_spread":0.315772958299269,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2057931901","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0045318627,0.00054460904,0.9927931,0.00036963954,0.000057460635,0.00004079507,0.00025474635,0.00045518373,0.0009526491],"genre_scores_gemma":[0.21889494,0.003059929,0.7688921,0.00032734033,0.0003039922,0.0008629119,0.0013765616,0.0006832447,0.0055990387],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9966085,0.0022652338,0.0001321463,0.00033841777,0.00055042317,0.000105363346],"domain_scores_gemma":[0.9875979,0.009781452,0.00086226146,0.0006660422,0.0009466423,0.00014564837],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007746878,0.00095944933,0.0008822595,0.0015276434,0.0005745802,0.0013069386,0.0010322586,0.00096479023,0.0037953407],"category_scores_gemma":[0.0326163,0.0004327769,0.0010476091,0.0030101973,0.0007606304,0.00091332395,0.0015194026,0.00243328,0.0012784244],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017731654,0.00015604035,0.010829689,0.000434134,0.0004243888,0.0008449104,0.000736502,0.32850528,0.004307956,0.3604489,0.017730184,0.27540475],"study_design_scores_gemma":[0.000015965174,0.00005388719,0.0023441932,0.000056874866,0.000023559853,0.00011133467,0.000089083966,0.9253388,0.00062060193,0.06192989,0.009378963,0.000036771853],"about_ca_topic_score_codex":0.016935574,"about_ca_topic_score_gemma":0.015313406,"teacher_disagreement_score":0.9830644,"about_ca_system_score_codex":0.0010378689,"about_ca_system_score_gemma":0.0023617058,"threshold_uncertainty_score":0.04096991},"labels":[],"label_agreement":null},{"id":"W2068388041","doi":"10.1214/09-aoas318","title":"Modeling hourly ozone concentration fields","year":2010,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Economic and Environmental Valuation","field":"Economics, Econometrics and Finance","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"BC Cancer Agency; University of British Columbia","funders":"","keywords":"Bayesian probability; Kriging; Computer science; Kalman filter; Computation; Point process; Algorithm; Machine learning; Mathematics; Statistics; Artificial intelligence","score_opus":0.1730773880426103,"score_gpt":0.2705241227258004,"score_spread":0.0974467346831901,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2068388041","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.2229747,0.00038728782,0.76331145,0.0007944046,0.000053330725,0.00008470894,0.0013753651,0.0007803679,0.01023845],"genre_scores_gemma":[0.9046593,0.00030238068,0.091265164,0.00008810721,0.00003937528,0.000105153464,0.0009122515,0.000052545834,0.0025757444],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997211,0.00011257897,0.000009097871,0.000046106008,0.0000787786,0.000032369942],"domain_scores_gemma":[0.99964654,0.00020609214,0.000043646123,0.000030404044,0.000053431435,0.000019990466],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007043211,0.00025069105,0.0003339209,0.00045558138,0.00024992687,0.0005016704,0.0009225458,0.000679598,0.0011103388],"category_scores_gemma":[0.0019010686,0.00032499217,0.0004251439,0.00056979764,0.00021972785,0.0007775957,0.0005121649,0.0005534284,0.00013033322],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000023891824,0.000033486478,0.0031874576,0.000017164659,0.000022959523,0.000012664636,0.000033161563,0.9675766,0.00056680763,0.010519418,0.00044901253,0.017557343],"study_design_scores_gemma":[0.0000060889697,0.0000073913843,0.0008883316,0.000002004634,0.0000024807077,0.0000023988393,0.000005797217,0.99657613,0.0001404836,0.0020424058,0.00032296698,0.0000036205613],"about_ca_topic_score_codex":0.045508716,"about_ca_topic_score_gemma":0.04487104,"teacher_disagreement_score":0.045508716,"about_ca_system_score_codex":0.00084083603,"about_ca_system_score_gemma":0.0011474087,"threshold_uncertainty_score":0.09048766},"labels":[],"label_agreement":null},{"id":"W2070512361","doi":"10.1214/12-aoas566","title":"Dating medieval English charters","year":2012,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Image Processing and 3D Reconstruction","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Documentation; History; Politics; Genealogy; Phrase; Variation (astronomy); Classics; Law; Computer science; Political science; Natural language processing","score_opus":0.05168577421080451,"score_gpt":0.30302702096211015,"score_spread":0.25134124675130565,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2070512361","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.94438136,0.0027034741,0.013025005,0.00025082874,0.00015120168,0.00013430584,0.018960658,0.0001751411,0.020218082],"genre_scores_gemma":[0.9550959,0.0019002364,0.017291794,0.000029689552,0.000078689925,0.00011830558,0.014443427,0.000094109564,0.010947914],"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9987808,0.00030069432,0.0001763293,0.00026610942,0.00035270283,0.00012337964],"domain_scores_gemma":[0.99061394,0.0023072646,0.0018215147,0.0011131988,0.0038855695,0.0002585784],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001679475,0.0002402172,0.00027399714,0.006045381,0.0008487469,0.0013841696,0.0004520846,0.00021806567,0.0028622304],"category_scores_gemma":[0.0110420175,0.0001843891,0.0001302906,0.006311666,0.0007287977,0.0009616214,0.0010112089,0.0003686116,0.0008766071],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005005631,0.00007865059,0.49454677,0.0008124716,0.000085312146,0.00090236025,0.041825626,0.004229955,0.0053370865,0.018851865,0.029409144,0.40342024],"study_design_scores_gemma":[0.000013854654,0.00004998825,0.7807599,0.00024625222,0.00004148134,0.0003278762,0.009463099,0.0027228787,0.0068512983,0.0013743193,0.19809671,0.000052402385],"about_ca_topic_score_codex":0.06280338,"about_ca_topic_score_gemma":0.13724455,"teacher_disagreement_score":0.06280338,"about_ca_system_score_codex":0.0014716038,"about_ca_system_score_gemma":0.0010727048,"threshold_uncertainty_score":0.124875665},"labels":[],"label_agreement":null},{"id":"W2076514995","doi":"10.1214/12-aoas584","title":"Robust VIF regression with application to variable selection in large data sets","year":2013,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":37,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"HEC Montréal","funders":"","keywords":"Outlier; Robustness (evolution); Feature selection; Computer science; Estimator; Regression; Robust regression; Lasso (programming language); Covariate; Regression analysis; Variance (accounting); Variance inflation factor; Econometrics; Artificial intelligence; Machine learning; Statistics; Mathematics; Multicollinearity; Economics","score_opus":0.20471608224767654,"score_gpt":0.4366883251623959,"score_spread":0.23197224291471938,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2076514995","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0015708044,0.00056877965,0.99686724,0.00021599972,0.00004556744,0.000044104658,0.00007655106,0.00038523762,0.00022565406],"genre_scores_gemma":[0.06811411,0.0013318072,0.9270338,0.00020402628,0.00028846855,0.0005853163,0.0006604818,0.00045643552,0.0013256223],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98697877,0.009645884,0.0004703641,0.0010867144,0.0015178364,0.0003003939],"domain_scores_gemma":[0.95035,0.041790497,0.0018668516,0.0032410466,0.0024697413,0.00028198754],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02301944,0.0021013562,0.0029732664,0.0036387276,0.0010706332,0.0017908436,0.0029733132,0.0024400346,0.0028308465],"category_scores_gemma":[0.08605167,0.0010406852,0.0027124058,0.0048927455,0.0016713708,0.0016988728,0.002935951,0.004473626,0.0012251678],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018713929,0.00012299785,0.0055504707,0.0005405953,0.00080881617,0.000552516,0.00031573628,0.5704059,0.0019353757,0.11416374,0.008422008,0.29699466],"study_design_scores_gemma":[0.00002971961,0.00004770424,0.00073224807,0.00005094403,0.000030995478,0.00009466768,0.00002372389,0.93725,0.00065708853,0.056677833,0.0043699457,0.000035033994],"about_ca_topic_score_codex":0.00813426,"about_ca_topic_score_gemma":0.0066318675,"teacher_disagreement_score":0.02301944,"about_ca_system_score_codex":0.0010769761,"about_ca_system_score_gemma":0.0027598669,"threshold_uncertainty_score":0.121739924},"labels":[],"label_agreement":null},{"id":"W2088906176","doi":"10.1214/13-aoas626","title":"Travel time estimation for ambulances using Bayesian data augmentation","year":2013,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Transportation Planning and Optimization","field":"Social Sciences","cited_by":109,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Global Positioning System; Bayesian probability; Computer science; Estimation; Path (computing); Data mining; Statistics; Mathematics; Artificial intelligence; Engineering","score_opus":0.14421433261847197,"score_gpt":0.3997035024737231,"score_spread":0.25548916985525116,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2088906176","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.021558158,0.0001532271,0.97678566,0.00017718229,0.00001869487,0.000039631956,0.0003838831,0.0003042557,0.0005792917],"genre_scores_gemma":[0.50933075,0.000540765,0.48413217,0.00018487619,0.00010770064,0.0005107994,0.0032457197,0.00022536334,0.0017218903],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99782515,0.0011473974,0.000098443525,0.00035891478,0.00046439405,0.000105728155],"domain_scores_gemma":[0.9926692,0.004686665,0.0007757705,0.0009430746,0.0007798451,0.00014533392],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0031064644,0.00079811894,0.0014896061,0.00162453,0.00051512645,0.0011300052,0.002742977,0.0013605134,0.0020920876],"category_scores_gemma":[0.020791534,0.0011219436,0.0012323058,0.0021270157,0.0007502825,0.0027469702,0.0015740043,0.002289332,0.0006763603],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00010045869,0.00007947709,0.00498691,0.00007272308,0.00009332896,0.000045665976,0.000105346786,0.92726237,0.0007815151,0.01678413,0.001666253,0.048021816],"study_design_scores_gemma":[0.00000996061,0.000012533215,0.0008149136,0.000010881965,0.0000074954887,0.000016872635,0.0000075062544,0.98944014,0.00020925679,0.008900558,0.0005539518,0.000016057169],"about_ca_topic_score_codex":0.021686926,"about_ca_topic_score_gemma":0.020451069,"teacher_disagreement_score":0.021686926,"about_ca_system_score_codex":0.0012092348,"about_ca_system_score_gemma":0.0017761895,"threshold_uncertainty_score":0.043121338},"labels":[],"label_agreement":null},{"id":"W2098683668","doi":"10.1214/08-aoas222","title":"Handbook for the GREAT08 Challenge: An image analysis competition for cosmological lensing","year":2009,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Galaxies: Formation, Evolution, Phenomena","field":"Physics and Astronomy","cited_by":121,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria; University of British Columbia","funders":"Science and Technology Facilities Council; European Commission; National Aeronautics and Space Administration; University College London; California Institute of Technology; Jet Propulsion Laboratory","keywords":"Weak gravitational lensing; Dark energy; Dark matter; Galaxy; Inference; Strong gravitational lensing; Cosmology; Gravitational lens; Physics; Data science; Astrophysics; Computer science; Theoretical physics; Astronomy; Artificial intelligence; Redshift","score_opus":0.05355674815407757,"score_gpt":0.3134809803572959,"score_spread":0.2599242322032183,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2098683668","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0041481215,0.0390286,0.40685302,0.15244131,0.03661472,0.0018808107,0.04366672,0.06288806,0.25247872],"genre_scores_gemma":[0.033531085,0.020617524,0.37513515,0.020177804,0.024016881,0.002610619,0.102961354,0.030595502,0.3903541],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.987038,0.0026196237,0.00081717846,0.0014120304,0.007301448,0.0008118002],"domain_scores_gemma":[0.96691173,0.009538485,0.00063180446,0.005148883,0.012800198,0.0049689887],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.019579057,0.0025412394,0.0033075917,0.005091679,0.0025894986,0.010081799,0.0057929778,0.0056058806,0.116793446],"category_scores_gemma":[0.044117156,0.0014203253,0.0018825433,0.00556711,0.0018966267,0.01053015,0.008729979,0.006309089,0.1079672],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000027699483,0.000018175868,0.000058979746,0.000065562715,0.0000087939625,0.000015598876,0.00001627614,0.00027796935,0.00013082795,0.0041648964,0.9501196,0.045095578],"study_design_scores_gemma":[0.000053078867,0.000037552076,0.00056034,0.00011188112,0.000008430893,0.00017619459,0.00006564137,0.007934963,0.0003583345,0.02675904,0.9638863,0.000048205682],"about_ca_topic_score_codex":0.013159919,"about_ca_topic_score_gemma":0.019263957,"teacher_disagreement_score":0.116793446,"about_ca_system_score_codex":0.0032484205,"about_ca_system_score_gemma":0.00681085,"threshold_uncertainty_score":0.39071304},"labels":[],"label_agreement":null},{"id":"W2107904222","doi":"10.1214/10-aoas398l","title":"Discussion of: A statistical analysis of multiple temperature proxies: Are reconstructions of surface temperatures over the last 1000 years reliable?","year":2011,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Tree-ring climate responses","field":"Earth and Planetary Sciences","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Guelph","funders":"","keywords":"Proxy (statistics); Autocorrelation; Statistical analysis; Paleoclimatology; Econometrics; History; Geology; Psychology; Statistics; Mathematics; Climate change; Oceanography","score_opus":0.03359346518537585,"score_gpt":0.2685703369750909,"score_spread":0.23497687178971507,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2107904222","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.003680706,0.016711647,0.12497468,0.81887454,0.030979753,0.00007064998,0.0011438254,0.00018905725,0.0033751815],"genre_scores_gemma":[0.22089157,0.023286317,0.12255456,0.36742467,0.24995272,0.0006282001,0.0012867746,0.00088832935,0.013086845],"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.96718276,0.022471473,0.0022803512,0.004134366,0.0035827658,0.00034831872],"domain_scores_gemma":[0.74776256,0.21790832,0.009063208,0.008781556,0.014105683,0.0023787648],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.06718966,0.0011229733,0.0018133947,0.0024681624,0.0023730383,0.005923705,0.005250442,0.00883307,0.008092367],"category_scores_gemma":[0.2772078,0.00048167983,0.0024996533,0.004704736,0.00909135,0.009556366,0.0026507757,0.009903645,0.0021308144],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002015025,0.00009035044,0.008224595,0.0012367591,0.001049718,0.00083514967,0.00092096557,0.007430037,0.0007707033,0.39668146,0.49416173,0.08839698],"study_design_scores_gemma":[0.000041603696,0.00009112944,0.0064494363,0.0006659693,0.00020965758,0.00096655264,0.0007650537,0.016106104,0.0007975634,0.67272615,0.30102524,0.00015557413],"about_ca_topic_score_codex":0.0047415528,"about_ca_topic_score_gemma":0.0033202593,"teacher_disagreement_score":0.06718966,"about_ca_system_score_codex":0.0027155846,"about_ca_system_score_gemma":0.004311086,"threshold_uncertainty_score":0.35533714},"labels":[],"label_agreement":null},{"id":"W2126292488","doi":"10.1214/09-aoas285","title":"BART: Bayesian additive regression trees","year":2010,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1515,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Division of Mathematical Sciences; Natural Sciences and Engineering Research Council of Canada; Isaac Newton Institute for Mathematical Sciences; National Science Foundation","keywords":"Frequentist inference; Bayesian probability; Nonparametric regression; Bayesian inference; Bayesian linear regression; Markov chain Monte Carlo; Boosting (machine learning); Feature selection; Inference; Bayesian average","score_opus":0.13013747162264705,"score_gpt":0.41145148035021956,"score_spread":0.2813140087275725,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2126292488","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0011087293,0.0004862959,0.995223,0.00038186388,0.00008228606,0.000028684619,0.00025620923,0.00084069406,0.0015923185],"genre_scores_gemma":[0.09928289,0.0019681638,0.8833504,0.0010839901,0.0005087338,0.00048915815,0.0017891085,0.00096841267,0.010559123],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9961551,0.0020649591,0.000117413125,0.00036844044,0.001085178,0.00020893659],"domain_scores_gemma":[0.9948836,0.003257942,0.00040085582,0.00055102614,0.00067799486,0.00022847984],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0058581834,0.0011799616,0.0018380763,0.0017705053,0.0008235721,0.002335827,0.003989236,0.0024495486,0.0068296264],"category_scores_gemma":[0.01747052,0.0011981198,0.0016474655,0.0025818679,0.0010535964,0.0029266388,0.002792375,0.004306383,0.0059792106],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00020175049,0.00011730544,0.0017066386,0.00028028263,0.00025150555,0.0001575882,0.0002067004,0.3565176,0.0015903783,0.34776902,0.032899283,0.25830194],"study_design_scores_gemma":[0.000025406527,0.00003496357,0.00023386144,0.000043409207,0.000031752617,0.00008617915,0.000010269708,0.8395257,0.00043633478,0.14373517,0.015806628,0.00003019958],"about_ca_topic_score_codex":0.0035501553,"about_ca_topic_score_gemma":0.004721936,"teacher_disagreement_score":0.0068296264,"about_ca_system_score_codex":0.00092438515,"about_ca_system_score_gemma":0.0015070711,"threshold_uncertainty_score":0.030981421},"labels":[],"label_agreement":null},{"id":"W2148429219","doi":"10.1214/10-aoas378","title":"Detecting multiple authorship of United States Supreme Court