{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":31,"total_is_capped":false,"direct_labels_cover":4,"predictions_cover":31,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"query_hash":"a02b92407db7","filters":{"venue":"Data Science Journal"}},"results":[{"id":"W3048186973","doi":"10.5334/dsj-2020-043","title":"The CARE Principles for Indigenous Data Governance","year":2020,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1206,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Laurentian University","funders":"Rural Development Administration; European Commission","keywords":"Indigenous; Sovereignty; Data governance; Data sharing; Corporate governance; Stewardship (theology); Political science; Public administration; Public relations; Sociology; Law; Business; Medicine; Data quality","retraction":null,"screen_n_in":null,"score":{"opus":0.3437787918939087,"gpt":0.4254753768765994,"spread":0.08169658498269072,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.0068816,0.0001074647,0.0001003795,0.00004539864,0.002599995,0.01663155,0.0692704,0.00001615007,0.000003322821],"category_scores_gemma":[0.00756462,0.00006919435,0.00001808042,0.001027167,0.0003855879,0.1058463,0.02386449,0.0003618239,0.00002741393],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008139827,"about_ca_system_score_gemma":0.001392241,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002073636,"about_ca_topic_score_gemma":0.00005825762,"domain_scores_codex":[0.9963055,0.00007221621,0.0003108673,0.0009715131,0.001663143,0.0006767356],"domain_scores_gemma":[0.9930814,0.0003578861,0.0004000109,0.005540348,0.0002698512,0.0003504913],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007383252,0.00008483884,0.002294803,0.0000964118,0.00007883446,0.000169792,0.004045895,0.0002448619,0.0008598361,0.2047193,0.09816121,0.6891704],"study_design_scores_gemma":[0.0001636146,0.00008467791,0.002148876,0.00001009439,0.000006029367,0.00003913936,0.0004993922,0.07463297,0.00008463098,0.0001062778,0.9221182,0.0001060448],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001506938,0.003043496,0.9333755,0.05633418,0.001130232,0.0008137605,0.001950406,0.00008380703,0.0017617],"genre_scores_gemma":[0.1881604,0.03180178,0.7688311,0.006244382,0.003110113,0.00004126283,0.001062811,0.00005429417,0.0006938366],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.823957,"threshold_uncertainty_score":0.9986985,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2036014579","doi":"10.2481/dsj.5.178","title":"Geography Markup Language","year":2006,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Geographic Information Systems Studies","field":"Social Sciences","cited_by":105,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Galdos Systems (Canada)","funders":"","keywords":"Markup language; XML; Geospatial analysis; Geomatics; Computer science; Document type definition; World Wide Web; Interoperability; Information retrieval; Geography; Document Structure Description; Remote sensing","retraction":null,"screen_n_in":null,"score":{"opus":0.02436389931910291,"gpt":0.3337672918235531,"spread":0.3094033925044501,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.006307485,0.00006965171,0.00009503196,0.0004394811,0.003037226,0.0007754082,0.001878441,0.00002936925,0.0001748279],"category_scores_gemma":[0.0002550465,0.00005728228,0.00003989557,0.001540107,0.00135599,0.003751356,0.0002978636,0.0001560703,0.00009503751],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004793714,"about_ca_system_score_gemma":0.0002771805,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003408645,"about_ca_topic_score_gemma":0.001336835,"domain_scores_codex":[0.9977407,0.00004683043,0.0002888066,0.0001759214,0.001266714,0.0004809696],"domain_scores_gemma":[0.9989511,0.00003922375,0.0002006184,0.0004102593,0.0002658715,0.0001328984],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001101243,0.0001199432,0.6487464,0.00001545308,0.00003834819,0.00006451304,0.02984191,0.00002724829,0.0006315579,0.1048355,0.1887604,0.02690769],"study_design_scores_gemma":[0.0002774777,0.00001788786,0.3053447,0.00002942415,0.00001348316,0.00008390946,0.04944879,0.00009565691,0.000027144,0.002118951,0.6422946,0.0002480375],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.3803984,0.0009756615,0.001130918,0.003777822,0.002909934,0.0002845026,0.0001382434,0.0001555763,0.610229],"genre_scores_gemma":[0.9967363,0.0001158226,0.001794394,0.000158041,0.0007286178,0.000001595061,0.00001592006,0.000002983444,0.0004463477],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6163379,"threshold_uncertainty_score":0.9982607,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2908272686","doi":"10.5334/dsj-2019-001","title":"Understanding Human Mobility Patterns in a Developing Country Using Mobile Phone Data","year":2019,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":47,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"","keywords":"TRIPS architecture; Mobile phone; Phone; Computer science; Destinations; Centroid; Entropy (arrow of time); Statistics; Geography; Mathematics; Telecommunications","retraction":null,"screen_n_in":null,"score":{"opus":0.3301334090330965,"gpt":0.4317011090950576,"spread":0.1015677000619611,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.01463965,0.0001109057,0.0002094543,0.0003467741,0.002131554,0.0008474777,0.003994029,0.00006115107,0.0008030343],"category_scores_gemma":[0.0004206378,0.0001075493,0.0000229832,0.001735153,0.0007786518,0.004630253,0.0008444102,0.000372066,0.00002626227],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00157749,"about_ca_system_score_gemma":0.002635871,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01126586,"about_ca_topic_score_gemma":0.04723522,"domain_scores_codex":[0.9968832,0.0002214649,0.0004645216,0.000697947,0.001171812,0.0005610907],"domain_scores_gemma":[0.9978449,0.0001343569,0.000229211,0.00147537,0.000129377,0.0001868156],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001373442,0.0002986943,0.9740191,0.00006309094,0.0000319434,0.00002319752,0.0114537,0.002882206,0.002625375,0.006210117,0.0004439712,0.001934892],"study_design_scores_gemma":[0.002670481,0.0001389009,0.1764921,0.00144513,0.0001883663,0.00007157381,0.3688538,0.4096285,0.0002443343,0.01160854,0.02640326,0.002255055],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9748257,0.00004358898,0.02355181,0.000236265,0.0003423913,0.0002314602,0.0001756389,0.00001937791,0.0005737013],"genre_scores_gemma":[0.9986168,0.00007916093,0.000808985,0.000140828,0.000177489,0.000001345686,0.0001514881,0.000005709075,0.00001815265],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.797527,"threshold_uncertainty_score":0.9991676,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2125590838","doi":"10.2481/dsj.4.106","title":"The