legal decisions using function words","year":2011,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Authorship Attribution and Profiling","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Supreme court; Style (visual arts); Function (biology); Legal writing; Writing style; Statistical analysis","score_opus":0.24488643153322526,"score_gpt":0.347889792177648,"score_spread":0.10300336064442273,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2148429219","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9905393,0.00018142392,0.007059358,0.00007097789,0.000027630424,0.00002321938,0.000244178,0.00006341421,0.001790432],"genre_scores_gemma":[0.9962519,0.000037193142,0.003156318,0.000010058264,0.000019952206,0.000014789499,0.00022455075,0.000014410849,0.00027087875],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.992353,0.0038578245,0.0011695819,0.0009074754,0.0013089498,0.00040317318],"domain_scores_gemma":[0.8277453,0.1250159,0.026607294,0.009003758,0.0094291475,0.0021985483],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0057574096,0.0004954014,0.0004931457,0.010259413,0.0010464725,0.0026715167,0.00046905765,0.00073879317,0.001435474],"category_scores_gemma":[0.080438785,0.00022833288,0.0003710158,0.0055254037,0.00118372,0.0027378688,0.0014764625,0.0007525908,0.00055529835],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005703649,0.00015513733,0.8283071,0.00017978884,0.00013526891,0.00050704373,0.0085396115,0.0017534255,0.008934806,0.003688864,0.0010653995,0.14616321],"study_design_scores_gemma":[0.000053595228,0.0004381701,0.8638601,0.0001906005,0.00020160149,0.0022429945,0.013267135,0.063424334,0.022178095,0.027196432,0.0067506414,0.00019636391],"about_ca_topic_score_codex":0.0008307722,"about_ca_topic_score_gemma":0.0012413528,"teacher_disagreement_score":0.010259413,"about_ca_system_score_codex":0.00054201676,"about_ca_system_score_gemma":0.00062076416,"threshold_uncertainty_score":0.030448437},"labels":[],"label_agreement":null},{"id":"W2492804687","doi":"10.1214/16-aoas909","title":"A Bayesian graphical model for genome-wide association studies (GWAS)","year":2016,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Lunenfeld-Tanenbaum Research Institute; York University; University of Toronto","funders":"National Cancer Institute","keywords":"Genome-wide association study; Computer science; Bayesian probability; Genetic association; Graphical model; Single-nucleotide polymorphism; Computational biology; Data mining; Machine learning; Artificial intelligence; Biology; Genetics","score_opus":0.0591072333190735,"score_gpt":0.3411212849881219,"score_spread":0.28201405166904836,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2492804687","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0014603386,0.0003964225,0.99607456,0.00049227907,0.0000368297,0.000036501573,0.0005482687,0.00030853046,0.00064629945],"genre_scores_gemma":[0.17411163,0.003118771,0.81037,0.0012060464,0.00047487413,0.001262889,0.004008169,0.00039571128,0.005051814],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9952678,0.0031456647,0.0002051311,0.0006664172,0.00055364537,0.00016132586],"domain_scores_gemma":[0.98780847,0.010246232,0.00068297726,0.00053832756,0.00052893994,0.00019503509],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00845036,0.0013220904,0.002114203,0.0028732314,0.00083779177,0.0025631483,0.004028929,0.0032680316,0.006593019],"category_scores_gemma":[0.026557272,0.001228271,0.0024031845,0.003913591,0.0024628066,0.0029243575,0.0018965112,0.0037459896,0.0024149881],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00014596757,0.000054097443,0.0026773333,0.00034330512,0.00021846537,0.0002746547,0.0002233976,0.4591268,0.0010170941,0.47139123,0.008050198,0.0564773],"study_design_scores_gemma":[0.00005575615,0.000033385026,0.00049845286,0.000050065843,0.000052236486,0.00014162173,0.000020375457,0.5703768,0.00011861704,0.42260104,0.0060128565,0.000038758604],"about_ca_topic_score_codex":0.011685903,"about_ca_topic_score_gemma":0.011592122,"teacher_disagreement_score":0.011685903,"about_ca_system_score_codex":0.0017404841,"about_ca_system_score_gemma":0.0022046093,"threshold_uncertainty_score":0.04469031},"labels":[],"label_agreement":null},{"id":"W2569003638","doi":"10.1214/16-aoas964","title":"Inferring rooted population trees using asymmetric neighbor joining","year":2016,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Genetic diversity and population structure","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Population; Artificial intelligence; Computer science; Mathematics; Geography; Demography; Sociology","score_opus":0.06098441619541242,"score_gpt":0.31021719994415936,"score_spread":0.24923278374874694,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2569003638","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.026802488,0.00009626319,0.97227174,0.00006277881,0.000015606927,0.000021435777,0.00015180359,0.00017105218,0.00040681355],"genre_scores_gemma":[0.3629756,0.00019408618,0.63463485,0.000085139865,0.00006789496,0.00011173944,0.0012505641,0.00014501893,0.00053513766],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99768,0.0010776503,0.00012170553,0.0004901133,0.00052653486,0.00010410503],"domain_scores_gemma":[0.99103945,0.0057146694,0.0012277792,0.0012028267,0.0006030822,0.00021219636],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0034366536,0.00061569596,0.0010087572,0.0025415292,0.000922864,0.0011050628,0.002581632,0.0009772559,0.001063446],"category_scores_gemma":[0.018814547,0.0005900748,0.00080517464,0.0021153751,0.00079112884,0.0017420771,0.0015803394,0.0020653356,0.0004809645],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00021994287,0.00016368179,0.021125736,0.00019605715,0.00029237807,0.0003866803,0.0008033587,0.6522746,0.009544293,0.10786435,0.0020206906,0.20510826],"study_design_scores_gemma":[0.000015289696,0.000016556938,0.0006950986,0.000010628088,0.000014661473,0.00007214848,0.000031634943,0.9392643,0.0012606244,0.05732427,0.0012796354,0.000015111322],"about_ca_topic_score_codex":0.0017442652,"about_ca_topic_score_gemma":0.0020735136,"teacher_disagreement_score":0.0034366536,"about_ca_system_score_codex":0.0005427894,"about_ca_system_score_gemma":0.00074398215,"threshold_uncertainty_score":0.018175006},"labels":[],"label_agreement":null},{"id":"W2602208471","doi":"10.1214/18-aoas1162","title":"Exact spike train inference via $\\ell_{0}$ optimization","year":2018,"lang":"en","type":"preprint","venue":"The Annals of Applied Statistics","topic":"Neural dynamics and brain function","field":"Neuroscience","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; National Institutes of Health; National Science Foundation","keywords":"Python (programming language); Computer science; TRACE (psycholinguistics); Inference; Software; Optimization problem; Spike (software development); Convex optimization; Measure (data warehouse); Algorithm; Mathematical optimization; Regular polygon; Mathematics; Artificial intelligence; Programming language","score_opus":0.10499761598033654,"score_gpt":0.3403495868487135,"score_spread":0.235351970868377,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2602208471","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0073886546,0.00028011893,0.9852889,0.0006226281,0.00007761726,0.000036556543,0.0003851224,0.0038559579,0.0020644239],"genre_scores_gemma":[0.2693916,0.00048549648,0.71164715,0.0011557799,0.0002478563,0.00035933897,0.0035825297,0.0027136584,0.0104166325],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99912053,0.00030223234,0.00005409397,0.00021596378,0.00018431133,0.00012292438],"domain_scores_gemma":[0.9969837,0.002161008,0.00014712977,0.00030567133,0.0002605838,0.00014177669],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0020974493,0.0012527341,0.0018184623,0.00071760465,0.0006482018,0.0018754491,0.0027311046,0.0019183524,0.010607657],"category_scores_gemma":[0.009949487,0.0009990442,0.0013839363,0.0010836093,0.0014041341,0.0021519565,0.0022124283,0.0037501508,0.003516207],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000211859,0.00006615103,0.00084357406,0.00012751177,0.000081191196,0.000074558964,0.000047584363,0.87511957,0.0010650486,0.03222102,0.010024078,0.0801179],"study_design_scores_gemma":[0.00001192635,0.000005634765,0.000044485754,0.0000053450426,0.0000033455983,0.0000072223997,0.0000033527174,0.98530835,0.00031649732,0.013681881,0.0006073921,0.0000045334414],"about_ca_topic_score_codex":0.013361345,"about_ca_topic_score_gemma":0.018754024,"teacher_disagreement_score":0.013361345,"about_ca_system_score_codex":0.0017360481,"about_ca_system_score_gemma":0.0037751535,"threshold_uncertainty_score":0.03548616},"labels":[],"label_agreement":null},{"id":"W2605027696","doi":"10.1214/16-aoas1002","title":"Electricity price dependence in New York State zones: A robust detrended correlation approach","year":2017,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Electric Power System Optimization","field":"Engineering","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"HEC Montréal","funders":"","keywords":"Econometrics; Electricity; Detrended fluctuation analysis; Grid; Estimator; Electricity price; Computer science; Economics; Statistics; Mathematics","score_opus":0.0574632283859776,"score_gpt":0.26386248696745934,"score_spread":0.20639925858148173,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2605027696","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.69239557,0.0008523366,0.30114907,0.00097308983,0.000082854654,0.000070838876,0.0011058134,0.00024677493,0.00312357],"genre_scores_gemma":[0.98306376,0.00024938976,0.014067891,0.000051847204,0.000054826673,0.000043537068,0.0010303042,0.000042288055,0.0013960643],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991804,0.00042948374,0.000041128464,0.00018208375,0.0001016395,0.00006537404],"domain_scores_gemma":[0.99175996,0.0058643045,0.00091775856,0.0006880751,0.0005955024,0.00017434881],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002686613,0.0004489879,0.00062607066,0.0010554494,0.0003373578,0.0011837324,0.0013403294,0.0006855129,0.0013255072],"category_scores_gemma":[0.013728826,0.00047098866,0.0007519173,0.0009972702,0.0007583509,0.0012688248,0.0009776908,0.0014092964,0.000169325],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00025825747,0.00016775286,0.08888382,0.00009161066,0.00054072635,0.0007252579,0.00022242176,0.8296993,0.001291934,0.040519238,0.0028098503,0.034789898],"study_design_scores_gemma":[0.0000035765975,0.000007711304,0.0059295627,0.000005639494,0.000010760264,0.0000148055115,0.000018385139,0.99032825,0.00011103848,0.0032833933,0.00027682597,0.000010004696],"about_ca_topic_score_codex":0.028378814,"about_ca_topic_score_gemma":0.021790719,"teacher_disagreement_score":0.028378814,"about_ca_system_score_codex":0.00077629887,"about_ca_system_score_gemma":0.0006117352,"threshold_uncertainty_score":0.05642724},"labels":[],"label_agreement":null},{"id":"W2761069967","doi":"10.1214/17-aoas1044","title":"Latent class modeling using matrix covariates with application to identifying early placebo responders based on EEG signals","year":2017,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Mental Health Research Topics","field":"Psychology","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"National Institute of Mental Health","keywords":"Covariate; Latent class model; Latent variable; Statistics; Placebo; Missing data; Bayesian probability; Matrix decomposition; Artificial intelligence; Computer science; Econometrics; Psychology; Mathematics; Medicine","score_opus":0.25930561809344677,"score_gpt":0.49458103777685225,"score_spread":0.23527541968340548,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2761069967","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007648695,0.00025166,0.9912327,0.00024331572,0.000022062823,0.00006236428,0.000099141405,0.00017949364,0.0002605691],"genre_scores_gemma":[0.40162066,0.0013061728,0.5906352,0.00033367603,0.00030494496,0.0007683183,0.0009701118,0.0002412097,0.0038198],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.995749,0.0028982987,0.000114062605,0.00064408657,0.00038620422,0.0002082327],"domain_scores_gemma":[0.9876188,0.009817757,0.00091915333,0.00079265604,0.0006174801,0.0002340956],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008829072,0.0012188659,0.0015136928,0.0013487225,0.00069842674,0.0014992001,0.0015337389,0.0015555879,0.0033226428],"category_scores_gemma":[0.021482134,0.0005536099,0.0019088463,0.0015619796,0.001311034,0.0017355169,0.0018081899,0.0029518397,0.0007717163],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008779685,0.0004105387,0.0144597655,0.00042464002,0.0005131113,0.00047274848,0.0010479625,0.47219232,0.006305649,0.19019271,0.004721247,0.3083814],"study_design_scores_gemma":[0.000028508122,0.00008088174,0.0009771802,0.000023531507,0.000025192014,0.00005345457,0.000034638328,0.9576948,0.0003700178,0.039412543,0.0012738514,0.000025330175],"about_ca_topic_score_codex":0.0041438937,"about_ca_topic_score_gemma":0.003923295,"teacher_disagreement_score":0.008829072,"about_ca_system_score_codex":0.00070298556,"about_ca_system_score_gemma":0.0016860787,"threshold_uncertainty_score":0.046693146},"labels":[],"label_agreement":null},{"id":"W2767505813","doi":"10.1214/21-aoas1471","title":"Improving exoplanet detection power: Multivariate Gaussian process models for stellar activity","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Spectroscopy and Chemometric Analyses","field":"Chemistry","cited_by":36,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Astrophysics Division; Pennsylvania Space Grant Consortium; Natural Sciences and Engineering Research Council of Canada; European Commission; National Aeronautics and Space Administration; University of Pennsylvania; Simons Foundation Autism Research Initiative; Pennsylvania State University; Simons Foundation; Institute for Computational and Data Sciences, Pennsylvania State University; Division of Mathematical Sciences; National Science Foundation","keywords":"Exoplanet; Radial velocity; Physics; Planet; Astrophysics; Astronomy; Astron; Gaussian process; Gaussian; Stars","score_opus":0.05106795244628423,"score_gpt":0.325273856738226,"score_spread":0.2742059042919418,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2767505813","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.037421044,0.00091885886,0.9587833,0.0007307088,0.00009036652,0.000051824674,0.00038453925,0.0005909731,0.0010284393],"genre_scores_gemma":[0.7857145,0.0021139816,0.2030461,0.0005539725,0.0006135321,0.00024172825,0.001783405,0.00031174047,0.005621055],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99870324,0.0005744547,0.000060235445,0.0003260542,0.00019184362,0.00014423118],"domain_scores_gemma":[0.99401855,0.004106997,0.0006902198,0.0005108749,0.00050467084,0.00016871002],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0044784606,0.0016262718,0.0016829803,0.0016080552,0.00059555063,0.0017688643,0.002710437,0.001617335,0.001628924],"category_scores_gemma":[0.012482779,0.0007886393,0.0023149862,0.0015113566,0.0011780643,0.001870603,0.0017642918,0.0038385305,0.0007107361],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00021523796,0.00020619163,0.012516487,0.00010614564,0.0003971265,0.00017505232,0.00028621248,0.8536628,0.0015521724,0.049479965,0.0031558422,0.078246705],"study_design_scores_gemma":[0.000007971167,0.000010780679,0.0005945447,0.00000563576,0.000015343168,0.000012617949,0.000004642637,0.989902,0.00007560582,0.009044407,0.00031531553,0.000011145362],"about_ca_topic_score_codex":0.014357846,"about_ca_topic_score_gemma":0.009155998,"teacher_disagreement_score":0.014357846,"about_ca_system_score_codex":0.001113638,"about_ca_system_score_gemma":0.0010525184,"threshold_uncertainty_score":0.028548539},"labels":[],"label_agreement":null},{"id":"W2789596487","doi":"10.1214/17-aoas1108","title":"Simultaneous modelling of movement, measurement error, and observer dependence in mark-recapture distance sampling: An application to Arctic bird surveys","year":2018,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Wildlife Ecology and Conservation","field":"Environmental Science","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Environment and Climate Change Canada","funders":"","keywords":"Estimator; Distance sampling; Mark and recapture; Observer (physics); Statistics; Abundance estimation; Sampling (signal processing); Observational error; Computer science; Abundance (ecology); Mathematics; Econometrics; Ecology; Biology; Computer vision","score_opus":0.10828061340485357,"score_gpt":0.3023662759143145,"score_spread":0.1940856625094609,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2789596487","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.067826055,0.00033075106,0.9305486,0.00029310133,0.000029228775,0.000052906595,0.000113384274,0.00013355156,0.0006723753],"genre_scores_gemma":[0.82370174,0.000632785,0.17008358,0.00012529669,0.00012137074,0.00021593364,0.0003439669,0.00011475688,0.0046605784],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99702686,0.0017836868,0.00011875062,0.0005228143,0.00032512436,0.00022290047],"domain_scores_gemma":[0.9814555,0.014576886,0.002165729,0.00076338096,0.0007737987,0.0002646318],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008121854,0.0009433949,0.0011225304,0.00071165815,0.00068581547,0.0011283101,0.0025014644,0.0016193335,0.0011268406],"category_scores_gemma":[0.025196133,0.0011281499,0.0017481686,0.0010774836,0.0013010157,0.00140847,0.0018381588,0.0017314249,0.0002028049],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00010696526,0.00006572064,0.037216384,0.00011864753,0.0002234801,0.00043710778,0.00052210136,0.89674217,0.0015266459,0.0299078,0.0005322289,0.03260076],"study_design_scores_gemma":[0.000008308454,0.000037132362,0.0049778637,0.00000996966,0.00003653848,0.00008204492,0.000027360371,0.9880194,0.00020087027,0.005851951,0.00072496076,0.000023622919],"about_ca_topic_score_codex":0.065877855,"about_ca_topic_score_gemma":0.04660486,"teacher_disagreement_score":0.065877855,"about_ca_system_score_codex":0.0015989229,"about_ca_system_score_gemma":0.0018734355,"threshold_uncertainty_score":0.13098878},"labels":[],"label_agreement":null},{"id":"W2790880830","doi":"10.1214/17-aoas1092","title":"Automated threshold selection for extreme value analysis via ordered goodness-of-fit tests with adjustment for false discovery rate","year":2018,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Hydrology and Drought Analysis","field":"Environmental Science","cited_by":82,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Environment and Climate Change Canada","funders":"","keywords":"Generalized Pareto distribution; Goodness of fit; False discovery rate; Statistics; Extreme value theory; Sample size determination; Mathematics; Selection (genetic algorithm); Multiple comparisons problem; Threshold limit value; Computer science; Model selection; Artificial intelligence","score_opus":0.05727779193429081,"score_gpt":0.3207032615150507,"score_spread":0.26342546958075985,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2790880830","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009392961,0.00014815971,0.98752534,0.0001303849,0.00008568932,0.00035177649,0.0001811524,0.0016938332,0.0004906249],"genre_scores_gemma":[0.13445942,0.00009861619,0.8619163,0.00020384236,0.000068048445,0.0018659773,0.00053240126,0.0005316762,0.00032371737],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.93410754,0.045472473,0.004991998,0.005921488,0.0085177105,0.0009887932],"domain_scores_gemma":[0.7391957,0.20814766,0.013431737,0.025303185,0.012745859,0.0011758357],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.08926175,0.0016087006,0.0035161464,0.005204268,0.0016681992,0.0037017388,0.004938726,0.002433501,0.0037822095],"category_scores_gemma":[0.26150733,0.00095635443,0.002744787,0.00487991,0.0031735078,0.0027216892,0.003015887,0.0049441913,0.0016891131],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0019506835,0.0008791346,0.061031457,0.0020163106,0.0024972677,0.001972268,0.0021581352,0.07691275,0.028880367,0.07263017,0.018210065,0.7308614],"study_design_scores_gemma":[0.0005429525,0.0014034989,0.029553458,0.00035249468,0.00047829899,0.0013743028,0.00033511716,0.756821,0.031255595,0.16229597,0.0151351895,0.0004522182],"about_ca_topic_score_codex":0.0012652695,"about_ca_topic_score_gemma":0.0017558968,"teacher_disagreement_score":0.08926175,"about_ca_system_score_codex":0.0012468238,"about_ca_system_score_gemma":0.0038528573,"threshold_uncertainty_score":0.47206688},"labels":[],"label_agreement":null},{"id":"W2884750441","doi":"10.1214/18-aoas1160sf","title":"When should modes of inference disagree? Some simple but challenging examples","year":2018,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Bayesian Modeling and Causal Inference","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Frequentist inference; Fiducial inference; Inference; Bayes' theorem; Computer science; Simple (philosophy); Bayesian inference; Statistical inference; Artificial intelligence; Machine learning; Predictive inference; Econometrics; Bayesian probability; Mathematics; Epistemology; Statistics; Philosophy","score_opus":0.20253955059033146,"score_gpt":0.36163241960983933,"score_spread":0.15909286901950787,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2884750441","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.048321176,0.013721266,0.6191606,0.19694026,0.0014356653,0.00029754132,0.0005840005,0.00055402797,0.11898555],"genre_scores_gemma":[0.74825364,0.0061769607,0.21927515,0.014313792,0.0026977926,0.0006533566,0.0004149919,0.00039895036,0.00781529],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.967478,0.023300983,0.0013465002,0.0022708264,0.004569087,0.0010345274],"domain_scores_gemma":[0.8215755,0.16012433,0.0038025687,0.005852664,0.0075026946,0.0011422191],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.041076265,0.0012951043,0.00222852,0.0031648024,0.008244418,0.0066874577,0.004150317,0.010859937,0.005212804],"category_scores_gemma":[0.20168886,0.0010504073,0.001578622,0.005045661,0.0153731555,0.015408616,0.004873789,0.009961688,0.000954547],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016342038,0.000037229263,0.0014619733,0.00021306456,0.00007233881,0.0009066187,0.0028721313,0.0042678555,0.00013123873,0.9500809,0.01494855,0.024844622],"study_design_scores_gemma":[0.000025402063,0.0000065183913,0.0002778601,0.00008295096,0.0000070403216,0.00018060314,0.00041990922,0.004903016,0.000055542816,0.9883994,0.0056212177,0.00002054616],"about_ca_topic_score_codex":0.0025266754,"about_ca_topic_score_gemma":0.0019954578,"teacher_disagreement_score":0.95892376,"about_ca_system_score_codex":0.0025213296,"about_ca_system_score_gemma":0.0012942418,"threshold_uncertainty_score":0.21723461},"labels":[],"label_agreement":null},{"id":"W2890389585","doi":"10.1214/17-aoas1130","title":"Estimating and comparing cancer progression risks under varying surveillance protocols","year":2018,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"National Cancer Institute; National Institutes of Health; Prostate Cancer Canada; Genomic Health; Canary Foundation; U.S. Department of Defense","keywords":"Prostate cancer; Schema (genetic algorithms); Medicine; Cancer; Computer science; Econometrics; Machine learning; Internal medicine; Mathematics","score_opus":0.47324418958584247,"score_gpt":0.5406745580878244,"score_spread":0.06743036850198192,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2890389585","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.74509805,0.0012779293,0.25047407,0.0009348978,0.000049026687,0.000298998,0.00068640226,0.00013663706,0.0010440524],"genre_scores_gemma":[0.92712504,0.00071543554,0.07018586,0.00014630497,0.000064181055,0.00039959105,0.00091824547,0.00002791837,0.00041734288],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9815231,0.0141381705,0.00074901804,0.0023408479,0.00086000195,0.0003889239],"domain_scores_gemma":[0.7810054,0.19921514,0.010316656,0.0072430465,0.0016608727,0.000558889],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.059869707,0.0006917927,0.0011541047,0.0015857319,0.00046202517,0.0019007256,0.0017752639,0.0017777957,0.00063343265],"category_scores_gemma":[0.18379173,0.0006673409,0.0015763275,0.001365474,0.0015854164,0.0022062242,0.0017591562,0.001836671,0.00009038324],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.001439496,0.00047678922,0.14800727,0.00042935327,0.00201636,0.00025649348,0.000592467,0.71863836,0.0017933439,0.03224095,0.0006983701,0.09341065],"study_design_scores_gemma":[0.00029438452,0.0010968469,0.04962424,0.00014696235,0.00090420985,0.00017316728,0.0002890056,0.86910063,0.0028181153,0.07436682,0.0010882566,0.0000973441],"about_ca_topic_score_codex":0.005874543,"about_ca_topic_score_gemma":0.0033464055,"teacher_disagreement_score":0.059869707,"about_ca_system_score_codex":0.0016593349,"about_ca_system_score_gemma":0.002253203,"threshold_uncertainty_score":0.31662506},"labels":[],"label_agreement":null},{"id":"W2890985400","doi":"10.1214/18-aoas1138","title":"A frequency-calibrated Bayesian search for new particles","year":2018,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Gaussian Processes and Bayesian Inference","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"False discovery rate; Computer science; Inference; Bayesian probability; Bayesian inference; Statistical inference; Data mining; Bayesian hierarchical modeling; Nuisance parameter; Function (biology); Machine learning; Artificial intelligence; Statistics; Mathematics","score_opus":0.08347519627341762,"score_gpt":0.33547479827747323,"score_spread":0.2519996020040556,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2890985400","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007843829,0.00021063123,0.9896079,0.0005478233,0.000029451829,0.000034411336,0.00004596331,0.00009316553,0.0015867901],"genre_scores_gemma":[0.21868697,0.0006287947,0.7758299,0.0005376725,0.00018250276,0.00022760582,0.00021510581,0.00010691297,0.0035845244],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99731976,0.0012321237,0.00008350737,0.00057416497,0.0006547968,0.00013564386],"domain_scores_gemma":[0.99013305,0.0075017866,0.000800608,0.00073511095,0.000612958,0.0002164861],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007037996,0.0006216849,0.0015631998,0.0023828752,0.0015542464,0.0021251966,0.0031117154,0.0027189946,0.0037292447],"category_scores_gemma":[0.025378257,0.0009285729,0.00114382,0.0016663761,0.0024176743,0.004229398,0.0028978994,0.002553753,0.0007528892],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013668445,0.00013317887,0.001971623,0.00017452589,0.0000989692,0.00016268533,0.00020296742,0.18584013,0.0029951534,0.7119829,0.0026751617,0.09362607],"study_design_scores_gemma":[0.000044778684,0.000044356297,0.00036643972,0.00004315319,0.000022847256,0.00012977223,0.000031420164,0.67256176,0.0010854057,0.32338208,0.0022432557,0.000044718465],"about_ca_topic_score_codex":0.0018996999,"about_ca_topic_score_gemma":0.0018180558,"teacher_disagreement_score":0.007037996,"about_ca_system_score_codex":0.0014023754,"about_ca_system_score_gemma":0.002714868,"threshold_uncertainty_score":0.037220955},"labels":[],"label_agreement":null},{"id":"W2891725784","doi":"10.1214/17-aoas1119","title":"Topological data analysis of single-trial electroencephalographic signals","year":2018,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Topological and Geometric Data Analysis","field":"Computer Science","cited_by":93,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institute of Biomedical Imaging and Bioengineering; King Abdullah University of Science and Technology; McMaster University; University of Wisconsin-Madison; Emory University; National Institutes of Health; National Science Foundation","keywords":"Electroencephalography; Pattern recognition (psychology); Computer science; Resampling; Neurophysiology; Artificial intelligence; Speech recognition; Neuroscience; Psychology","score_opus":0.16827890641685442,"score_gpt":0.3625785741721692,"score_spread":0.1942996677553148,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2891725784","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.5351376,0.00034935903,0.46070278,0.00018837306,0.000052416373,0.00010025114,0.0012950576,0.0006533746,0.0015207744],"genre_scores_gemma":[0.9443266,0.00015358526,0.053347144,0.000021258196,0.000028936134,0.00008413202,0.001683276,0.000050594375,0.00030455567],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9992362,0.00025203076,0.00007027017,0.0001696789,0.00019712553,0.00007469956],"domain_scores_gemma":[0.99530596,0.0026320026,0.0006619441,0.0006740516,0.00055462576,0.00017143205],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012633438,0.00039465743,0.00045841315,0.002723515,0.00032949867,0.00070102065,0.00052943884,0.00038915878,0.0014135379],"category_scores_gemma":[0.010536909,0.00011141963,0.0005785851,0.001517828,0.00076293346,0.0010563718,0.0008638753,0.00047487576,0.00022955048],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0015948989,0.0003283516,0.065154366,0.0011184132,0.00069520844,0.0011091076,0.0010817158,0.2287464,0.11051172,0.045102097,0.0039230073,0.5406347],"study_design_scores_gemma":[0.000029775874,0.0006138701,0.069146164,0.000041812847,0.00010928675,0.0008303504,0.0004757445,0.8630608,0.02288131,0.03917649,0.0035228122,0.00011165086],"about_ca_topic_score_codex":0.0006619696,"about_ca_topic_score_gemma":0.00071082107,"teacher_disagreement_score":0.002723515,"about_ca_system_score_codex":0.00030470363,"about_ca_system_score_gemma":0.00039706146,"threshold_uncertainty_score":0.006681323},"labels":[],"label_agreement":null},{"id":"W2949319659","doi":"10.1214/19-aoas1248","title":"Correction to: Statistical modeling and analysis of trace element concentrations in forensic glass evidence","year":2019,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Forensic and Genetic Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"Engineering and Physical Sciences Research Council","keywords":"TRACE (psycholinguistics); Forensic science; Computer science; Statistical analysis; Statistics; Data science; Archaeology; Mathematics; History; Philosophy; Linguistics","score_opus":0.05227684764868486,"score_gpt":0.3627472880181723,"score_spread":0.31047044036948745,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2949319659","genre_codex":"editorial","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0030559923,0.0021196753,0.06790641,0.03704989,0.63123333,0.00065406825,0.22153962,0.01918791,0.017253151],"genre_scores_gemma":[0.09668558,0.006339001,0.19903745,0.036136847,0.0586496,0.0032436794,0.24669991,0.04779256,0.30541536],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.98930234,0.0020898443,0.0013756502,0.0012819576,0.00533518,0.00061490614],"domain_scores_gemma":[0.8546616,0.04087999,0.0041335905,0.01269616,0.08581246,0.0018161114],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010918233,0.0022137056,0.0017811236,0.0058750277,0.0024157062,0.003776597,0.0049775606,0.0035258806,0.23682193],"category_scores_gemma":[0.21656552,0.001542244,0.0026087067,0.00822988,0.0013440165,0.0028583547,0.0031137767,0.00717818,0.08684257],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000050506165,0.000006989036,0.00026518916,0.00014627745,0.000020880516,0.000038119,0.000030368084,0.00029089634,0.00006605527,0.0007666793,0.99002355,0.0082944445],"study_design_scores_gemma":[0.00009408644,0.000030983898,0.0040266546,0.00059197564,0.000071448565,0.00025742623,0.00012380822,0.0027216119,0.0007064863,0.005913905,0.9853681,0.00009346721],"about_ca_topic_score_codex":0.06556838,"about_ca_topic_score_gemma":0.07234015,"teacher_disagreement_score":0.23682193,"about_ca_system_score_codex":0.0055475887,"about_ca_system_score_gemma":0.011624022,"threshold_uncertainty_score":0.7922483},"labels":[],"label_agreement":null},{"id":"W2951194651","doi":"10.1214/18-aoas1236","title":"Early diagnosis of neurological disease using peak degeneration ages of multiple biomarkers","year":2019,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institute of Biomedical Imaging and Bioengineering; Canadian Institutes of Health Research; National Institutes of Health; Genentech; National Institute of Neurological Disorders and Stroke; IXICO; National Institute of General Medical Sciences; H. Lundbeck A/S; Servier; Eisai; Northern California Institute for Research and Education; Pfizer; Biogen; BioClinica; F. Hoffmann-La Roche; University of Southern California; Eli Lilly and Company; U.S. Department of Defense; Meso Scale Diagnostics; Alzheimer's Disease Neuroimaging Initiative; Novartis Pharmaceuticals Corporation; Bristol-Myers Squibb; National Institute on Aging; Alzheimer's Association; Foundation for the National Institutes of Health","keywords":"Biomarker; Disease; Medicine; Pathology; Biology","score_opus":0.21261294380341322,"score_gpt":0.3913457883170323,"score_spread":0.1787328445136191,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2951194651","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.2001934,0.0024209863,0.7943785,0.0012004597,0.00007466423,0.00010168853,0.0005121573,0.00028292564,0.0008351467],"genre_scores_gemma":[0.7838752,0.0010850285,0.21319069,0.00014799646,0.000057618803,0.000107972235,0.0004297766,0.000035252182,0.0010703998],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9989587,0.0005586423,0.000070997325,0.00026978028,0.000089521294,0.00005232254],"domain_scores_gemma":[0.99023587,0.0073513053,0.0011772866,0.0005105428,0.0005082926,0.00021666102],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0058360584,0.00073524343,0.0009616255,0.0013249872,0.00038003104,0.0010425772,0.00082550687,0.0010884617,0.00069550023],"category_scores_gemma":[0.020767415,0.00032908964,0.00087461603,0.0009553618,0.0007255442,0.0015472679,0.0011903419,0.0014108422,0.00021887179],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012994909,0.00032013218,0.34344015,0.00055937254,0.00073234725,0.00076502014,0.00090357865,0.276686,0.009434381,0.03357984,0.0022112255,0.3300685],"study_design_scores_gemma":[0.00007762598,0.00040099732,0.055755362,0.00021859535,0.00021497867,0.0007148015,0.00022964641,0.8604091,0.007846566,0.07168072,0.002341213,0.00011049514],"about_ca_topic_score_codex":0.0021895398,"about_ca_topic_score_gemma":0.0030242587,"teacher_disagreement_score":0.0058360584,"about_ca_system_score_codex":0.0007001437,"about_ca_system_score_gemma":0.0011547392,"threshold_uncertainty_score":0.030864358},"labels":[],"label_agreement":null},{"id":"W2959430680","doi":"10.1214/21-aoas1457","title":"Markov-switching state space models for uncovering musical interpretation","year":2021,"lang":"en","type":"preprint","venue":"The Annals of Applied Statistics","topic":"Music and Audio Processing","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Interpretation (philosophy); Musical; Performing arts; Computer science; Active listening; Space (punctuation); Cognitive psychology; Subject (documents); Improvisation; Psychology; Linguistics; Cognitive science; Communication; Visual arts; Art; World Wide