long-term preservation of accurate and authentic digital data: the INTERPARES project","year":2005,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":46,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Scope (computer science); Computer science; Usability; Transparency (behavior); Implementation; Metadata; Digital preservation; Open data; Data science; Reuse; World Wide Web; Data publishing; Electronic publishing; Publishing; Open science; Data curation; Political science; Software engineering; Engineering; The Internet; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.1383611772689184,"gpt":0.3153712717204733,"spread":0.1770100944515549,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0006010377,0.00005217662,0.00004470365,0.00004623296,0.0008391248,0.003074821,0.002039686,0.000001990089,0.00001614487],"category_scores_gemma":[0.00009573165,0.00002337088,0.00001094142,0.00004708541,0.001239047,0.01115321,0.001179028,0.0000805684,0.000005583385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000007263287,"about_ca_system_score_gemma":0.00006701525,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007380647,"about_ca_topic_score_gemma":0.00009935794,"domain_scores_codex":[0.9991996,0.00001087382,0.0001768641,0.0001483735,0.0003319795,0.0001322699],"domain_scores_gemma":[0.9992553,0.00008725054,0.000125036,0.0004563351,0.00004769373,0.00002840515],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005490146,0.0002129315,0.001634297,0.00003198324,0.00009267937,0.000006438699,0.009719093,0.00001263478,0.00002119266,0.4779778,0.03215406,0.478082],"study_design_scores_gemma":[0.0003588116,0.0001840977,0.1382102,0.0001682535,0.00005863684,0.0001176531,0.00351674,0.01732137,0.000018211,0.02022396,0.8195981,0.0002239937],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8409565,0.0004070523,0.0006918022,0.01452676,0.001101797,0.0006038491,0.004029919,0.00003146939,0.1376509],"genre_scores_gemma":[0.9984393,0.0001128123,0.00002406557,0.00006143337,0.0003057558,0.000001196558,0.0001831411,0.000002238743,0.0008701152],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7874441,"threshold_uncertainty_score":0.9979601,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2891686592","doi":"10.5334/dsj-2018-020","title":"Using Classified and Unclassified Land Cover Data to Estimate the Footprint of Human Settlement","year":2018,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Impact of Light on Environment and Health","field":"Environmental Science","cited_by":45,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Esri (Canada)","funders":"","keywords":"Geography; Population; Footprint; Land cover; Cartography; Settlement (finance); Orthophoto; Ancillary data; Georeference; Census; Remote sensing; Land use; Computer science; Physical geography; Archaeology; Ecology; World Wide Web; Demography","retraction":null,"screen_n_in":null,"score":{"opus":0.2387426298426108,"gpt":0.4369638167012191,"spread":0.1982211868586084,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004134521,0.0001082535,0.0001288562,0.00005799741,0.001101543,0.0001956278,0.002527294,0.00002359419,0.0006476028],"category_scores_gemma":[0.0001035781,0.00006728544,0.000009488956,0.0002760113,0.001452478,0.001020496,0.002937724,0.0001646692,0.00005904817],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001599346,"about_ca_system_score_gemma":0.00008982723,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001615523,"about_ca_topic_score_gemma":0.0001770816,"domain_scores_codex":[0.9979424,0.00004319806,0.000305061,0.0004147622,0.0008305092,0.0004640447],"domain_scores_gemma":[0.9981157,0.00002684542,0.0002048832,0.001318809,0.00001123365,0.0003225186],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00004901968,0.0001299974,0.6243751,0.000008937919,0.00001502845,0.000006882702,0.0009675651,0.0001622054,0.3593679,0.0002353468,0.006989289,0.0076927],"study_design_scores_gemma":[0.000379484,0.0002633452,0.9581269,0.00003374194,0.00002802046,0.0000809878,0.0001439768,0.01217866,0.003296906,0.0002128703,0.02508836,0.0001667398],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9942934,0.00001094614,0.001488734,0.001907212,0.0001851506,0.0001635429,0.000146377,0.000005286494,0.001799365],"genre_scores_gemma":[0.9908052,0.00001681447,0.008605475,0.000366142,0.0001324352,3.809013e-7,0.0000203058,0.000006514644,0.00004670237],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.356071,"threshold_uncertainty_score":0.8472292,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1978333903","doi":"10.2481/dsj.1.45","title":"A Mexican case study on a centralised database from world natural history museums","year":2002,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":33,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Distribution (mathematics); Geography; Ornithology; Natural history; Species richness; Endemism; Database; Collections management; Library science; Ethnology; Ecology; Archaeology; History; Computer science; Biology; Southern Hemisphere","retraction":null,"screen_n_in":null,"score":{"opus":0.1465141139966224,"gpt":0.3052122749726442,"spread":0.1586981609760218,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.0008370852,0.0001551797,0.0001354755,0.0001131423,0.0006013488,0.0002198819,0.001760869,0.00001633015,0.131237],"category_scores_gemma":[0.0001516482,0.0001240367,0.0000337977,0.0006518186,0.0006603919,0.002014641,0.0008182747,0.0004045866,0.001256145],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00198072,"about_ca_system_score_gemma":0.0000297971,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00174713,"about_ca_topic_score_gemma":0.003861384,"domain_scores_codex":[0.9974725,0.0000663614,0.0002640958,0.0006007298,0.001092289,0.0005040583],"domain_scores_gemma":[0.9982836,0.0000411567,0.0001526381,0.001065921,0.00001657545,0.000440077],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004856466,0.00219683,0.05145061,0.000001707332,0.00002262252,0.01028361,0.004184839,0.000009404587,0.01292112,0.00009869719,0.9129282,0.005853851],"study_design_scores_gemma":[0.004073201,0.0004499004,0.2677686,0.00005259111,0.0001213675,0.006229777,0.05635346,0.01229061,0.000551304,0.00002645001,0.6508609,0.001221796],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.98927,0.0001106885,0.00001421678,0.0005988429,0.001204034,0.0001619935,0.0008938364,0.00003216529,0.007714191],"genre_scores_gemma":[0.9983959,0.00003386961,0.0001157531,0.0008747341,0.0001062098,0.000002591499,0.0001120216,0.000007164209,0.0003517467],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2620672,"threshold_uncertainty_score":0.9995215,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2133007173","doi":"10.2481/dsj.4.127","title":"Multi-sensor data fusion for land vehicle attitude estimation using a fuzzy expert system","year":2005,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Inertial Sensor and Navigation","field":"Engineering","cited_by":31,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Gyroscope; Accelerometer; Kalman filter; Sensor fusion; Compass; Computer science; Inertial measurement unit; Fuzzy logic; Dead reckoning; Vibrating structure gyroscope; Artificial intelligence; Position (finance); Field (mathematics); Inertial navigation system; Global Positioning System; Computer vision; Control theory (sociology); Real-time computing; Inertial frame of reference; Engineering; Mathematics; Geography; Telecommunications; Aerospace engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.1043321370756325,"gpt":0.3515910503201103,"spread":0.2472589132444778,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000987898,0.00008854179,0.00009877604,0.0001005525,0.0004528069,0.000266168,0.0009498479,0.00003682464,0.000005856706],"category_scores_gemma":[0.0001181165,0.00007548454,0.00001390081,0.0002287601,0.0000501258,0.002983,0.0001848482,0.0001163746,0.00001533748],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001765838,"about_ca_system_score_gemma":0.00005139531,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003849059,"about_ca_topic_score_gemma":0.00002917913,"domain_scores_codex":[0.9989182,0.00001212966,0.0002493616,0.0002331319,0.0003254593,0.0002616543],"domain_scores_gemma":[0.9991177,0.00002305234,0.00006397993,0.0006032234,0.00008379332,0.00010829],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000247422,0.00004337715,0.0005533852,0.00006181093,0.00001287733,0.000007193788,0.0003632848,0.2213658,0.7274735,0.00002458036,0.002252854,0.04781663],"study_design_scores_gemma":[0.0003843583,0.000009696599,0.0007336836,0.00008315879,0.00001180654,0.0001153629,0.00008194189,0.9918023,0.004314865,0.000001438007,0.002356883,0.0001045535],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6659661,0.0003000441,0.3317428,0.0001374902,0.0009703157,0.0002378339,0.0004588946,0.0001248484,0.00006165227],"genre_scores_gemma":[0.7768784,0.00003004973,0.2221823,0.00002302032,0.0006313758,9.723291e-7,0.000237712,0.00001214934,0.000004005419],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7704365,"threshold_uncertainty_score":0.348267,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2585472482","doi":"10.5334/dsj-2017-003","title":"Legal and Ethical Issues around Incorporating Traditional Knowledge in Polar Data Infrastructures","year":2017,"lang":"en","type":"article","venue":"Data Science Journal","topic":"International Maritime Law Issues","field":"Environmental Science","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University; University of Ottawa","funders":"","keywords":"Traditional knowledge; Acknowledgement; Knowledge management; Interoperability; Context (archaeology); Sociology of scientific knowledge; Knowledge sharing; Inclusion (mineral); Body of knowledge; Computer science; Engineering ethics; Indigenous; Sociology; Social science; World Wide Web; Engineering; Computer security; Geography","retraction":null,"screen_n_in":null,"score":{"opus":0.07870823258626229,"gpt":0.3662956398552014,"spread":0.2875874072689391,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication","open_science"],"consensus_categories":[],"category_scores_codex":[0.003575016,0.000120407,0.0001343467,0.00008433343,0.001451489,0.001865201,0.005788098,0.00007899713,0.0005766268],"category_scores_gemma":[0.001313209,0.0001037132,0.00001001189,0.000129588,0.002577154,0.007557402,0.004914184,0.0008427703,0.00006335417],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001390311,"about_ca_system_score_gemma":0.0001614084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001978735,"about_ca_topic_score_gemma":0.002346664,"domain_scores_codex":[0.9978966,0.00005969519,0.0002966865,0.0005651304,0.0008864232,0.0002955166],"domain_scores_gemma":[0.9982844,0.00007926808,0.0002317109,0.001195249,0.0000283306,0.0001810598],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00004575794,0.0001704166,0.8215747,0.00001330194,0.00001746899,0.0003044931,0.0007296831,0.0001577078,0.01433233,0.1003501,0.03898946,0.02331458],"study_design_scores_gemma":[0.0003002,0.00003393197,0.916895,0.00005548968,0.000006026657,0.000733214,0.0001517237,0.0313087,0.0001486983,0.02691956,0.02325012,0.0001972804],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9643266,0.0001542717,0.0002384602,0.006707413,0.0007382738,0.0001096629,0.0007520493,0.0000147654,0.02695855],"genre_scores_gemma":[0.9831789,0.00003254638,0.01594931,0.0002229645,0.0004305294,6.06873e-7,0.0001260281,0.000006146502,0.00005300186],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.09532033,"threshold_uncertainty_score":0.9998485,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2141093030","doi":"10.2481/dsj.3.202","title":"Appraising digital records for long-term preservation","year":2004,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Scope (computer science); Computer science; Digital preservation; Usability; Transparency (behavior); Metadata; Implementation; Open data; World Wide Web; Data science; Reuse; Data curation; Data publishing; Electronic publishing; Publishing; Political science; Software engineering; The Internet; Computer security; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.1352581856858978,"gpt":0.3013021351824907,"spread":0.1660439494965929,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0002986707,0.00006120263,0.00005408098,0.0001075579,0.0007667891,0.003097451,0.0008317063,0.000003119873,0.00004804169],"category_scores_gemma":[0.00007477966,0.00004698223,0.00002793282,0.0000508993,0.0003953896,0.01229572,0.0002102138,0.00006432678,0.0000268885],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003085606,"about_ca_system_score_gemma":0.0001027945,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004467136,"about_ca_topic_score_gemma":0.00001849017,"domain_scores_codex":[0.9991291,0.000001948864,0.0001568321,0.0001877499,0.0003267768,0.0001975593],"domain_scores_gemma":[0.9995371,0.00002385352,0.00008332559,0.0001951264,0.00007889087,0.00008170452],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00002503786,0.0002051465,0.000566358,0.00002411439,0.00002065374,0.00001162415,0.001496654,0.00008172121,0.00002172711,0.8512712,0.004269432,0.1420063],"study_design_scores_gemma":[0.001197671,0.0004221939,0.07667181,0.0002198342,0.00003272942,0.0001379987,0.0006614539,0.0007757785,0.00004507093,0.4855567,0.4338547,0.0004240657],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.4352731,0.0000343514,0.05849253,0.005652582,0.003204986,0.0005021509,0.002288348,0.0000878168,0.4944641],"genre_scores_gemma":[0.9964588,0.000008269363,0.0009111961,0.0001694228,0.0009237953,0.000002929451,0.0004393305,0.000005401108,0.001080827],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5611857,"threshold_uncertainty_score":0.9979374,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2123356220","doi":"10.2481/dsj.5.174","title":"The Crystallographic Information File (CIF)","year":2006,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University; Brockhouse Institute for Materials Research","funders":"","keywords":"Computer science; Scope (computer science); Transparency (behavior); Usability; Open data; Metadata; World Wide Web; Data science; Reuse; Data publishing; Data curation; Implementation; Software; Publishing; Political science; Software engineering; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.01438017795474202,"gpt":0.2244732268565835,"spread":0.2100930489018415,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.00208839,0.00005889447,0.00004426562,0.00007852489,0.001586647,0.001764883,0.005072091,0.0000202893,0.00008145432],"category_scores_gemma":[0.0004003712,0.00003607815,0.00001924482,0.0008630502,0.0003303222,0.006450942,0.0009428125,0.0001893019,0.00003110138],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001625294,"about_ca_system_score_gemma":0.0002109128,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001840131,"about_ca_topic_score_gemma":0.000007347141,"domain_scores_codex":[0.9988,0.0000169234,0.000217235,0.0001557569,0.0004915458,0.0003185547],"domain_scores_gemma":[0.9986336,0.00007030243,0.0001669518,0.0008569684,0.0002006249,0.00007156489],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000003729475,0.00003780509,0.00152476,0.00000680055,0.000005811754,0.00001948101,0.0002033477,0.0009088664,0.002711408,0.05737445,0.8562754,0.08092816],"study_design_scores_gemma":[0.00007136831,0.00001292816,0.01128023,0.000007501288,0.000001097934,0.0003736666,0.00006927886,0.04582079,0.0001762957,0.01429427,0.9278135,0.00007904135],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01389526,0.0004131574,0.6820613,0.04730189,0.003323414,0.0003236115,0.0003480311,0.0003303547,0.252003],"genre_scores_gemma":[0.9635534,0.00008047044,0.03444013,0.0006032989,0.0003846908,0.000004542931,0.0001587397,0.000001264135,0.0007734713],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9496582,"threshold_uncertainty_score":0.9997131,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3127642832","doi":"10.5334/dsj-2021-007","title":"Stewardship Maturity Assessment Tools for Modernization of Climate Data Management","year":2021,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Climate variability and models","field":"Environmental Science","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Environment and Climate Change Canada","funders":"Environment and Climate Change Canada; Centrum fÖr Personcentrerad Vård; National Centers for Environmental Information; National Oceanic and Atmospheric Administration; Grains Research and Development Corporation; National Aeronautics and Space Administration","keywords":"Stewardship (theology); Data management; Maturity (psychological); Scope (computer science); Data quality; Computer science; Process (computing); Quality (philosophy); Usability; Process management; Environmental resource management; Business; Data science; Database; Environmental science; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.132977886381932,"gpt":0.3695392086253232,"spread":0.2365613222433913,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00494847,0.00008953389,0.0001385919,0.00004126104,0.0004582422,0.000427813,0.002660736,0.00002942476,0.0006552727],"category_scores_gemma":[0.000304637,0.00007914195,0.00002423686,0.0004094121,0.0002788423,0.00517791,0.004904248,0.0001276942,0.00001233571],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001565725,"about_ca_system_score_gemma":0.0001256997,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000140905,"about_ca_topic_score_gemma":0.00003685146,"domain_scores_codex":[0.9978228,0.00003834646,0.0003633086,0.0006051083,0.0008156914,0.0003547359],"domain_scores_gemma":[0.9976196,0.0001154981,0.0001872854,0.001894491,0.00005420211,0.0001289696],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002350684,0.003458559,0.08252691,0.0008736347,0.0002066396,0.0001256137,0.00092652,0.1227513,0.07437627,0.02928841,0.02493816,0.6602929],"study_design_scores_gemma":[0.0006390645,0.00004659519,0.0560965,0.0000632916,0.0001068636,0.00008169524,0.0005826178,0.9182319,0.0008006504,0.003402154,0.01971875,0.0002299384],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2265876,0.0001228593,0.7437834,0.002394564,0.001555833,0.0009831034,0.008196896,0.00004619704,0.01632952],"genre_scores_gemma":[0.6832484,0.002417604,0.312406,0.0002412589,0.00008525471,0.000006712516,0.001523407,0.00001295582,0.00005841366],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7954805,"threshold_uncertainty_score":0.7174777,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3111376437","doi":"10.5334/dsj-2020-047","title":"39 Hints to Facilitate the Use of Semantics for Data on Agriculture and Nutrition","year":2020,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Agriculture and Agri-Food Canada; Rural Development Administration; Agence Nationale de la Recherche; Ministry of Agriculture of the People's Republic of China; Department for International Development; Bill and Melinda Gates Foundation","keywords":"Computer science; Semantic