Web","score_opus":0.06173092111457535,"score_gpt":0.31619146310697216,"score_spread":0.2544605419923968,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2959430680","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03783044,0.0006001633,0.9573591,0.0007032563,0.00006095793,0.000100304525,0.0008601243,0.00041284983,0.002072769],"genre_scores_gemma":[0.8636342,0.0013457021,0.12156057,0.000348175,0.00024143104,0.00078835257,0.003261897,0.00017389144,0.008645769],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9977435,0.00125506,0.00009470202,0.0005115139,0.00022612698,0.00016920081],"domain_scores_gemma":[0.97617745,0.021342903,0.0011076966,0.00066682894,0.0004918037,0.00021336235],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0053588697,0.0015296878,0.0015773021,0.0023829092,0.00071465806,0.0018154761,0.0018953335,0.001830562,0.004864873],"category_scores_gemma":[0.01988769,0.0008005934,0.0016842922,0.0017760524,0.0022772027,0.002312405,0.0015773035,0.003229479,0.0009423712],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00023839291,0.00015914396,0.008211746,0.0001783093,0.00020516268,0.000168754,0.0004385808,0.81786203,0.000791308,0.14172661,0.0018801711,0.02813983],"study_design_scores_gemma":[0.0000141589835,0.000026811284,0.000672444,0.000018792754,0.000019009238,0.0000130188255,0.00003490181,0.9308464,0.00011057225,0.06775058,0.00047134177,0.000022000624],"about_ca_topic_score_codex":0.011850726,"about_ca_topic_score_gemma":0.012820398,"teacher_disagreement_score":0.011850726,"about_ca_system_score_codex":0.0021578039,"about_ca_system_score_gemma":0.0011217212,"threshold_uncertainty_score":0.028340757},"labels":[],"label_agreement":null},{"id":"W2962800545","doi":"10.1214/16-aoas993","title":"A continuous-time stochastic block model for basketball networks","year":2017,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":38,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Acadia University; University of Waterloo; Royal Bank of Canada","funders":"","keywords":"Basketball; Analytics; Computer science; Stochastic block model; Data science; Offensive; Block (permutation group theory); Cluster analysis; Machine learning; Artificial intelligence; Operations research; Engineering; Geography; Mathematics","score_opus":0.09067351114725328,"score_gpt":0.28776109313606096,"score_spread":0.1970875819888077,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2962800545","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.051223435,0.0016674288,0.92800915,0.0022988834,0.00037239568,0.00022906403,0.0035863817,0.00052305055,0.012090091],"genre_scores_gemma":[0.85691506,0.0044008577,0.06083363,0.00073610095,0.0005141386,0.0011722913,0.0049193255,0.00033496632,0.070173636],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988268,0.00038604075,0.00003745214,0.00033103937,0.00016024163,0.00025831404],"domain_scores_gemma":[0.9965952,0.002022437,0.00051599543,0.00011812577,0.00043450695,0.00031377096],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024591591,0.001945377,0.0022614466,0.0013072598,0.00094654295,0.0027992588,0.0039680293,0.0027140055,0.014827894],"category_scores_gemma":[0.0069850073,0.0011441197,0.0016496597,0.0015582964,0.0018869101,0.0028003545,0.0018766106,0.0033060906,0.0027969275],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00015637257,0.000060592713,0.002229649,0.000110792265,0.000070962786,0.00018364371,0.00013679583,0.8541661,0.00080634496,0.13145512,0.004206965,0.006416624],"study_design_scores_gemma":[0.000025163949,0.000024109908,0.00030423666,0.000015848047,0.000021324497,0.000023094955,0.000023021417,0.97899616,0.000053732107,0.019296635,0.0012001523,0.0000166487],"about_ca_topic_score_codex":0.04532421,"about_ca_topic_score_gemma":0.028568959,"teacher_disagreement_score":0.04532421,"about_ca_system_score_codex":0.002952008,"about_ca_system_score_gemma":0.0018186017,"threshold_uncertainty_score":0.09012079},"labels":[],"label_agreement":null},{"id":"W2983854231","doi":"10.1214/21-aoas1565","title":"Measurement error correction in particle tracking microrheology","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Streptococcal Infections and Treatments","field":"Medicine","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta; University of Waterloo","funders":"","keywords":"Microrheology; Estimator; Range (aeronautics); Computer science; Power law; Bandwidth (computing); Viscoelasticity; Algorithm; Statistical physics; Physics; Acoustics; Mathematics; Statistics; Engineering; Telecommunications; Aerospace engineering","score_opus":0.15190217414061702,"score_gpt":0.36921828068977763,"score_spread":0.21731610654916061,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2983854231","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0035183532,0.00026640887,0.9956911,0.00008526901,0.000055121473,0.000017014228,0.000017841747,0.00016816576,0.00018078394],"genre_scores_gemma":[0.20777501,0.0010001247,0.78809935,0.00020745219,0.00017945687,0.00020696272,0.00021609853,0.00026226768,0.0020533549],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99775165,0.0008679257,0.00015422779,0.00040570577,0.00072642294,0.00009405063],"domain_scores_gemma":[0.99131197,0.0064834272,0.0005766279,0.0007280964,0.00081145973,0.00008841859],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005665807,0.0010191251,0.0010086889,0.0009460809,0.00061358535,0.0008167314,0.001233597,0.0015213034,0.00084588473],"category_scores_gemma":[0.02418296,0.00047327153,0.0006313093,0.0011080023,0.0011541685,0.0009959642,0.0012274017,0.0015593101,0.0004915105],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00051298464,0.00010981199,0.0042824745,0.0005796224,0.00023630726,0.0002718363,0.00037524354,0.50048757,0.031171603,0.06617672,0.0027879367,0.39300784],"study_design_scores_gemma":[0.00001797055,0.0000818593,0.00084870437,0.000039616032,0.000027092727,0.00012753841,0.000023524384,0.9698821,0.013483847,0.012073403,0.003361371,0.000032954165],"about_ca_topic_score_codex":0.0032895757,"about_ca_topic_score_gemma":0.002937318,"teacher_disagreement_score":0.005665807,"about_ca_system_score_codex":0.00068464095,"about_ca_system_score_gemma":0.0016238673,"threshold_uncertainty_score":0.02996403},"labels":[],"label_agreement":null},{"id":"W2990428857","doi":"10.1214/19-aoas1274","title":"Joint model of accelerated failure time and mechanistic nonlinear model for censored covariates, with application in HIV/AIDS","year":2019,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; City University of New York; National Science Foundation","keywords":"Covariate; Proportional hazards model; Accelerated failure time model; Econometrics; Inference; Statistics; Survival analysis; Computer science; Outcome (game theory); Mathematics; Artificial intelligence","score_opus":0.08529347782029319,"score_gpt":0.3467134122240529,"score_spread":0.2614199344037597,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2990428857","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012961065,0.00040069496,0.9847964,0.00041841465,0.00005612848,0.000058614274,0.00020792578,0.00013881722,0.00096196873],"genre_scores_gemma":[0.59703386,0.0021041387,0.37603536,0.00057553535,0.00039414564,0.0010905692,0.0010808178,0.00023100156,0.02145456],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9958144,0.0025965986,0.00013986348,0.00070694933,0.00046115855,0.00028101212],"domain_scores_gemma":[0.98737144,0.009418635,0.0014240822,0.0007392115,0.0007432006,0.00030344687],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011805109,0.0013448892,0.0020195607,0.0014676207,0.000756833,0.0016077422,0.0036634754,0.0022722073,0.0041671745],"category_scores_gemma":[0.024891395,0.0010412749,0.002600108,0.0017877469,0.0023425182,0.002873084,0.0021917892,0.0032355986,0.00067431194],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018491603,0.00008512332,0.005823607,0.0002305283,0.0002268632,0.00046554214,0.00042476816,0.52689517,0.0007841106,0.43772626,0.0014197372,0.025733437],"study_design_scores_gemma":[0.000033730892,0.000069814756,0.00081253634,0.000023349226,0.00006567872,0.00011951901,0.000031832322,0.8978777,0.00015952998,0.09891262,0.0018644186,0.000029323968],"about_ca_topic_score_codex":0.010161136,"about_ca_topic_score_gemma":0.007418862,"teacher_disagreement_score":0.011805109,"about_ca_system_score_codex":0.0017029664,"about_ca_system_score_gemma":0.0022243455,"threshold_uncertainty_score":0.06243211},"labels":[],"label_agreement":null},{"id":"W2990681736","doi":"10.1214/19-aoas1269","title":"Robust elastic net estimators for variable selection and identification of proteomic biomarkers","year":2019,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":35,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"University of British Columbia; Genome Canada","keywords":"Estimator; Outlier; Robustness (evolution); Feature selection; Computer science; Robust statistics; Robust regression; Elastic net regularization; Lasso (programming language); Mathematics; Statistics; Artificial intelligence; Biology","score_opus":0.12250591961529245,"score_gpt":0.3935837860991451,"score_spread":0.2710778664838527,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2990681736","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0033812427,0.00015327525,0.99604577,0.000088953486,0.00001753667,0.000023165434,0.00003939555,0.000101049285,0.00014963157],"genre_scores_gemma":[0.26743823,0.0009144762,0.72505623,0.00035075302,0.00026061028,0.000788148,0.0010347851,0.0003161257,0.0038406344],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99743193,0.0016317945,0.00011830931,0.0003267002,0.00038769827,0.000103533275],"domain_scores_gemma":[0.9854005,0.011407626,0.0012578923,0.0007924785,0.00097317115,0.0001682555],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009866947,0.0015396139,0.0013474707,0.0020165087,0.00049997424,0.0010772665,0.0018862707,0.0014263509,0.0017773693],"category_scores_gemma":[0.028213376,0.00072772993,0.0011278372,0.0013599395,0.0015245983,0.001805266,0.0018291606,0.0023894531,0.00055838947],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002059577,0.00009523383,0.0026118567,0.00017019818,0.00021446725,0.000116048366,0.000055468405,0.8771178,0.001956559,0.041982386,0.0019178183,0.073556185],"study_design_scores_gemma":[0.000010600341,0.000022029153,0.00023377954,0.000011438115,0.000008350685,0.00001681483,0.0000043795767,0.98454696,0.00043033613,0.014301907,0.00040235452,0.000011058409],"about_ca_topic_score_codex":0.0016352237,"about_ca_topic_score_gemma":0.0012882905,"teacher_disagreement_score":0.009866947,"about_ca_system_score_codex":0.0008064747,"about_ca_system_score_gemma":0.0011274774,"threshold_uncertainty_score":0.05218208},"labels":[],"label_agreement":null},{"id":"W3018701121","doi":"10.1214/20-aoas1348","title":"Markov decision processes with dynamic transition probabilities: An analysis of shooting strategies in basketball","year":2020,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":27,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Basketball; Transition (genetics); Markov chain; Computer science; Machine learning; History; Chemistry","score_opus":0.051517805710940424,"score_gpt":0.2760733188156001,"score_spread":0.22455551310465968,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3018701121","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.4325787,0.0006084729,0.5558171,0.0012002803,0.00006252014,0.00014733832,0.00038124042,0.0001623071,0.009041965],"genre_scores_gemma":[0.9781797,0.0003545569,0.01573916,0.00009358401,0.000023392633,0.00009302937,0.00015652327,0.00003955713,0.005320502],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99926215,0.00028322992,0.000027484028,0.00012689496,0.00013012068,0.00017009515],"domain_scores_gemma":[0.9951794,0.0035928758,0.0005851621,0.000121483645,0.00022977973,0.00029131572],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002411613,0.000701418,0.0009523337,0.0008608297,0.00053564546,0.0014056152,0.0014811857,0.0011939616,0.0047204294],"category_scores_gemma":[0.009307036,0.00063323975,0.000862905,0.00067941996,0.0012129338,0.0017533554,0.0010911204,0.0018290875,0.0002725462],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00004867548,0.000038583734,0.0023017726,0.000023668052,0.000037315236,0.000104892766,0.000097199794,0.92942446,0.00043009518,0.064034455,0.0003509364,0.0031078907],"study_design_scores_gemma":[0.00000801619,0.000018271397,0.0004586357,0.000006103132,0.0000073860365,0.000011958997,0.00003214234,0.9818886,0.00006531749,0.017311513,0.0001845108,0.0000076245215],"about_ca_topic_score_codex":0.017468506,"about_ca_topic_score_gemma":0.01174392,"teacher_disagreement_score":0.017468506,"about_ca_system_score_codex":0.0016042205,"about_ca_system_score_gemma":0.0015013946,"threshold_uncertainty_score":0.034733653},"labels":[],"label_agreement":null},{"id":"W3021847850","doi":"10.1214/21-aoas1526","title":"Permutation tests under a rotating sampling plan with clustered data","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Forest ecology and management","field":"Environmental Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Permutation (music); Sampling (signal processing); Resampling; Statistics; Parametric statistics; Inference; Econometrics; Sampling distribution; Computer science; Mathematics; Artificial intelligence","score_opus":0.12410127655508822,"score_gpt":0.3257595050175824,"score_spread":0.20165822846249415,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3021847850","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.072907306,0.0001094887,0.9230503,0.00035830872,0.00011433639,0.000842845,0.0005373936,0.00064141466,0.0014386674],"genre_scores_gemma":[0.5067963,0.00014097718,0.48627344,0.00038453066,0.0001721444,0.00284942,0.0014929438,0.00015459425,0.0017357738],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8424625,0.11817869,0.005025911,0.022357816,0.009633167,0.0023419615],"domain_scores_gemma":[0.6207936,0.27150548,0.029902361,0.06574157,0.010015836,0.0020412412],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.09458126,0.0017499635,0.0028177036,0.0023582142,0.0019505509,0.0026519785,0.00669221,0.003017192,0.0056971377],"category_scores_gemma":[0.3216661,0.0012492518,0.0036208902,0.0040749847,0.0063514677,0.005143799,0.002987018,0.0043668333,0.0011445016],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0051267906,0.0009910834,0.060845852,0.0007100078,0.003074319,0.0020452503,0.0021686677,0.24366301,0.0064746253,0.36922738,0.0064750295,0.2991981],"study_design_scores_gemma":[0.00069335924,0.0021305298,0.01450444,0.00009619241,0.0003608525,0.00046850607,0.00040557826,0.8398683,0.0033085577,0.13509318,0.0029001038,0.00017040524],"about_ca_topic_score_codex":0.0034831828,"about_ca_topic_score_gemma":0.0019880391,"teacher_disagreement_score":0.09458126,"about_ca_system_score_codex":0.0020852834,"about_ca_system_score_gemma":0.0040444783,"threshold_uncertainty_score":0.5001995},"labels":[],"label_agreement":null},{"id":"W3037512058","doi":"10.1214/20-aoas1331","title":"Focused model selection for linear mixed models with an application to whale ecology","year":2020,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Toronto Metropolitan University","funders":"Norges Forskningsråd","keywords":"Whaling; Generalized linear mixed model; Model selection; Mixed model; Estimator; Linear model; Selection (genetic algorithm); Computer science; Whale; Information Criteria; Ecology; Econometrics; Mathematics; Statistics; Artificial intelligence; Machine learning; Biology","score_opus":0.19487171623464805,"score_gpt":0.40556105054850106,"score_spread":0.210689334313853,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3037512058","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0006595029,0.00042713585,0.9981927,0.00027140413,0.000026559726,0.00002590506,0.000033303637,0.00006351927,0.0002999003],"genre_scores_gemma":[0.05881895,0.001814671,0.93515515,0.00064173125,0.0005088179,0.00085013837,0.00041119868,0.000309546,0.0014897602],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.97538495,0.021030663,0.0006109961,0.0011889193,0.0015172403,0.00026716533],"domain_scores_gemma":[0.88856083,0.10335989,0.0023438435,0.0026627511,0.0026608284,0.0004119197],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03598071,0.0018901376,0.0023356748,0.0040169097,0.0011163338,0.0023754882,0.003364263,0.0025275575,0.0040904097],"category_scores_gemma":[0.104575515,0.0011931816,0.0034391545,0.0031166049,0.0039742473,0.003305654,0.0043372433,0.004864979,0.00083234115],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000077141216,0.0000626704,0.0015718366,0.0006238105,0.0004832358,0.0003498229,0.0005830312,0.09062429,0.0010275358,0.824905,0.0038133264,0.075878434],"study_design_scores_gemma":[0.000024143681,0.000086115244,0.0003880698,0.00014208595,0.000069295456,0.0001212413,0.000052806216,0.3555411,0.0005164742,0.63701326,0.006000423,0.000044924185],"about_ca_topic_score_codex":0.0029494548,"about_ca_topic_score_gemma":0.0029938307,"teacher_disagreement_score":0.03598071,"about_ca_system_score_codex":0.0019647942,"about_ca_system_score_gemma":0.0022316247,"threshold_uncertainty_score":0.19028652},"labels":[],"label_agreement":null},{"id":"W3106574156","doi":"10.1214/23-aoas1737","title":"Predictive inference for travel time on transportation networks","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Traffic Prediction and Management Techniques","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"HEC Montréal; University of Toronto","funders":"","keywords":"Computer science; Predictive inference; Population; Global Positioning System; Inference; Statistical inference; Statistics; Data mining; Mathematics; Artificial intelligence; Bayesian inference","score_opus":0.03528464756629771,"score_gpt":0.2870721727432698,"score_spread":0.25178752517697206,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3106574156","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.024761612,0.00054909824,0.9711415,0.00047313244,0.000052351188,0.000067225636,0.0010040342,0.0008151866,0.0011358066],"genre_scores_gemma":[0.823397,0.0018814221,0.16294679,0.00040698235,0.0004042467,0.00060552463,0.0055523743,0.000444256,0.004361364],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99729055,0.0012372251,0.00010763255,0.0007723642,0.00040851388,0.00018368886],"domain_scores_gemma":[0.9673716,0.027588457,0.0017650111,0.001749669,0.0012524448,0.00027289544],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008068961,0.001192826,0.0015223266,0.0024760256,0.00063860597,0.0017364738,0.0035453527,0.0017461468,0.0041947938],"category_scores_gemma":[0.054761745,0.001116405,0.0016495404,0.0030106914,0.0020937417,0.0035894457,0.001848155,0.0039682025,0.0007473913],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00007824516,0.000028397659,0.0045459494,0.000089379915,0.000086696185,0.00006739496,0.000114049464,0.92030555,0.0002069092,0.052771095,0.001784849,0.019921582],"study_design_scores_gemma":[0.0000074435925,0.0000067201863,0.0005232242,0.000021685373,0.0000091178445,0.000012679668,0.000014073071,0.96558905,0.00011113753,0.03331475,0.00038258094,0.000007576821],"about_ca_topic_score_codex":0.023599422,"about_ca_topic_score_gemma":0.014253051,"teacher_disagreement_score":0.023599422,"about_ca_system_score_codex":0.0024967159,"about_ca_system_score_gemma":0.001274509,"threshold_uncertainty_score":0.046924114},"labels":[],"label_agreement":null},{"id":"W3168647111","doi":"10.1214/22-aoas1726","title":"Evaluating the use of generalized dynamic weighted ordinary least squares for individualized HIV treatment strategies","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Agence Nationale de la Recherche","keywords":"Categorical variable; Estimator; Ordinary least squares; Human immunodeficiency virus (HIV); Mathematics; Population; Mathematical optimization; Robustness (evolution); Computer science; Medicine; Statistics; Econometrics; Algorithm; Immunology","score_opus":0.6063138857192424,"score_gpt":0.5398848109691714,"score_spread":0.06642907475007098,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3168647111","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.14368768,0.0020325708,0.8480628,0.0011827071,0.00009305489,0.00038661915,0.00019333494,0.0003858762,0.0039754095],"genre_scores_gemma":[0.7748069,0.00051420496,0.22278605,0.00025374134,0.00004696349,0.00035040893,0.00022833556,0.00009114848,0.00092229026],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98791933,0.009859153,0.0003396042,0.00065290456,0.0010186582,0.00021028607],"domain_scores_gemma":[0.9040384,0.08937324,0.002472102,0.0017142475,0.001960357,0.00044160118],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.022166196,0.0012566862,0.0016210852,0.0013342967,0.0004245512,0.0015611709,0.0015386641,0.0016709714,0.0017130027],"category_scores_gemma":[0.08644311,0.00064995646,0.0009513054,0.0010198788,0.0010849491,0.0017106527,0.0013684384,0.0014964275,0.00018450763],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003384024,0.00013317872,0.004176367,0.00016740507,0.00034071843,0.000039186376,0.0000779288,0.9277818,0.0003454479,0.010907743,0.00044000102,0.055251654],"study_design_scores_gemma":[0.000036689667,0.00018463007,0.00045987786,0.000022893339,0.000040584695,0.000011648887,0.000025946129,0.99175465,0.00025390342,0.00694284,0.0002562743,0.000010106541],"about_ca_topic_score_codex":0.0058653215,"about_ca_topic_score_gemma":0.005556967,"teacher_disagreement_score":0.022166196,"about_ca_system_score_codex":0.0015260865,"about_ca_system_score_gemma":0.0027872545,"threshold_uncertainty_score":0.117227435},"labels":[],"label_agreement":null},{"id":"W3174385715","doi":"10.1214/21-aoas1570","title":"Estimation of the marginal effect of antidepressants on body mass index under confounding and endogenous covariate-driven monitoring times","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Covariate; Endogeneity; Confounding; Statistics; Body mass index; Propensity score matching; Marginal structural model; Econometrics; Selection bias; Index (typography); Medicine; Computer science; Mathematics; Internal medicine","score_opus":0.17584202781014677,"score_gpt":0.41534126538013316,"score_spread":0.2394992375699864,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3174385715","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.081939064,0.0030075342,0.9105866,0.0013649974,0.00016968344,0.00028910846,0.0009431789,0.00020079102,0.0014990692],"genre_scores_gemma":[0.6937077,0.004872694,0.2925583,0.0008060886,0.0004710756,0.0011112794,0.00176031,0.00011577654,0.004596738],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.98673356,0.009455225,0.00064188836,0.0019386344,0.00084363687,0.00038699323],"domain_scores_gemma":[0.8961616,0.09132943,0.0053975764,0.005286582,0.0014735374,0.0003512397],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.029537143,0.0010824157,0.0020344404,0.0014464841,0.00045296372,0.0016448277,0.002583616,0.001810678,0.0036884556],"category_scores_gemma":[0.12355511,0.00075036724,0.0034128423,0.0025237242,0.0018932772,0.0026150406,0.0019396594,0.0024443285,0.00037432194],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011324373,0.00035718826,0.15997376,0.0021846408,0.0055319634,0.001007261,0.0015445647,0.20340107,0.0017577886,0.35106912,0.0034839169,0.26855627],"study_design_scores_gemma":[0.0002779999,0.0006854186,0.052843165,0.00059706415,0.0025030815,0.0005724834,0.00028535913,0.5066868,0.0022196437,0.42273536,0.010472121,0.000121491736],"about_ca_topic_score_codex":0.007760918,"about_ca_topic_score_gemma":0.0045684464,"teacher_disagreement_score":0.029537143,"about_ca_system_score_codex":0.0013295321,"about_ca_system_score_gemma":0.0026861222,"threshold_uncertainty_score":0.15620923},"labels":[],"label_agreement":null},{"id":"W3181775732","doi":"10.1214/20-aoas1423","title":"A compositional model to assess expression changes from single-cell RNA-seq data","year":2021,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"National Institute of Allergy and Infectious Diseases; National Institute of General Medical Sciences; National Science Foundation; National Institutes of Health; National Cancer Institute; National Institute of Dental and Craniofacial Research; University of Wisconsin-Madison","keywords":"Cluster analysis; Bayesian probability; Expression (computer science); Mixture model; Computer science; Prior probability; Hierarchical clustering; Posterior probability; Probability distribution; Computational biology; Mathematics; Data mining; Artificial intelligence; Statistics; Biology","score_opus":0.21491683278284487,"score_gpt":0.3373738237349338,"score_spread":0.12245699095208892,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3181775732","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.014670985,0.00008960123,0.984447,0.00008964345,0.000010394622,0.000038798757,0.00012873717,0.0001715274,0.00035328793],"genre_scores_gemma":[0.41514093,0.00054224295,0.57872474,0.00031385428,0.000099860365,0.0005512266,0.001362901,0.00026077288,0.0030034077],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984067,0.000682727,0.00007695166,0.00033401072,0.00040346407,0.000096069576],"domain_scores_gemma":[0.9941158,0.004108523,0.000608767,0.00056109147,0.00044361743,0.00016218382],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006643904,0.0008264289,0.0011121051,0.001569165,0.0005632261,0.001531622,0.0021205815,0.0013341074,0.0019120004],"category_scores_gemma":[0.017691338,0.0006849523,0.001191453,0.0012540383,0.0014441155,0.0020153718,0.0019593253,0.0016576508,0.00075578113],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00034593235,0.00012654276,0.008762974,0.00032154802,0.00022476951,0.00021715362,0.00038264165,0.71376634,0.03716441,0.15302074,0.001417962,0.08424908],"study_design_scores_gemma":[0.000008958033,0.000032969147,0.00080167496,0.00000891763,0.000018590155,0.00005064831,0.000016657785,0.9535383,0.0017140736,0.043273207,0.0005166157,0.000019361072],"about_ca_topic_score_codex":0.0030921632,"about_ca_topic_score_gemma":0.0038234685,"teacher_disagreement_score":0.006643904,"about_ca_system_score_codex":0.001019839,"about_ca_system_score_gemma":0.0012688695,"threshold_uncertainty_score":0.03513676},"labels":[],"label_agreement":null},{"id":"W3191733181","doi":"10.1214/21-aoas1501","title":"The ASA president’s task force statement on statistical significance and replicability","year":2021,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":41,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Task force; Statement (logic); Task (project management); Statistical analysis; Statistics; Computer science; Econometrics; Natural language processing; Mathematics; Political science; Law; Management; Economics","score_opus":0.5158428041741353,"score_gpt":0.5546070747217429,"score_spread":0.03876427054760756,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3191733181","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0015366148,0.041681185,0.18402164,0.46947026,0.243812,0.008840113,0.0048821275,0.0016104938,0.044145606],"genre_scores_gemma":[0.029673824,0.041351333,0.33849984,0.35213616,0.17285727,0.03843954,0.0037820197,0.002105745,0.021154247],"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.5791425,0.24945755,0.06759895,0.0124815265,0.086480774,0.004838741],"domain_scores_gemma":[0.2028281,0.5724922,0.037598338,0.049584284,0.13058175,0.0069154254],"candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.48481098,0.003085638,0.007781122,0.010247799,0.006377546,0.01495982,0.006354643,0.031677775,0.013417003],"category_scores_gemma":[0.6991994,0.004084573,0.0076310793,0.008229179,0.018795019,0.007347844,0.005701226,0.04422818,0.01791636],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00087022776,0.00014789547,0.0011906213,0.003292616,0.00052114594,0.00035184214,0.0011350206,0.00094217696,0.00070092274,0.06591138,0.8377728,0.087163374],"study_design_scores_gemma":[0.0008507982,0.00047585706,0.0032779258,0.014006471,0.0006347808,0.0009325843,0.00033672308,0.0035585838,0.0014122863,0.13350499,0.84058523,0.00042379214],"about_ca_topic_score_codex":0.004157059,"about_ca_topic_score_gemma":0.0035823802,"teacher_disagreement_score":0.51518905,"about_ca_system_score_codex":0.0050111115,"about_ca_system_score_gemma":0.035747044,"threshold_uncertainty_score":0.6353198},"labels":[],"label_agreement":null},{"id":"W4220666549","doi":"10.1214/21-aoas1499","title":"Bayesian adjustment for preferential testing in estimating infection fatality rates, as motivated by the COVID-19 pandemic","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Institute of Genetics; Canadian Institutes of Health Research; European Commission","keywords":"Coronavirus disease 2019 (COVID-19); Bayesian probability; Pandemic; Statistics; Case fatality rate; Econometrics; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Identifiability; Demography; 2019-20 coronavirus outbreak; Population; Statistical hypothesis testing; Mathematics; Geography; Actuarial science; Medicine; Economics; Virology; Sociology; Outbreak","score_opus":0.5179263313108267,"score_gpt":0.4970631588051286,"score_spread":0.02086317250569808,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4220666549","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04845202,0.0010670678,0.9445403,0.0028451562,0.00016766941,0.00025058378,0.00022370498,0.00025331817,0.0022002286],"genre_scores_gemma":[0.628529,0.00078474346,0.36532876,0.0015877074,0.00028068258,0.00064274686,0.00062375626,0.00013671629,0.0020859055],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.88210773,0.105373636,0.0022727887,0.005405396,0.0036858192,0.0011546196],"domain_scores_gemma":[0.6324457,0.3276676,0.015097243,0.017826997,0.0058094007,0.0011531692],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.13482149,0.0012538675,0.0024668903,0.0019487026,0.0013731859,0.0028416421,0.004420705,0.0029998089,0.0019600964],"category_scores_gemma":[0.3909401,0.0014802121,0.0024091983,0.0027654436,0.0055601983,0.004015801,0.0036415707,0.005745742,0.0003064639],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0014005469,0.00019039099,0.08502714,0.00071157695,0.0024303796,0.001119086,0.0019060284,0.38573918,0.0021023257,0.3593467,0.006629892,0.15339684],"study_design_scores_gemma":[0.00014540953,0.00037565164,0.016883148,0.00028824058,0.0003213401,0.0006612658,0.0002143289,0.6568577,0.0013710634,0.31854194,0.0042002657,0.00013953152],"about_ca_topic_score_codex":0.012879487,"about_ca_topic_score_gemma":0.011022559,"teacher_disagreement_score":0.13482149,"about_ca_system_score_codex":0.002331473,"about_ca_system_score_gemma":0.003190097,"threshold_uncertainty_score":0.7130127},"labels":[],"label_agreement":null},{"id":"W4221109862","doi":"10.1214/21-aoas1518","title":"Accounting for drop-out using inverse probability censoring weights in longitudinal clustered data with informative cluster size","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Censoring (clinical trials); Statistics; Inverse probability; Inference; Generalized estimating equation; Mathematics; Random effects model; Estimating equations; Marginal structural model; Econometrics; Causal inference; Computer science; Medicine; Estimator; Artificial intelligence; Meta-analysis","score_opus":0.16111533411320592,"score_gpt":0.35644827601533313,"score_spread":0.19533294190212722,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4221109862","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04520948,0.00043208376,0.9528109,0.00047667368,0.00004132887,0.00014148586,0.0001741428,0.00026759825,0.0004464088],"genre_scores_gemma":[0.63819563,0.0007335305,0.35752437,0.00039553872,0.00010612623,0.00052615395,0.0007896649,0.00013291449,0.0015960215],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9862845,0.010589731,0.00046347122,0.0014169585,0.0008502156,0.00039519224],"domain_scores_gemma":[0.917083,0.06802259,0.0052231313,0.0069272467,0.0021205996,0.0006234881],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.038128514,0.0009916291,0.001727818,0.0022263085,0.0009447747,0.0015909638,0.003099265,0.0023403617,0.0013019671],"category_scores_gemma":[0.107896626,0.0008018413,0.0020323754,0.0024914192,0.0015849061,0.0023214633,0.0025160918,0.002944033,0.00029044593],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0006134006,0.0003138929,0.083761945,0.00038644832,0.0010464978,0.0005674789,0.0013279546,0.56252396,0.0012777618,0.094542824,0.003035865,0.250602],"study_design_scores_gemma":[0.00006771568,0.00013610834,0.0069812094,0.0001043341,0.00018719093,0.0001417794,0.00014133903,0.91387206,0.00081809604,0.07555223,0.0019312683,0.000066735505],"about_ca_topic_score_codex":0.0119699035,"about_ca_topic_score_gemma":0.012868774,"teacher_disagreement_score":0.038128514,"about_ca_system_score_codex":0.0011102754,"about_ca_system_score_gemma":0.0018596975,"threshold_uncertainty_score":0.20164526},"labels":[],"label_agreement":null},{"id":"W4287780226","doi":"10.1214/22-aoas1701","title":"The scalable birth–death MCMC algorithm for mixed graphical model learning with application to genomic data integration","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; York University; Lunenfeld-Tanenbaum Research Institute; University of New Brunswick","funders":"National Center for Advancing Translational Sciences; National Cancer Institute","keywords":"Subtyping; Lasso (programming language); Computer science; Inference; Scalability; Biological data; Feature selection; Computational biology; Data mining; Machine learning; Artificial intelligence; Bioinformatics; Biology","score_opus":0.062091907064135005,"score_gpt":0.33707138606872744,"score_spread":0.2749794790045924,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4287780226","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0022819138,0.00021320833,0.996349,0.00024356681,0.00003225165,0.00006386393,0.00008332733,0.00046515252,0.00026760198],"genre_scores_gemma":[0.116347134,0.0003950732,0.8784335,0.00050922384,0.00015658214,0.00094655453,0.00094327965,0.00034197196,0.0019267246],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99761117,0.001429433,0.00010335307,0.0002986391,0.0004214431,0.00013589562],"domain_scores_gemma":[0.9877264,0.009960955,0.0004829379,0.00052456016,0.0009426751,0.0003624695],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005941883,0.0013801043,0.0019297034,0.0019337842,0.0012899891,0.0015650316,0.003856547,0.002363866,0.0038765692],"category_scores_gemma":[0.019185277,0.0013373285,0.0020438938,0.002088329,0.001408291,0.0012386021,0.0030071249,0.004579017,0.0010318378],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016677135,0.000078237914,0.0023454123,0.00015648769,0.00023795146,0.00018951092,0.0001353782,0.860866,0.00094877166,0.042544115,0.0038915072,0.08843993],"study_design_scores_gemma":[0.000013122438,0.000006912626,0.00004300918,0.0000041742396,0.0000049831146,0.000009405331,0.0000027454207,0.992645,0.000082513194,0.006857586,0.00032538507,0.000005056343],"about_ca_topic_score_codex":0.018476965,"about_ca_topic_score_gemma":0.018889679,"teacher_disagreement_score":0.018476965,"about_ca_system_score_codex":0.0018088244,"about_ca_system_score_gemma":0.0036149314,"threshold_uncertainty_score":0.036738813},"labels":[],"label_agreement":null},{"id":"W4297333613","doi":"10.1214/22-aoas1600","title":"Bayesian hierarchical random-effects meta-analysis and design of phase I clinical trials","year":2022,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"National Cancer