interoperability; Interoperability; Conceptualization; Linked data; Knowledge management; Data science; Semantics (computer science); World Wide Web; Data sharing; Standardization; Semantic Web","retraction":null,"screen_n_in":null,"score":{"opus":0.2965619348551438,"gpt":0.3499252635567615,"spread":0.0533633287016177,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004380371,0.00004694858,0.00006122729,0.00001376747,0.0001457805,0.00008163723,0.001029871,0.00003244594,0.000001268631],"category_scores_gemma":[0.001631234,0.00002511802,0.00000977222,0.0001098217,0.000254642,0.00002258309,0.0006291388,0.00007208018,7.32055e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000002275094,"about_ca_system_score_gemma":0.00004545297,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005255373,"about_ca_topic_score_gemma":0.000006337761,"domain_scores_codex":[0.999363,0.00001562962,0.0001060267,0.000240675,0.0001598228,0.0001148663],"domain_scores_gemma":[0.9993479,0.00002865791,0.0000513654,0.0004004735,0.00006729506,0.000104272],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000130384,0.00005291786,0.0002454911,0.00002975654,0.00001750493,0.000002138698,0.0002572611,0.00001725183,0.4129091,0.00002622017,0.5379817,0.04833025],"study_design_scores_gemma":[0.0002423488,0.0005657869,0.001119058,0.00003063188,0.00001228429,0.00003567033,0.0003419913,0.002643021,0.009223532,0.00002380636,0.9856941,0.00006784405],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8424088,0.0008228849,0.08568249,0.06267826,0.0004870404,0.0005510685,0.007323998,0.0000127069,0.00003270734],"genre_scores_gemma":[0.8966928,0.001437364,0.09337672,0.005962473,0.0007835758,0.000004925877,0.001625632,0.000009126178,0.0001074253],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4477123,"threshold_uncertainty_score":0.1952857,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4301396488","doi":"10.5334/dsj-2022-017","title":"A Survey on Publicly Available Open Datasets Derived From Electronic Health Records (EHRs) of Patients with Neuroblastoma","year":2022,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Neuroblastoma Research and Treatments","field":"Medicine","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Neuroblastoma; Computer science; Health records; Open science; License; The Internet; Electronic health record; Open data; World Wide Web; Internet privacy; Medicine; Information retrieval; Data mining; Political science; Mathematics; Statistics; Health care","retraction":null,"screen_n_in":null,"score":{"opus":0.07305559528067233,"gpt":0.3504970068041415,"spread":0.2774414115234691,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002857057,0.0001493583,0.0003222227,0.0003372727,0.0009012067,0.0002854741,0.002741999,0.00001149033,0.001453],"category_scores_gemma":[0.0004917475,0.000109303,0.00002019123,0.001281552,0.0002250439,0.001605544,0.002434728,0.0006999192,0.00004382752],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004026068,"about_ca_system_score_gemma":0.005009867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00360217,"about_ca_topic_score_gemma":0.001634632,"domain_scores_codex":[0.9959581,0.0003654963,0.0003506356,0.0007080496,0.001834701,0.000782961],"domain_scores_gemma":[0.9972745,0.0001210484,0.0003709691,0.00143447,0.0002028244,0.0005961609],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.008404824,0.00473583,0.6944162,0.00001703237,0.0001862987,0.0002755148,0.00008853035,0.00002532378,0.001545126,0.00001421331,0.2717462,0.01854486],"study_design_scores_gemma":[0.004738842,0.01035989,0.9714262,0.00004344098,0.00001677347,0.00008529951,0.00003786686,0.0003308833,0.0005484685,0.00001275752,0.01228374,0.0001158507],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9797359,0.00004012203,0.00001694145,0.0006629621,0.0001290784,0.000570018,0.01870537,0.000009823865,0.0001297909],"genre_scores_gemma":[0.9890721,0.00007237559,0.0004722965,0.0008504962,0.00002497538,0.00001016863,0.009386744,0.00002036718,0.00009043954],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.27701,"threshold_uncertainty_score":0.9994598,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2154198926","doi":"10.2481/dsj.14-047","title":"Data-PE: A Framework for Evaluating Data Publication Policies at Scholarly Journals","year":2015,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Scope (computer science); Usability; Transparency (behavior); Metadata; Open data; Data science; Data publishing; Implementation; Reuse; Publishing; World Wide Web; Political science; Software engineering; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.7646680752783171,"gpt":0.5850843913237225,"spread":0.1795836839545947,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication","open_science"],"consensus_categories":["metaresearch","scholarly_communication","open_science"],"category_scores_codex":[0.09832495,0.0002101452,0.0002424734,0.0008999013,0.002200963,0.07931364,0.09842829,0.00006954568,0.00004324852],"category_scores_gemma":[0.1405156,0.0001723954,0.00002327301,0.003131325,0.0004518173,0.5295488,0.06948273,0.001027364,0.0001308629],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003573455,"about_ca_system_score_gemma":0.003144836,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006507427,"about_ca_topic_score_gemma":0.00004087663,"domain_scores_codex":[0.9905491,0.0004240604,0.0008371482,0.001898546,0.005167422,0.00112372],"domain_scores_gemma":[0.9675272,0.001252176,0.001725026,0.02474114,0.003322104,0.001432369],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00006281883,0.0002409634,0.002266322,0.00003132016,0.0001000675,0.00004661496,0.0005304582,0.00008508888,0.001376014,0.1026383,0.7540167,0.1386053],"study_design_scores_gemma":[0.0005222979,0.0001618431,0.001782224,0.00008859458,0.00003761835,0.0006855375,0.0004032616,0.4110035,0.00004463927,0.009835939,0.5751445,0.0002899849],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001301176,0.0008485469,0.9456808,0.04713186,0.00106938,0.0004595178,0.002694956,0.00007401028,0.0007397709],"genre_scores_gemma":[0.005308094,0.001001022,0.9880139,0.00172198,0.001014589,0.00001107501,0.002550244,0.00002000188,0.0003591258],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4502352,"threshold_uncertainty_score":0.9990981,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2032869846","doi":"10.2481/dsj.6.od26","title":"Canadian National Consultation on Access to Scientific Research