Institute","keywords":"Meta-analysis; Random effects model; Computer science; Bayesian probability; Parametric statistics; Clinical study design; Statistics; Clinical trial; Data mining; Artificial intelligence; Mathematics; Medicine","score_opus":0.8487982915840147,"score_gpt":0.6507765650971692,"score_spread":0.19802172648684546,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4297333613","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0017367491,0.0068056164,0.9881512,0.0006869654,0.00019494185,0.0011032068,0.00033847577,0.00044655547,0.00053634663],"genre_scores_gemma":[0.11585664,0.004770738,0.870261,0.0009578235,0.0003414775,0.0063328617,0.00070501753,0.00018008437,0.00059430936],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.7909615,0.192903,0.004296406,0.005802768,0.005367464,0.0006688659],"domain_scores_gemma":[0.86643565,0.11148594,0.0077962317,0.01002133,0.0035366567,0.00072426646],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.13166603,0.0031255735,0.007433944,0.007775874,0.0009088187,0.0041700425,0.0065771886,0.003182594,0.0030620468],"category_scores_gemma":[0.217425,0.0029703705,0.01106321,0.005801319,0.002007113,0.003665709,0.0029988363,0.004657174,0.0006022953],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0036028947,0.00031346508,0.0053896965,0.014078812,0.049019612,0.00035419848,0.00048441137,0.48181498,0.0013159267,0.14853212,0.00870699,0.28638688],"study_design_scores_gemma":[0.004015224,0.0011785628,0.0026232377,0.0016945449,0.020620609,0.00023383202,0.00005205981,0.592342,0.0016815013,0.36467472,0.010570962,0.0003127771],"about_ca_topic_score_codex":0.0033128823,"about_ca_topic_score_gemma":0.00387966,"teacher_disagreement_score":0.13166603,"about_ca_system_score_codex":0.0035429108,"about_ca_system_score_gemma":0.0063757477,"threshold_uncertainty_score":0.6963248},"labels":[],"label_agreement":null},{"id":"W4320169867","doi":"10.1214/22-aoas1639","title":"Modeling panels of extremes","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Financial Risk and Volatility Modeling","field":"Economics, Econometrics and Finance","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"HEC Montréal","funders":"Natural Sciences and Engineering Research Council of Canada; Fondation HEC; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Pooling; Extreme value theory; Inference; Computer science; Regression; Econometrics; Statistics; Regression analysis; Data mining; Mathematics; Machine learning; Artificial intelligence","score_opus":0.23901888810594185,"score_gpt":0.3158408307269341,"score_spread":0.07682194262099223,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4320169867","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.061625667,0.0002629044,0.93286383,0.00039347174,0.00006266153,0.000071868446,0.0010217294,0.00040659908,0.003291258],"genre_scores_gemma":[0.8213528,0.00056214974,0.16713542,0.0004193989,0.00020991082,0.00046854976,0.003363904,0.00020327451,0.006284636],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99807274,0.0009990291,0.000049537917,0.0005167287,0.00018817319,0.00017374502],"domain_scores_gemma":[0.9944805,0.003484528,0.0006778569,0.0008735896,0.00029117055,0.00019240538],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0032424678,0.00059021043,0.0011291401,0.0009728071,0.00062567694,0.0017554417,0.001911147,0.0017679502,0.0048746057],"category_scores_gemma":[0.013388232,0.0006159064,0.0013614511,0.0014814776,0.00094121153,0.0019237128,0.0019874764,0.0023068388,0.00092692964],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011162113,0.00008729324,0.018327065,0.00008339244,0.00020995249,0.0002571521,0.0003281436,0.8099087,0.0019152955,0.12676145,0.004167259,0.037842702],"study_design_scores_gemma":[0.000014189661,0.000023783481,0.0029434727,0.000016059714,0.000021836213,0.000040696174,0.000037867943,0.9046796,0.0002700464,0.08965204,0.002281695,0.000018772773],"about_ca_topic_score_codex":0.0044075344,"about_ca_topic_score_gemma":0.0037540402,"teacher_disagreement_score":0.0048746057,"about_ca_system_score_codex":0.0006499506,"about_ca_system_score_gemma":0.0005024349,"threshold_uncertainty_score":0.017147958},"labels":[],"label_agreement":null},{"id":"W4320176121","doi":"10.1214/22-aoas1624","title":"Regularized fingerprinting in detection and attribution of climate change with weight matrix optimizing the efficiency in scaling factor estimation","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Spatial and Panel Data Analysis","field":"Economics, Econometrics and Finance","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Environment and Climate Change Canada","funders":"National Science Foundation","keywords":"Covariance matrix; Mathematics; Statistics; Estimator; Matrix (chemical analysis); Linear regression; Covariance; Estimation of covariance matrices; Applied mathematics","score_opus":0.08214424114205639,"score_gpt":0.2889005372703241,"score_spread":0.20675629612826774,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4320176121","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.015567848,0.00015011067,0.98356044,0.0001381557,0.000021224496,0.000026922116,0.000032953936,0.00016670543,0.0003356916],"genre_scores_gemma":[0.43043578,0.0005375487,0.56573915,0.0001811586,0.0001325282,0.00024114926,0.00036018103,0.00020955765,0.0021628304],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9970854,0.0014750819,0.00016866758,0.0006212664,0.00046086815,0.0001887849],"domain_scores_gemma":[0.98135424,0.013066851,0.0013771001,0.0023335423,0.0015571875,0.00031110397],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007453694,0.0012296875,0.0014321731,0.0012362807,0.00058347086,0.0013894235,0.0018314632,0.0014950943,0.0015560411],"category_scores_gemma":[0.044809867,0.00084720197,0.001061762,0.001208929,0.002104425,0.0026956878,0.0022637544,0.0020791364,0.0004591715],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00022422586,0.00008953143,0.003516536,0.00020658485,0.00014547986,0.0001472391,0.00019261484,0.83496034,0.0056579458,0.05242708,0.0010638131,0.10136854],"study_design_scores_gemma":[0.000013675691,0.000022480306,0.00035999474,0.000015290054,0.00001221768,0.000021859734,0.0000106641255,0.9812284,0.0008656382,0.017144108,0.00029062186,0.00001504346],"about_ca_topic_score_codex":0.0056439987,"about_ca_topic_score_gemma":0.0036353164,"teacher_disagreement_score":0.007453694,"about_ca_system_score_codex":0.0008521557,"about_ca_system_score_gemma":0.0018068444,"threshold_uncertainty_score":0.039419353},"labels":[],"label_agreement":null},{"id":"W4367598633","doi":"10.1214/22-aoas1656","title":"Robust joint modelling of left-censored longitudinal data and survival data with application to HIV vaccine studies","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"National Institute of Allergy and Infectious Diseases","keywords":"Outlier; Estimator; Inference; Multivariate statistics; Longitudinal data; Computer science; Statistics; Human immunodeficiency virus (HIV); Longitudinal study; Data mining; Econometrics; Mathematics; Artificial intelligence; Medicine; Immunology","score_opus":0.6228817504563653,"score_gpt":0.4455431750804737,"score_spread":0.1773385753758916,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4367598633","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0050300383,0.00034777337,0.99399924,0.0002796521,0.00002820069,0.000017942843,0.00006403056,0.000091933915,0.00014127142],"genre_scores_gemma":[0.43063226,0.002257108,0.5613516,0.00039834433,0.0005249254,0.0005874391,0.000830194,0.00024970152,0.0031684588],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.991876,0.006383824,0.00027059866,0.00062942476,0.0006040917,0.00023592313],"domain_scores_gemma":[0.9590328,0.03397471,0.0030849862,0.0021991418,0.0012559044,0.00045258546],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01680571,0.00085007306,0.0017853965,0.0014060356,0.00050434965,0.0017753046,0.002637652,0.0017798879,0.0014146159],"category_scores_gemma":[0.05212561,0.00073122006,0.0022195124,0.0019263321,0.0016032174,0.0021683532,0.002692189,0.0029599483,0.00033075886],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018990746,0.00007588778,0.0045990297,0.00017720187,0.00041457056,0.00039272997,0.00028317602,0.79960483,0.0014420479,0.15197851,0.0013503107,0.039491802],"study_design_scores_gemma":[0.000015136042,0.000039384293,0.00047980165,0.000017210303,0.000023968287,0.000041121824,0.000021789872,0.93400216,0.00023003032,0.06441785,0.00068989414,0.000021654887],"about_ca_topic_score_codex":0.0042695235,"about_ca_topic_score_gemma":0.002941712,"teacher_disagreement_score":0.9831943,"about_ca_system_score_codex":0.000780292,"about_ca_system_score_gemma":0.0016710565,"threshold_uncertainty_score":0.088878155},"labels":[],"label_agreement":null},{"id":"W4367598648","doi":"10.1214/22-aoas1675","title":"Mixed-frequency extreme value regression: Estimating the effect of mesoscale convective systems on extreme rainfall intensity","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Hydrology and Drought Analysis","field":"Environmental Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"HEC Montréal","funders":"Natural Sciences and Engineering Research Council of Canada; Fondation HEC","keywords":"Extreme value theory; Flash flood; Mesoscale meteorology; Generalized extreme value distribution; Maxima; Environmental science; Climatology; Intensity (physics); Covariate; Statistics; Mesoscale convective system; Regression analysis; Meteorology; Flood myth; Geography; Mathematics; Geology","score_opus":0.06433382728355574,"score_gpt":0.29693473832307515,"score_spread":0.2326009110395194,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4367598648","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.697387,0.0003123823,0.30004567,0.00030871617,0.000043511773,0.000076821925,0.000797452,0.00037248968,0.0006559649],"genre_scores_gemma":[0.95572746,0.000102820886,0.04245623,0.000045395806,0.00006112479,0.000094563824,0.000728899,0.000031813204,0.0007516775],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99695814,0.002122306,0.00008615766,0.0005329536,0.0001402013,0.00016016592],"domain_scores_gemma":[0.98269767,0.013605833,0.001569233,0.0013791858,0.00045521534,0.00029293474],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006658774,0.0006825888,0.0007746517,0.0012114109,0.00036738304,0.0008135073,0.0012910697,0.00083993876,0.001580149],"category_scores_gemma":[0.022865964,0.00038965215,0.0011830324,0.001179633,0.00051250745,0.00088030007,0.0008086163,0.0009659399,0.00034694074],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008461032,0.00050822453,0.6170397,0.00009351942,0.0015277324,0.0003597041,0.00027476245,0.25936675,0.0026244824,0.010001415,0.002028529,0.10532911],"study_design_scores_gemma":[0.000031277654,0.00028081145,0.06395778,0.00001843366,0.00010879528,0.00011112701,0.000092766626,0.92777,0.000861726,0.005840327,0.00089714094,0.000029801065],"about_ca_topic_score_codex":0.0072029005,"about_ca_topic_score_gemma":0.0052409144,"teacher_disagreement_score":0.0072029005,"about_ca_system_score_codex":0.0003441948,"about_ca_system_score_gemma":0.00055977283,"threshold_uncertainty_score":0.035215378},"labels":[],"label_agreement":null},{"id":"W4386514870","doi":"10.1214/23-aoas1729","title":"A Bayesian growth mixture model for complex survey data: Clustering postdisaster PTSD trajectories","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"National Institute of Environmental Health Sciences; National Institute on Aging; National Institutes of Health","keywords":"Computer science; Mixture model; Statistics; Cluster analysis; Bayesian probability; Artificial intelligence; Data mining; Mathematics","score_opus":0.41289666394850244,"score_gpt":0.45495789631561145,"score_spread":0.042061232367109014,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4386514870","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01862088,0.00023600759,0.97784346,0.00064807246,0.000050847735,0.00027967748,0.0010117195,0.0005415452,0.00076785986],"genre_scores_gemma":[0.31934008,0.0010405413,0.6631229,0.00069013954,0.00019625075,0.0027571337,0.005346641,0.0004494189,0.007056873],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99261206,0.0048235306,0.00026431508,0.0014165925,0.00057977164,0.00030363174],"domain_scores_gemma":[0.9805936,0.014136499,0.001477176,0.0019596117,0.0014850677,0.00034796732],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.018355697,0.0012763895,0.0022050405,0.0028744559,0.0012121085,0.0024472403,0.0045264782,0.0025512655,0.0050101224],"category_scores_gemma":[0.04838102,0.0016072576,0.0027163557,0.0029941325,0.0023585483,0.003114187,0.0030469352,0.0037244891,0.0015058819],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00037216634,0.00015755284,0.034080643,0.00032956296,0.00047594716,0.00024482352,0.0015566176,0.530484,0.0011196482,0.308846,0.009458609,0.112874396],"study_design_scores_gemma":[0.0000546188,0.00005247411,0.0034721796,0.000080457125,0.00007181019,0.00007344704,0.00011987619,0.90445036,0.0002268163,0.086711556,0.004624765,0.000061651735],"about_ca_topic_score_codex":0.030575043,"about_ca_topic_score_gemma":0.025814971,"teacher_disagreement_score":0.030575043,"about_ca_system_score_codex":0.0025838222,"about_ca_system_score_gemma":0.0026983928,"threshold_uncertainty_score":0.09707534},"labels":[],"label_agreement":null},{"id":"W4388054702","doi":"10.1214/23-aoas1776","title":"Continuous-time modelling of behavioural responses in animal movement","year":2023,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Marine animal studies overview","field":"Environmental Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"Bureau of Medicine and Surgery; National Marine Fisheries Service; U.S. Department of Defense; U.S. Fleet Forces Command; Office of Naval Research; U.S. Navy","keywords":"Baseline (sea); Nonparametric statistics; Computer science; Covariate; Stochastic modelling; Sound exposure; Statistics; Machine learning; Mathematics","score_opus":0.13857183501729953,"score_gpt":0.31002161568547715,"score_spread":0.17144978066817762,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4388054702","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1669488,0.0005554424,0.82695246,0.0007288929,0.00013892775,0.00008514143,0.0009397817,0.0004443916,0.0032061993],"genre_scores_gemma":[0.9576732,0.0004260046,0.03301469,0.000108074666,0.000056479916,0.00027314312,0.00080477784,0.00006973514,0.0075737983],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999151,0.00032201593,0.000053975346,0.0002450308,0.00012732839,0.00010069353],"domain_scores_gemma":[0.9954389,0.0032643941,0.00055651733,0.00022593605,0.00034880723,0.00016547389],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0023220167,0.00077683094,0.0007468545,0.0006846233,0.00029362438,0.0012089274,0.00239179,0.0016144563,0.002442226],"category_scores_gemma":[0.008423836,0.0005166671,0.0013702811,0.0008361923,0.0011476272,0.00093347294,0.0009411649,0.0016371695,0.0003494266],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00003652053,0.00004286222,0.0035003314,0.00005809875,0.000057235233,0.000078607,0.00012296879,0.9763509,0.00076272537,0.0152979195,0.00021577538,0.0034761187],"study_design_scores_gemma":[0.0000028701463,0.0000075056446,0.0005088284,0.000002570085,0.0000044075673,0.000004613064,0.000006648922,0.9971802,0.00003208441,0.002111746,0.00013448914,0.0000039640627],"about_ca_topic_score_codex":0.024190411,"about_ca_topic_score_gemma":0.010814148,"teacher_disagreement_score":0.024190411,"about_ca_system_score_codex":0.00097390835,"about_ca_system_score_gemma":0.000800423,"threshold_uncertainty_score":0.04809922},"labels":[],"label_agreement":null},{"id":"W4391418111","doi":"10.1214/23-aoas1795","title":"Network-level traffic flow prediction: Functional time series vs. functional neural network approach","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Traffic Prediction and Management Techniques","field":"Engineering","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Time series; Computer science; Series (stratigraphy); Artificial neural network; Traffic flow (computer networking); Artificial intelligence; Machine learning; Geology; Computer network","score_opus":0.045327854834515825,"score_gpt":0.2309017812429498,"score_spread":0.18557392640843398,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4391418111","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.09252863,0.0010932655,0.90199417,0.0007637327,0.00011770054,0.00003052717,0.00031184196,0.00030941653,0.0028507016],"genre_scores_gemma":[0.96931165,0.00083764567,0.027691577,0.00010773923,0.00016240562,0.000051856743,0.00046551827,0.000035474826,0.0013361943],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995084,0.00018720469,0.00002479968,0.00014784331,0.00008003183,0.000051725387],"domain_scores_gemma":[0.9984592,0.0008998669,0.00018404938,0.00011095465,0.00027892427,0.000067070774],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016031693,0.001132102,0.0006795155,0.00090553146,0.00025043663,0.00077849434,0.0014573386,0.0010113717,0.0009507082],"category_scores_gemma":[0.003875378,0.0002929723,0.0005628341,0.0011050653,0.0005762049,0.0019777694,0.0006598177,0.0011482539,0.00016049026],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000050052484,0.00003874254,0.0026707137,0.000039095023,0.00003736765,0.00004170126,0.000023523851,0.96545535,0.00029242883,0.008687883,0.00044049593,0.02222264],"study_design_scores_gemma":[4.5806388e-7,0.0000036602971,0.00014535336,0.0000014874256,0.0000024657338,0.0000022504332,0.0000024248109,0.99861157,0.00002977704,0.0011572749,0.00004184596,0.0000014195931],"about_ca_topic_score_codex":0.014798764,"about_ca_topic_score_gemma":0.0071456507,"teacher_disagreement_score":0.014798764,"about_ca_system_score_codex":0.0008808265,"about_ca_system_score_gemma":0.00059988856,"threshold_uncertainty_score":0.029425263},"labels":[],"label_agreement":null},{"id":"W4393975851","doi":"10.1214/23-aoas1854","title":"How