Data","year":2007,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"National Research Council Canada; Université de Montréal","funders":"National Research Council Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Strong","keywords":"Scope (computer science); Computer science; Transparency (behavior); Usability; Metadata; Data science; Open data; Reuse; Implementation; World Wide Web; Engineering; Computer security; Software engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.6443846780180462,"gpt":0.5804227013872811,"spread":0.06396197663076508,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication","open_science"],"consensus_categories":["metaresearch","scholarly_communication","open_science"],"category_scores_codex":[0.1153449,0.0001042604,0.00008555254,0.004424845,0.003721329,0.05571972,0.0610731,0.00002882001,0.000042437],"category_scores_gemma":[0.0286426,0.00009212034,0.000008597463,0.009378552,0.0006809213,0.1791516,0.01571452,0.0007426076,0.0005058338],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009058783,"about_ca_system_score_gemma":0.008137026,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006504286,"about_ca_topic_score_gemma":0.06561691,"domain_scores_codex":[0.9893884,0.0001973602,0.0003665139,0.001566692,0.007161078,0.001319925],"domain_scores_gemma":[0.9892991,0.001009704,0.0001581076,0.005559531,0.002416912,0.001556627],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003228793,0.000146129,0.001204467,0.000008186363,0.00001602259,0.0003074604,0.0003529262,0.0002003047,0.001937291,0.1968256,0.683045,0.1159243],"study_design_scores_gemma":[0.0002300197,0.00008314685,0.0227707,0.00004181271,0.000002021133,0.0001048072,0.0002524629,0.03378573,0.0004238111,0.0018177,0.9402866,0.0002011373],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02979269,0.00008459127,0.6988483,0.1551097,0.007028805,0.001855234,0.002705235,0.0001342414,0.1044412],"genre_scores_gemma":[0.8362272,0.0002277161,0.1535375,0.004962541,0.00110115,0.000009323553,0.001544237,0.00002517569,0.002365097],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8064345,"threshold_uncertainty_score":0.9975757,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1985018337","doi":"10.2481/dsj.4.21","title":"Rescuing and recovering lost or endangered data","year":2005,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Digital and Cyber Forensics","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dominion Astrophysical Observatory; Herzberg Institute of Astrophysics","funders":"","keywords":"Scope (computer science); Computer science; Transparency (behavior); Open data; Metadata; Usability; Reuse; Data science; Implementation; World Wide Web; Data curation; Open science; Engineering; Software engineering; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.1006452456693089,"gpt":0.3114684138730922,"spread":0.2108231682037833,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.001840256,0.00009339402,0.0001000224,0.0001196775,0.0004126667,0.002539361,0.007528788,0.00001803821,0.00002196971],"category_scores_gemma":[0.0002795089,0.00006570332,0.00000888563,0.000621687,0.0002886826,0.02111246,0.00657994,0.000193836,0.00004820587],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004215307,"about_ca_system_score_gemma":0.0003369909,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007468132,"about_ca_topic_score_gemma":0.00006239803,"domain_scores_codex":[0.998228,0.00001222256,0.0002209034,0.0005769674,0.0005829153,0.0003789991],"domain_scores_gemma":[0.9976349,0.00004189458,0.0001026348,0.001900333,0.0000643173,0.0002559084],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000005843674,0.00002481808,0.0003033376,0.000002959497,0.000004526695,0.00005407867,0.0002292279,0.00001282551,0.0006239468,0.003042965,0.01335666,0.9823388],"study_design_scores_gemma":[0.0005524473,0.0001297205,0.005094686,0.0001471533,0.00001208146,0.005457228,0.0001905631,0.3062987,0.001317035,0.002794568,0.6775013,0.0005045437],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2837887,0.002513954,0.611541,0.02397785,0.006986304,0.0005015326,0.001778075,0.0004200705,0.06849254],"genre_scores_gemma":[0.535657,0.0007955978,0.4597428,0.001546102,0.001247856,5.037006e-7,0.00007665662,0.00001658468,0.0009168095],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9818343,"threshold_uncertainty_score":0.9984961,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4386830764","doi":"10.5334/dsj-2023-035","title":"Umbrella Data Management Plans to Integrate FAIR Data: Lessons From the ISIDORe and BY-COVID Consortia for Pandemic Preparedness","year":2023,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer","funders":"Biotechnology and Biological Sciences Research Council","keywords":"Preparedness; Data management; Computer science; Multidisciplinary approach; Plan (archaeology); Download; Process (computing); Data management plan; Data science; Process management; Knowledge management; Engineering management; World Wide Web; Business; Engineering; Political science; Geography; Database","retraction":null,"screen_n_in":null,"score":{"opus":0.4303456883339171,"gpt":0.4883463625022655,"spread":0.05800067416834848,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.01811445,0.0002092041,0.0002069868,0.0002996529,0.001474167,0.01234649,0.0605602,0.00003372423,0.00001125888],"category_scores_gemma":[0.003022712,0.0001406859,0.00001544136,0.001915669,0.0005205619,0.0574757,0.05872685,0.0004146619,0.00006275514],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008032258,"about_ca_system_score_gemma":0.0005390171,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002910757,"about_ca_topic_score_gemma":0.0006445327,"domain_scores_codex":[0.9946046,0.0001963095,0.0004569856,0.002384598,0.001505197,0.0008522715],"domain_scores_gemma":[0.9871954,0.001069429,0.0002923318,0.01085734,0.0001168116,0.0004686951],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003812247,0.00004640682,0.001562356,0.00002544203,0.00009817055,0.00008164605,0.0004650117,0.00004189485,0.0005124847,0.008265208,0.9272674,0.0615958],"study_design_scores_gemma":[0.000470265,0.00005477865,0.003738831,0.00008534134,0.00005596958,0.00006942753,0.003096505,0.1567426,0.00001987293,0.001256254,0.8341447,0.0002653818],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007924392,0.0003826584,0.895521,0.0454405,0.001163222,0.001249609,0.04769893,0.0001964046,0.0004232271],"genre_scores_gemma":[0.33051,0.07910057,0.4685426,0.02039512,0.002579436,0.0003121855,0.09261571,0.0002059003,0.005738525],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4269785,"threshold_uncertainty_score":0.9998258,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W4402883940","doi":"10.5334/dsj-2024-046","title":"Knowledge