are PreLaunch online movie reviews related to box office revenues?","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Digital Marketing and Social Media","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University; York University","funders":"","keywords":"Box office; Revenue; Advertising; Computer science; Business; Data science; Accounting","score_opus":0.08587048036631538,"score_gpt":0.382266476417505,"score_spread":0.2963959960511896,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4393975851","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99164635,0.0008798188,0.0030680445,0.00042444933,0.000048201746,0.000024842062,0.0011994292,0.000040712097,0.002668081],"genre_scores_gemma":[0.9970517,0.00027680877,0.00063314533,0.000031283445,0.000086100365,0.000012528263,0.0011024215,0.0000128949505,0.0007931344],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99929786,0.00014454308,0.000055637593,0.00022725729,0.00016015522,0.00011454797],"domain_scores_gemma":[0.9869594,0.0050720023,0.005923763,0.00043916507,0.0010675548,0.0005380216],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011120457,0.00030745508,0.00048900594,0.0017023019,0.00039897376,0.001369613,0.00036079396,0.0005456772,0.0039910763],"category_scores_gemma":[0.014119327,0.00020588607,0.00037772235,0.0018718704,0.00046721956,0.0016863114,0.000548793,0.00095879677,0.0010200308],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00027114642,0.00009106864,0.9564269,0.00008251781,0.00013486737,0.00024906057,0.00039997284,0.0028776398,0.0009294443,0.0013586974,0.002330836,0.03484787],"study_design_scores_gemma":[0.0000045795696,0.00005028189,0.9837435,0.000020950216,0.000037839865,0.00015290096,0.00045595324,0.012759829,0.00031436465,0.0012171721,0.0012212186,0.000021394291],"about_ca_topic_score_codex":0.005049147,"about_ca_topic_score_gemma":0.0077041765,"teacher_disagreement_score":0.005049147,"about_ca_system_score_codex":0.00036886963,"about_ca_system_score_gemma":0.00023813857,"threshold_uncertainty_score":0.01335144},"labels":[],"label_agreement":null},{"id":"W4394006165","doi":"10.1214/23-aoas1825","title":"A population-aware retrospective regression to detect genome-wide variants with sex difference in allele frequency","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Hospital for Sick Children","funders":"","keywords":"Allele; Allele frequency; Regression; Genetics; Regression analysis; Population; Genome; Biology; Computer science; Medicine; Statistics; Gene; Machine learning; Mathematics; Environmental health","score_opus":0.02478594328100024,"score_gpt":0.3053104523246091,"score_spread":0.28052450904360887,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394006165","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.11075034,0.00052549096,0.88571984,0.0005005722,0.000092729184,0.00007708678,0.000995069,0.0009821556,0.00035675205],"genre_scores_gemma":[0.75294185,0.00030597832,0.24174753,0.00030551862,0.00013565997,0.00018587988,0.002368649,0.0002302721,0.0017787508],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.995739,0.0026136418,0.0001784741,0.0010830198,0.0002380514,0.00014787025],"domain_scores_gemma":[0.98393387,0.012305377,0.0010157196,0.0019357416,0.0006209496,0.00018828797],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011404923,0.00068597,0.0009527567,0.0009754295,0.0003768227,0.0010178816,0.0020023317,0.00085077924,0.0014848468],"category_scores_gemma":[0.03011712,0.0005170144,0.0016642704,0.0012259148,0.0006695416,0.0006332711,0.00091962644,0.0018330608,0.0004666574],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012931741,0.00027090235,0.37653136,0.000323413,0.0031725292,0.001233667,0.0004759761,0.27839723,0.013299414,0.018987035,0.0053808875,0.30063444],"study_design_scores_gemma":[0.00007470619,0.00017493275,0.0220556,0.000028854698,0.00028201487,0.00049812085,0.000062435254,0.96258235,0.0022807624,0.009307086,0.0026055323,0.000047603306],"about_ca_topic_score_codex":0.007713634,"about_ca_topic_score_gemma":0.0073144226,"teacher_disagreement_score":0.011404923,"about_ca_system_score_codex":0.00029695398,"about_ca_system_score_gemma":0.0009638976,"threshold_uncertainty_score":0.06031567},"labels":[],"label_agreement":null},{"id":"W4401373758","doi":"10.1214/24-aoas1877","title":"Probabilistic contrastive dimension reduction for case-control study data","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Center for Advancing Translational Sciences; National Human Genome Research Institute; National Heart, Lung, and Blood Institute; Canadian Institute for Advanced Research; National Cancer Institute; National Science Foundation; National Institutes of Health; National Institute of Environmental Health Sciences; Leona M. and Harry B. Helmsley Charitable Trust","keywords":"Dimensionality reduction; Dimension (graph theory); Reduction (mathematics); Probabilistic logic; Computer science; Statistics; Data reduction; Artificial intelligence; Natural language processing; Mathematics","score_opus":0.08856121255749513,"score_gpt":0.37194932954972393,"score_spread":0.28338811699222877,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401373758","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005045044,0.00038225966,0.9934002,0.00040809374,0.000034661545,0.00017434945,0.00015873411,0.00020833373,0.00018835158],"genre_scores_gemma":[0.23399848,0.0005284492,0.76179004,0.0005510615,0.00021872726,0.0014358672,0.00091721956,0.00010946678,0.00045068903],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9719257,0.021640653,0.00092708884,0.002969757,0.002326252,0.00021051691],"domain_scores_gemma":[0.9125114,0.073140204,0.003472788,0.008529861,0.0019846058,0.00036116893],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04075399,0.001089864,0.001597415,0.002722795,0.0010457474,0.0018706761,0.002595002,0.0019598792,0.0014482422],"category_scores_gemma":[0.11609815,0.00090468855,0.002394829,0.0020994542,0.0038222098,0.0014960065,0.0023520757,0.0039649154,0.00033551131],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011193714,0.00039450455,0.024564834,0.0010488679,0.0017175493,0.0008310105,0.0010016842,0.3031078,0.009804653,0.28364334,0.01037278,0.3623936],"study_design_scores_gemma":[0.00013226867,0.0001561858,0.0051280363,0.00007972323,0.00011964838,0.00035771835,0.000047023972,0.7510155,0.00262396,0.23531592,0.0049521658,0.00007181002],"about_ca_topic_score_codex":0.0022666287,"about_ca_topic_score_gemma":0.0018226244,"teacher_disagreement_score":0.04075399,"about_ca_system_score_codex":0.0016730814,"about_ca_system_score_gemma":0.0018778989,"threshold_uncertainty_score":0.21553028},"labels":[],"label_agreement":null},{"id":"W4403577681","doi":"10.1214/25-aoas2115","title":"K-contact distance for noisy nonhomogeneous spatial point data with application to repeating fast radio burst sources","year":2024,"lang":"en","type":"preprint","venue":"The Annals of Applied Statistics","topic":"Robotics and Sensor-Based Localization","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University; York University; McGill University Health Centre; Simon Fraser University; Canadian Institute for Theoretical Astrophysics","funders":"Smithsonian Astrophysical Observatory; University of Toronto","keywords":"Point (geometry); Computer science; Algorithm; Mathematics; Geometry","score_opus":0.032036673803525095,"score_gpt":0.28131462166736615,"score_spread":0.24927794786384105,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403577681","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027623223,0.00015818479,0.97151035,0.000120867895,0.000011251882,0.000033780685,0.00006109925,0.00018430696,0.000296841],"genre_scores_gemma":[0.6899943,0.00050593517,0.30607426,0.00014563986,0.00016975627,0.00028267884,0.0007140673,0.00023602151,0.0018773568],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9977241,0.00080901856,0.00015686314,0.00062092696,0.0005512806,0.00013777036],"domain_scores_gemma":[0.96455973,0.027072294,0.0032787612,0.0030708017,0.0013642373,0.0006540877],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006174157,0.00091423123,0.0013964395,0.0024695941,0.00083771715,0.0013838843,0.002682845,0.0018372778,0.0011316183],"category_scores_gemma":[0.034668155,0.0006869795,0.0014117009,0.0018208355,0.0020744167,0.0026664485,0.0027480323,0.0019474863,0.00041731537],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017812627,0.00008196446,0.01228916,0.00013010539,0.00013981292,0.00029813586,0.00042301702,0.9044409,0.0028827633,0.04516005,0.00069162564,0.033284344],"study_design_scores_gemma":[0.0000050525796,0.000021571854,0.00073044427,0.000005283965,0.0000059470217,0.000046538982,0.000022486225,0.98615825,0.0005597368,0.012154245,0.00028035752,0.000010167786],"about_ca_topic_score_codex":0.00501143,"about_ca_topic_score_gemma":0.0038930967,"teacher_disagreement_score":0.006174157,"about_ca_system_score_codex":0.0015557522,"about_ca_system_score_gemma":0.0011856756,"threshold_uncertainty_score":0.032652497},"labels":[],"label_agreement":null},{"id":"W4403924989","doi":"10.1214/24-aoas1917","title":"Statistical curve models for inferring 3D chromatin architecture","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Genomics and Chromatin Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"National Institute of Biomedical Imaging and Bioengineering; National Institute of General Medical Sciences; Natural Sciences and Engineering Research Council of Canada; University of Toronto; National Institutes of Health; National Science Foundation","keywords":"Computer science; Architecture; Artificial intelligence; Natural language processing; Computational biology; Geography; Biology; Archaeology","score_opus":0.030417245872816058,"score_gpt":0.30504001709156947,"score_spread":0.2746227712187534,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403924989","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.014584409,0.00030821696,0.98356426,0.00020596078,0.0000188301,0.000035513985,0.00028342952,0.0005179942,0.00048149025],"genre_scores_gemma":[0.48227873,0.0018983233,0.5046059,0.00029048638,0.0002358293,0.0006876045,0.0035010239,0.0006797859,0.0058222855],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99872106,0.0005153701,0.000069369904,0.00029799412,0.0003104289,0.00008576355],"domain_scores_gemma":[0.9883272,0.008611324,0.001026044,0.00089701934,0.00082368875,0.00031468453],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0042104786,0.0011314179,0.0012793983,0.0033907646,0.00072380767,0.0016607756,0.0029558432,0.002398502,0.0024190547],"category_scores_gemma":[0.016772203,0.0011170342,0.0021081243,0.0028566576,0.0018220268,0.0023650911,0.0022939916,0.0022774097,0.0012969793],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00003792349,0.000021319905,0.0018880268,0.000053246597,0.0000336367,0.00005477763,0.00007350483,0.94309664,0.00085304474,0.033496115,0.00067375036,0.019718057],"study_design_scores_gemma":[0.0000016333943,0.0000040023256,0.00010167911,0.0000032712478,0.0000014156125,0.00000904135,0.000004405318,0.9890057,0.000079478596,0.010517964,0.0002665023,0.0000048603497],"about_ca_topic_score_codex":0.009369604,"about_ca_topic_score_gemma":0.00643716,"teacher_disagreement_score":0.009369604,"about_ca_system_score_codex":0.0016026968,"about_ca_system_score_gemma":0.0013793743,"threshold_uncertainty_score":0.022267342},"labels":[],"label_agreement":null},{"id":"W4403925054","doi":"10.1214/24-aoas1933","title":"Learning risk preferences in Markov decision processes: An application to the fourth down decision in the national football league","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; University of British Columbia","funders":"","keywords":"Football; League; Computer science; Artificial intelligence; Econometrics; Economics; Political science","score_opus":0.0637202319053105,"score_gpt":0.3171072281836259,"score_spread":0.2533869962783154,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403925054","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.42808923,0.0015406222,0.56060606,0.0039196783,0.00012629089,0.00020767718,0.00035714713,0.0002524462,0.0049008406],"genre_scores_gemma":[0.9606999,0.00071800465,0.034826722,0.00034583837,0.00010786593,0.00015129027,0.00023560425,0.000037721056,0.002876986],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9969427,0.0020292911,0.00008856249,0.0004530036,0.00018800727,0.00029848592],"domain_scores_gemma":[0.9594123,0.036606513,0.0019471761,0.0005289057,0.0006788189,0.00082631677],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010272247,0.0011208443,0.0021873252,0.00096261,0.00091801165,0.0019761992,0.001532768,0.0025703607,0.0039120186],"category_scores_gemma":[0.034721486,0.0007664162,0.0017653317,0.0008295485,0.0022717111,0.0019109516,0.0022819769,0.004705676,0.00022763164],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00031554504,0.0002475223,0.014007,0.000094443756,0.00016088407,0.0003551326,0.00051030016,0.87473506,0.00042894768,0.09502869,0.0008992883,0.013217132],"study_design_scores_gemma":[0.000037808804,0.000042925523,0.0009787686,0.000014282311,0.000020484333,0.00001822799,0.00007011694,0.95570177,0.00008111676,0.042752743,0.00025991598,0.0000218287],"about_ca_topic_score_codex":0.030972524,"about_ca_topic_score_gemma":0.019012606,"teacher_disagreement_score":0.030972524,"about_ca_system_score_codex":0.0024459325,"about_ca_system_score_gemma":0.0023984269,"threshold_uncertainty_score":0.061584473},"labels":[],"label_agreement":null},{"id":"W4403936075","doi":"10.1214/24-aoas1942","title":"Learning brain connectivity in social cognition with dynamic network regression","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Functional Brain Connectivity Studies","field":"Neuroscience","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Cognition; Regression; Cognitive psychology; Artificial intelligence; Regression analysis; Psychology; Computer science; Machine learning; Neuroscience","score_opus":0.06496974438133032,"score_gpt":0.3407465099376238,"score_spread":0.2757767655562935,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403936075","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.076981544,0.00042265307,0.9203328,0.0007175861,0.000027257827,0.000055587145,0.00015252818,0.00033260236,0.0009774814],"genre_scores_gemma":[0.79355747,0.0007325328,0.20122479,0.0001873206,0.00013496431,0.00024425526,0.00059672183,0.00019707771,0.0031248028],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99897337,0.00067679404,0.00002350589,0.00020409063,0.00006810866,0.000054179654],"domain_scores_gemma":[0.9925764,0.0061940127,0.00054407446,0.0003459334,0.00021827278,0.000121259414],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038493595,0.00087955175,0.00081841653,0.0012408491,0.00040983988,0.0008607883,0.001029748,0.0009977045,0.0019331052],"category_scores_gemma":[0.01741025,0.00056428113,0.0009926542,0.0010507922,0.0011409259,0.0017268019,0.0013164029,0.0016525295,0.00033421614],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000111647256,0.0000881401,0.0061544827,0.00007767536,0.0001802857,0.00012891433,0.00017357727,0.9188042,0.0016294611,0.025202382,0.0011066473,0.04634262],"study_design_scores_gemma":[0.0000072343746,0.000015760015,0.00090143137,0.000005514024,0.000008998414,0.000016955322,0.000011916541,0.98233324,0.00015710248,0.016281202,0.00025311892,0.0000075587022],"about_ca_topic_score_codex":0.008931056,"about_ca_topic_score_gemma":0.010724711,"teacher_disagreement_score":0.008931056,"about_ca_system_score_codex":0.0009342671,"about_ca_system_score_gemma":0.0007525374,"threshold_uncertainty_score":0.020357609},"labels":[],"label_agreement":null},{"id":"W4403936079","doi":"10.1214/24-aoas1943","title":"Modeling trajectories using functional linear differential equations","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Control Systems and Identification","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institute of Neurological Disorders and Stroke; Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Applied mathematics; Mathematics; Computer science","score_opus":0.10356006756529802,"score_gpt":0.30158981968118304,"score_spread":0.19802975211588503,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403936079","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.02014619,0.0002742076,0.97596425,0.00044433938,0.000046276105,0.000054859876,0.00042232854,0.00035246043,0.0022950626],"genre_scores_gemma":[0.81488305,0.0011363329,0.1638572,0.0003494085,0.00012020079,0.0009933376,0.001448984,0.0002763515,0.01693521],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9994,0.0001984328,0.00003952267,0.00018276581,0.00011558642,0.000063630854],"domain_scores_gemma":[0.99758923,0.0015887723,0.00034519084,0.00009564113,0.0003197053,0.000061523715],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017480163,0.0010271383,0.00085236406,0.0010675193,0.00038686843,0.001160429,0.0013710179,0.0019578729,0.0030338345],"category_scores_gemma":[0.0064715017,0.00067040004,0.0011307105,0.00084833294,0.00078612874,0.00096806587,0.0011808336,0.0015147503,0.00065700716],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000012130272,0.000014109013,0.0009906961,0.0000405882,0.000021337217,0.000047634043,0.000059338767,0.9807356,0.00054031063,0.012950345,0.00030348016,0.0042844564],"study_design_scores_gemma":[0.0000018311845,0.0000048934894,0.00008236065,0.0000036919894,0.0000025787222,0.00000559383,0.000004370649,0.9973647,0.000057041612,0.0021762715,0.00029316216,0.000003577771],"about_ca_topic_score_codex":0.023672255,"about_ca_topic_score_gemma":0.0123416735,"teacher_disagreement_score":0.023672255,"about_ca_system_score_codex":0.0012541434,"about_ca_system_score_gemma":0.0012497016,"threshold_uncertainty_score":0.047068954},"labels":[],"label_agreement":null},{"id":"W4403936291","doi":"10.1214/24-aoas1915","title":"Multisite disease analytics with applications to estimating COVID-19 undetected cases in Canada","year":2024,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Victoria; Simon Fraser University","funders":"","keywords":"Coronavirus disease 2019 (COVID-19); 2019-20 coronavirus outbreak; Analytics; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Data science; Pandemic; Computer science; Disease; Medicine; Infectious disease (medical specialty); Virology; Outbreak","score_opus":0.34565989660508173,"score_gpt":0.4734745204622825,"score_spread":0.12781462385720077,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403936291","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.65562123,0.0014245784,0.3118364,0.0046094935,0.0000918883,0.00031250936,0.0177635,0.002481639,0.005858771],"genre_scores_gemma":[0.94766325,0.0004266178,0.04593604,0.00012067555,0.00002858213,0.000069517955,0.0038379992,0.00007429133,0.001843032],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992694,0.00021231332,0.000038721068,0.00020443497,0.00016420311,0.00011099121],"domain_scores_gemma":[0.9965101,0.0018338886,0.00043487342,0.000272902,0.0007678978,0.00018036841],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0020869342,0.0007494243,0.00066969375,0.0017838667,0.0007872851,0.0012896629,0.001523491,0.0006209914,0.0016633391],"category_scores_gemma":[0.009269441,0.00037499378,0.00076648523,0.0024856334,0.000688405,0.0006061988,0.0014218681,0.0011123514,0.00017709551],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00012357785,0.00007515797,0.12729456,0.00009241829,0.00016840536,0.00020366984,0.00035285964,0.81464654,0.00058624847,0.010212733,0.004216517,0.042027287],"study_design_scores_gemma":[0.000011486397,0.000010062334,0.013832056,0.00001647159,0.000014182561,0.000026583633,0.00012427794,0.9798605,0.00016577641,0.0047117495,0.001207069,0.000019814637],"about_ca_topic_score_codex":0.9408332,"about_ca_topic_score_gemma":0.9142351,"teacher_disagreement_score":0.05916679,"about_ca_system_score_codex":0.009583004,"about_ca_system_score_gemma":0.011002072,"threshold_uncertainty_score":0.119030416},"labels":[],"label_agreement":null},{"id":"W4408584528","doi":"10.1214/24-aoas1914","title":"Leveraging cellphone-derived mobility networks to assess Covid-19 travel risk","year":2025,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; University of Guelph","funders":"","keywords":"Coronavirus disease 2019 (COVID-19); Computer science; 2019-20 coronavirus outbreak; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Geography; Data science; Econometrics; Mathematics; Medicine; Virology","score_opus":0.11589663114758272,"score_gpt":0.4026240938907544,"score_spread":0.2867274627431716,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408584528","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.96544313,0.0004022102,0.030304076,0.0003452534,0.00002808956,0.000053082884,0.002117963,0.00006589798,0.0012403055],"genre_scores_gemma":[0.9922037,0.00017688092,0.0057105618,0.000025148396,0.000018820087,0.000025218404,0.0015574951,0.0000068253753,0.00027535358],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.999401,0.0003190004,0.000029295228,0.00012281706,0.000077953155,0.000049954386],"domain_scores_gemma":[0.99740916,0.0015240846,0.000459321,0.00019387716,0.0002947215,0.00011876987],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016940153,0.00048702754,0.00040222058,0.0018987176,0.00018597663,0.0008746556,0.00051644136,0.00064467464,0.00060368434],"category_scores_gemma":[0.008284803,0.00015019115,0.0004197215,0.0014379799,0.00020769013,0.0009474186,0.0009608078,0.00047590985,0.00019830442],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00021927828,0.00012471993,0.7724448,0.000081896826,0.0004484921,0.00019287124,0.00035433675,0.19280598,0.0010538028,0.0027310778,0.00093357434,0.028609302],"study_design_scores_gemma":[0.000011586444,0.00014210508,0.18645284,0.000039276994,0.00011162809,0.0001833371,0.00085916737,0.805215,0.00040730936,0.0052428874,0.0012965569,0.00003841042],"about_ca_topic_score_codex":0.022010433,"about_ca_topic_score_gemma":0.02225354,"teacher_disagreement_score":0.022010433,"about_ca_system_score_codex":0.0005115416,"about_ca_system_score_gemma":0.0004999418,"threshold_uncertainty_score":0.04376465},"labels":[],"label_agreement":null},{"id":"W4408584779","doi":"10.1214/24-aoas1964","title":"A three-state coupled Markov switching model for COVID-19 outbreaks across Quebec based on hospital admissions","year":2025,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University","funders":"","keywords":"Coronavirus disease 2019 (COVID-19); Outbreak; 2019-20 coronavirus outbreak; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Markov chain; Markov model; Statistics; State (computer science); Geography; Computer science; Econometrics; Virology; Medicine; Mathematics; Algorithm; Infectious disease (medical specialty)","score_opus":0.25073335402443636,"score_gpt":0.47628581849535473,"score_spread":0.22555246447091837,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408584779","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.703081,0.0012092586,0.26423526,0.0049898094,0.00018957224,0.00032378477,0.010240416,0.00061582756,0.015115032],"genre_scores_gemma":[0.97457856,0.00036749945,0.0113929715,0.00017469404,0.000050603536,0.00013150003,0.0023574207,0.000035242265,0.010911535],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99935323,0.00018582761,0.000026267384,0.00017529743,0.000059265636,0.00020024311],"domain_scores_gemma":[0.99826956,0.0008939226,0.00029652988,0.00007140537,0.00030146932,0.00016707124],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012829748,0.0006805073,0.000865703,0.0011823209,0.0010540416,0.0015980463,0.0027243854,0.0014976982,0.0063167345],"category_scores_gemma":[0.002897862,0.00055573037,0.0010630505,0.0011489372,0.0011020241,0.000768832,0.0008936513,0.0014612894,0.00043401693],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000205629,0.000117224,0.032579478,0.000065939945,0.00015804087,0.0005037874,0.0003250451,0.9082054,0.0012995111,0.042504273,0.003885863,0.010149866],"study_design_scores_gemma":[0.000019774376,0.0000147772525,0.004019934,0.000008595772,0.000028401873,0.000019712295,0.00004710191,0.99234456,0.00004375525,0.0027849027,0.00065047265,0.000018043002],"about_ca_topic_score_codex":0.79753524,"about_ca_topic_score_gemma":0.7121315,"teacher_disagreement_score":0.20246476,"about_ca_system_score_codex":0.008859769,"about_ca_system_score_gemma":0.0056194793,"threshold_uncertainty_score":0.40731418},"labels":[],"label_agreement":null},{"id":"W4408585636","doi":"10.1214/24-aoas1958","title":"Poisson cluster process models for detecting ultra-diffuse galaxies","year":2025,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo; University of Toronto","funders":"","keywords":"Computer science; Cluster (spacecraft); Poisson distribution; Statistical physics; Statistics; Mathematics; Physics","score_opus":0.05633146807322464,"score_gpt":0.34934960217812083,"score_spread":0.2930181341048962,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408585636","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03002056,0.00026726403,0.96787035,0.00038127764,0.00004510278,0.00007420069,0.00023205556,0.00029765113,0.0008115672],"genre_scores_gemma":[0.7515829,0.0008367721,0.24014485,0.00043297626,0.00028097132,0.0005196318,0.0014241773,0.00015399065,0.0046237484],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9981364,0.00078737695,0.00006926136,0.0004960323,0.0003412195,0.00016964981],"domain_scores_gemma":[0.98910797,0.00731562,0.0013436999,0.00078877015,0.0010625619,0.0003813501],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0057210256,0.0011735846,0.0014744239,0.002015111,0.0007916148,0.0017029853,0.0057814335,0.0016932139,0.0020473795],"category_scores_gemma":[0.015598539,0.0009404047,0.0017365175,0.0013630274,0.0021096687,0.0021238453,0.0020330176,0.0025212204,0.00062579685],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013499493,0.000054268665,0.0116968015,0.000082514445,0.0001394521,0.00013739541,0.00015339762,0.88483906,0.0007116627,0.076867215,0.0014327051,0.023750544],"study_design_scores_gemma":[0.00001097585,0.000012981174,0.000405543,0.0000058827873,0.000014120662,0.000020995627,0.000009865052,0.98262054,0.00016419863,0.016425692,0.000297502,0.000011673084],"about_ca_topic_score_codex":0.015177483,"about_ca_topic_score_gemma":0.009321512,"teacher_disagreement_score":0.015177483,"about_ca_system_score_codex":0.0016614302,"about_ca_system_score_gemma":0.0015930489,"threshold_uncertainty_score":0.030256093},"labels":[],"label_agreement":null},{"id":"W4410804643","doi":"10.1214/24-aoas2009","title":"A privacy-preserved and high-utility synthesis strategy for risk-based stratified subgroups of the Canadian scleroderma patient registry data","year":2025,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University; University of Alberta","funders":"","keywords":"Medicine; Data science; Family medicine; Computer science","score_opus":0.11185211933707181,"score_gpt":0.31742714614673184,"score_spread":0.20557502680966003,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4410804643","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020679101,0.0005589364,0.9723078,0.0013404805,0.00006638781,0.0011391189,0.00195866,0.0005614101,0.0013881365],"genre_scores_gemma":[0.43739107,0.0004718196,0.55369574,0.0005809095,0.00016356188,0.0018513965,0.0037104266,0.00019868827,0.0019363696],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9757547,0.016863031,0.0013572983,0.0026325919,0.0026629854,0.0007293077],"domain_scores_gemma":[0.9180627,0.06496577,0.001839497,0.010160117,0.0042806068,0.0006912609],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.044242606,0.0014806534,0.0030649332,0.0040547173,0.0014916877,0.0041955886,0.0030040315,0.0014591125,0.005951955],"category_scores_gemma":[0.12259642,0.001130787,0.0032362216,0.0037798013,0.0017694825,0.0036508073,0.0046139085,0.0028796184,0.0006198768],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0058152406,0.0006095465,0.014390293,0.0015426057,0.0021880113,0.00069602835,0.0020555349,0.20919953,0.0052169035,0.24547598,0.01080028,0.5020101],"study_design_scores_gemma":[0.0004374531,0.00044788062,0.001974344,0.00023341135,0.0008872931,0.00017547773,0.0005143517,0.5339448,0.0045167413,0.44977924,0.0069870087,0.00010199248],"about_ca_topic_score_codex":0.0076143323,"about_ca_topic_score_gemma":0.008476177,"teacher_disagreement_score":0.9923857,"about_ca_system_score_codex":0.002483283,"about_ca_system_score_gemma":0.0088023385,"threshold_uncertainty_score":0.23398006},"labels":[],"label_agreement":null},{"id":"W4417045360","doi":"10.1214/25-aoas2106","title":"Statistical learning of trade credit insurance network data with applications to ratemaking and reserving","year":2025,"lang":"","type":"article","venue":"The Annals of Applied Statistics","topic":"Financial Distress and Bankruptcy Prediction","field":"Business, Management and Accounting","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Statistical learning; Artificial neural network; Statistical analysis; Statistical model; Credit risk","score_opus":0.05797655912212767,"score_gpt":0.31871985938234126,"score_spread":0.2607433002602136,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417045360","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.39530304,0.0025323639,0.59153384,0.004340724,0.0003721473,0.00013334754,0.0019241577,0.001491643,0.002368764],"genre_scores_gemma":[0.93280643,0.00096347136,0.05964544,0.00016555506,0.00047115897,0.000101490754,0.0032482059,0.00008997322,0.00250829],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983784,0.00085989735,0.00012562131,0.00030143152,0.00023713888,0.00009756666],"domain_scores_gemma":[0.9704789,0.024481308,0.001062293,0.0018402308,0.0017430792,0.0003942501],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007192109,0.0007684906,0.0011073332,0.0018810807,0.0005852347,0.001684637,0.0014604185,0.0015535132,0.0013686137],"category_scores_gemma":[0.036873132,0.0005904155,0.0009429348,0.002192252,0.0011594798,0.0021018563,0.001406852,0.0027593512,0.0005028985],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00029088612,0.00044643562,0.02112364,0.00006772253,0.00012444421,0.00010283782,0.00011855324,0.767722,0.0006598869,0.009220692,0.0045510815,0.1955718],"study_design_scores_gemma":[0.0000048624415,0.000013347425,0.00093024847,0.0000036538268,0.0000043082623,0.000006936476,0.000010336206,0.994371,0.00012111807,0.0043509295,0.00017876089,0.000004562373],"about_ca_topic_score_codex":0.011166989,"about_ca_topic_score_gemma":0.010397265,"teacher_disagreement_score":0.011166989,"about_ca_system_score_codex":0.0011676591,"about_ca_system_score_gemma":0.0015225688,"threshold_uncertainty_score":0.03803593},"labels":[],"label_agreement":null},{"id":"W4417045508","doi":"10.1214/25-aoas2085","title":"The dynamic interplay of clan culture and socioeconomic factors on fertility: Evidence from China","year":2025,"lang":"","type":"article","venue":"The Annals of Applied Statistics","topic":"Family Dynamics and Relationships","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Clan; Socioeconomic status; Fertility; Population; China; Population growth; Birth rate","score_opus":0.040165505688895306,"score_gpt":0.35746637972926654,"score_spread":0.3173008740403712,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417045508","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99786216,0.0006763882,0.00018651852,0.00022929854,0.000005514899,0.0000075354,0.00025592546,0.0000023449702,0.000774344],"genre_scores_gemma":[0.99903643,0.00044945013,0.00008639846,0.000045318302,0.0000047524595,0.0000053213653,0.00022375527,0.0000010221893,0.0001475199],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9991448,0.0003807553,0.00006603464,0.00012514417,0.00013659839,0.00014667872],"domain_scores_gemma":[0.9967501,0.0010861572,0.00074818084,0.0004352888,0.0006003855,0.00037981593],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002651464,0.0002816474,0.0003993663,0.0011865988,0.0010625578,0.0005659816,0.00063243276,0.0002528551,0.0014897181],"category_scores_gemma":[0.0042637032,0.00020741326,0.0006660675,0.0018900334,0.0008498552,0.0003829041,0.0009020116,0.00046049527,0.00009919008],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00003432215,0.000028142496,0.98694885,0.00005928441,0.0002621487,0.00013151242,0.0018076502,0.00036150072,0.00008666293,0.0006857256,0.0003535046,0.009240708],"study_design_scores_gemma":[0.00000455723,0.000029204308,0.99596584,0.00004433506,0.0001150124,0.000036632577,0.0015643432,0.0011487148,0.00006365717,0.00019598892,0.00082301354,0.000008694703],"about_ca_topic_score_codex":0.18794818,"about_ca_topic_score_gemma":0.23707154,"teacher_disagreement_score":0.18794818,"about_ca_system_score_codex":0.0013443365,"about_ca_system_score_gemma":0.0023095917,"threshold_uncertainty_score":0.37370837},"labels":[],"label_agreement":null},{"id":"W4417071687","doi":"10.1214/25-aoas2091","title":"Tomographic reconstruction of disease transmission landscapes from GPS-recorded random paths","year":2025,"lang":"","type":"article","venue":"The Annals of Applied Statistics","topic":"Data-Driven Disease Surveillance","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Pattern recognition (psychology); Tomography; Transmission (telecommunications); Tomographic reconstruction; Iterative reconstruction","score_opus":0.02081823664402697,"score_gpt":0.2968701473228377,"score_spread":0.2760519106788107,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417071687","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6303674,0.0010251293,0.35413936,0.0016932541,0.0000982134,0.00011062805,0.008173912,0.0014552111,0.0029369483],"genre_scores_gemma":[0.9534002,0.0003077986,0.043746293,0.000047651523,0.000026508236,0.000062569095,0.0017806771,0.000050747327,0.0005775548],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9998274,0.00007148332,0.000010883924,0.000041882366,0.000027035048,0.000021268414],"domain_scores_gemma":[0.9985141,0.0010410422,0.00013010192,0.00013099394,0.00011927282,0.000064459564],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00043667122,0.0004254863,0.0004678018,0.0010832157,0.00024279588,0.0008845469,0.0010183079,0.0010175243,0.0019502948],"category_scores_gemma":[0.005180607,0.00065976614,0.00075600005,0.0015586235,0.00053656683,0.00060156523,0.0006900175,0.00065538334,0.0003819582],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009060128,0.000026431051,0.008967018,0.000059262085,0.000049732025,0.00016063587,0.00006962679,0.9788174,0.0007968424,0.0025932377,0.00080353953,0.0075655784],"study_design_scores_gemma":[0.000015008688,0.0000066574107,0.0015724599,0.000007294017,0.000007106417,0.00005488767,0.000023487606,0.99606615,0.000121873185,0.0018352721,0.00028223926,0.000007633559],"about_ca_topic_score_codex":0.020620935,"about_ca_topic_score_gemma":0.014742785,"teacher_disagreement_score":0.020620935,"about_ca_system_score_codex":0.0006158706,"about_ca_system_score_gemma":0.0008618105,"threshold_uncertainty_score":0.041001856},"labels":[],"label_agreement":null}]}