Infrastructures Are Growing Up: The Case for Institutional (Data) Repositories 10 Years After the Holdren Memo","year":2024,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"University at Buffalo; University of Minnesota","keywords":"Computer science; Business; World Wide Web; Data science","retraction":null,"screen_n_in":null,"score":{"opus":0.1327752469239644,"gpt":0.407377605795062,"spread":0.2746023588710976,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.009147289,0.0001288822,0.0000949206,0.0002398749,0.002061167,0.02548499,0.01762145,0.00002665378,0.00002635601],"category_scores_gemma":[0.002421483,0.00006868741,0.00003661862,0.001369077,0.0009493338,0.1035825,0.008841521,0.0005360123,0.0000277457],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009268914,"about_ca_system_score_gemma":0.001038491,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004153354,"about_ca_topic_score_gemma":0.0000939319,"domain_scores_codex":[0.9973193,0.0001432988,0.0002900816,0.0007977821,0.0009836063,0.0004658871],"domain_scores_gemma":[0.9955194,0.0005715686,0.0001433678,0.00342157,0.0002005066,0.0001435883],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007762791,0.00006396609,0.0005016797,0.000113775,0.000210945,0.007416906,0.002472865,0.000250086,0.0005179262,0.1671872,0.6434868,0.1777003],"study_design_scores_gemma":[0.000132032,0.00003317324,0.004543498,0.00008413425,0.0000380173,0.01068298,0.0006824459,0.08267478,0.00008837073,0.001839931,0.8990363,0.0001642667],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08161543,0.01645269,0.7887615,0.07082103,0.03063291,0.002113395,0.005332223,0.0004016923,0.003869097],"genre_scores_gemma":[0.9702263,0.0005855712,0.02438929,0.0005651977,0.002928636,0.00003884304,0.0001650025,0.00001763522,0.00108347],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8886109,"threshold_uncertainty_score":0.999238,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2785284360","doi":"10.5334/dsj-2018-001","title":"Science Metadata Management, Interoperability and Data Citations of the National Institute of Polar Research, Japan","year":2018,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Metadata; Data management plan; Interoperability; Directory; Stewardship (theology); Data center; Meta Data Services; Open science; Data management; Data mapping; Library science; Data as a service; Research center; World Wide Web; Metadata management; Computer science; Metadata repository; Database; Political science; Business; Service (business)","retraction":null,"screen_n_in":null,"score":{"opus":0.478091744716791,"gpt":0.5185561253619065,"spread":0.04046438064511554,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication","open_science"],"consensus_categories":["metaresearch","sts","scholarly_communication","open_science"],"category_scores_codex":[0.05292474,0.00009198726,0.0001268916,0.001296897,0.001901981,0.004876843,0.05046319,0.00001479925,0.00001072559],"category_scores_gemma":[0.01249736,0.00006125085,0.00001391077,0.007689485,0.01365335,0.1956799,0.05490604,0.0004173805,0.000005293442],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00010056,"about_ca_system_score_gemma":0.001952438,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001073867,"about_ca_topic_score_gemma":0.000137663,"domain_scores_codex":[0.9929287,0.0002789827,0.0004714578,0.001070055,0.004750406,0.0005004175],"domain_scores_gemma":[0.9908152,0.0002142397,0.0003806398,0.006241896,0.002130767,0.0002172096],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.00001950564,0.0003055972,0.007960822,0.00006809953,0.00006865123,0.000004177844,0.0005141192,0.000008422713,0.02220159,0.9405392,0.01020442,0.01810537],"study_design_scores_gemma":[0.001237699,0.0006465832,0.4873074,0.0004007885,0.00006800935,0.0003309563,0.002538114,0.1315599,0.00615588,0.02218223,0.3470151,0.000557267],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4610418,0.0003780674,0.4570993,0.03166175,0.003859209,0.001998345,0.004233794,0.00006445283,0.03966329],"genre_scores_gemma":[0.8977801,0.0003427218,0.1014898,0.00011068,0.00007961562,0.000001388287,0.00002984631,0.000003382027,0.0001624768],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.918357,"threshold_uncertainty_score":0.9993974,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2059472642","doi":"10.2481/dsj.5.143","title":"A hashing technique using separate binary tree","year":2006,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Scope (computer science); Usability; Data science; Metadata; Transparency (behavior); Data publishing; Implementation; World Wide Web; Publishing; Software engineering; Computer security; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.05802956779394889,"gpt":0.3299944953690478,"spread":0.2719649275750989,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":[],"category_scores_codex":[0.002588516,0.0001259706,0.0001264872,0.0003222381,0.001201362,0.001299588,0.006534043,0.0000371168,0.00001202778],"category_scores_gemma":[0.00004912259,0.00009739359,0.00002722336,0.001221676,0.0002285357,0.01226881,0.003068991,0.0003168499,0.00001534941],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009391203,"about_ca_system_score_gemma":0.0005358025,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001223061,"about_ca_topic_score_gemma":0.000005678585,"domain_scores_codex":[0.9978426,0.00004508674,0.0003027551,0.0005530382,0.0007848041,0.0004717178],"domain_scores_gemma":[0.998046,0.00002719993,0.0002136016,0.001441359,0.000106675,0.0001651312],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001402931,0.0003265747,0.002089148,0.00001247582,0.00000922145,0.0008866977,0.0002461896,0.003949196,0.8668347,0.01081442,0.03874526,0.07607212],"study_design_scores_gemma":[0.0001825495,0.00004637907,0.002447165,0.00008131468,0.000004581583,0.002751757,0.00002367483,0.9755055,0.005683639,0.003290381,0.009747108,0.0002359116],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02621412,0.0001254431,0.9719428,0.0002097897,0.0005360648,0.00008537449,0.0000266446,0.00007140768,0.0007883479],"genre_scores_gemma":[0.197221,0.00001573886,0.8021817,0.00009900163,0.0004082565,0.000001652107,0.00002195927,0.000007523606,0.00004315561],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9715564,"threshold_uncertainty_score":0.9997371,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2021867077","doi":"10.2481/dsj.8.27","title":"CISTI'S Activities in Support of Scientific Data Management in Canada 2008-2010","year":2009,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"National Research Council Canada","funders":"","keywords":"Scope (computer science); Computer science; Transparency (behavior); Usability; Metadata; Implementation; Open data; Reuse; Data management; Data science; World Wide Web; Publishing; Data publishing; Political science; Database; Engineering; Software engineering; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.1662144302179339,"gpt":0.3827343700927204,"spread":0.2165199398747866,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.01396079,0.0001343779,0.0002001972,0.001347581,0.0003092705,0.004411082,0.03220082,0.00001611021,0.00004062261],"category_scores_gemma":[0.0006568268,0.0001222156,0.00001037413,0.003705725,0.0003959801,0.1023877,0.009457158,0.0004597836,0.000006130207],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004685907,"about_ca_system_score_gemma":0.004557692,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.08816091,"about_ca_topic_score_gemma":0.4507587,"domain_scores_codex":[0.9946444,0.0001158097,0.0005752335,0.001141552,0.002702567,0.0008204466],"domain_scores_gemma":[0.9935417,0.0001030923,0.000337765,0.00570392,0.00008607248,0.0002274548],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"observational","study_design_scores_codex":[0.00006495135,0.0008541809,0.02460271,0.00008813247,0.00004096268,0.004332749,0.0003972666,0.0006636226,0.002638246,0.03760831,0.5165929,0.412116],"study_design_scores_gemma":[0.001581795,0.00019692,0.4637489,0.0001697189,0.00001937386,0.0003598611,0.001597063,0.1239855,0.0007930856,0.002060679,0.4047073,0.0007798312],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5688199,0.001243265,0.234402,0.09266604,0.01877156,0.003990431,0.00684258,0.0001701585,0.07309406],"genre_scores_gemma":[0.9353852,0.001132012,0.06177567,0.0005806049,0.00006222464,0.000002549104,0.0003366961,0.000007096435,0.0007179871],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4391462,"threshold_uncertainty_score":0.9985542,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2031989156","doi":"10.2481/dsj.2.25","title":"Data integration and knowledge discovery in biomedical databases. Reliable information from unreliable sources","year":2003,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Artificial Intelligence in Healthcare","field":"Health Professions","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"","keywords":"Scope (computer science); Computer science; Data science; Usability; Metadata; Transparency (behavior); Open data; Reuse; Implementation; World Wide Web; Data discovery; Data publishing; Publishing; Software engineering; Political science; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.2754938637313639,"gpt":0.5016500590355036,"spread":0.2261561953041397,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.009279463,0.0001363507,0.0002102831,0.0004363615,0.001773008,0.0002811909,0.001959139,0.0001049144,0.0002932659],"category_scores_gemma":[0.008379996,0.0001061065,0.000008638024,0.001120645,0.0005607523,0.02142639,0.001526225,0.001204732,0.0002505665],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002568796,"about_ca_system_score_gemma":0.0034468,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006725489,"about_ca_topic_score_gemma":0.007182514,"domain_scores_codex":[0.9968604,0.0003112753,0.00105772,0.0004868684,0.0006409665,0.0006427743],"domain_scores_gemma":[0.9967422,0.0006503429,0.0004001916,0.001601444,0.0002583877,0.0003474572],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001655665,0.0004155617,0.6449422,0.0002440427,0.00001861051,0.00003016296,0.0207428,0.00006320696,0.00203891,0.01804375,0.2351105,0.07818463],"study_design_scores_gemma":[0.00054743,0.00008433536,0.01476497,0.00153197,0.00002356034,0.00004109928,0.0609754,0.1445456,0.0003451213,0.00424817,0.772508,0.0003842963],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9100146,0.002224475,0.05944523,0.003495337,0.008950341,0.001187014,0.009279273,0.00008636026,0.005317365],"genre_scores_gemma":[0.9623731,0.002681874,0.02550796,0.00125341,0.000813539,0.00001865948,0.007174637,0.0000220284,0.0001548111],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6301773,"threshold_uncertainty_score":0.9999728,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2955083888","doi":"10.5334/dsj-2019-027","title":"Developing a Model Guidelines Addressing Legal Impediments to Open Access to Publicly Funded Research Data in Malaysia","year":2019,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Data access; Computer science; Business; Knowledge management; Database","retraction":null,"screen_n_in":null,"score":{"opus":0.8726138537959646,"gpt":0.6560074045530535,"spread":0.2166064492429111,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","scholarly_communication","open_science"],"consensus_categories":["metaresearch","scholarly_communication","open_science"],"category_scores_codex":[0.06316291,0.000208573,0.0002867829,0.002724296,0.000960773,0.1387359,0.1754812,0.00004054912,0.00004372932],"category_scores_gemma":[0.01252298,0.0001796003,0.00001346147,0.009334749,0.0001523621,0.4469834,0.2792544,0.0009194107,0.000276133],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005420236,"about_ca_system_score_gemma":0.004801103,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001941203,"about_ca_topic_score_gemma":0.0009227645,"domain_scores_codex":[0.9897661,0.0003772906,0.0008711441,0.002533969,0.004717786,0.001733664],"domain_scores_gemma":[0.9877329,0.0002715608,0.0002763317,0.009509586,0.001340059,0.0008695852],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002971794,0.0005134697,0.05038571,0.0001336461,0.00008816613,0.0006479616,0.0009614271,0.03064923,0.01172123,0.1298038,0.5911071,0.1836911],"study_design_scores_gemma":[0.0005050562,0.00007597431,0.006413233,0.0003457797,0.000002260421,0.00008665122,0.000357404,0.8766887,0.000212501,0.001082299,0.1138588,0.00037135],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.108871,0.00005487614,0.7881599,0.09499923,0.0008955941,0.002015272,0.0002501319,0.0000492368,0.004704692],"genre_scores_gemma":[0.1148152,0.000235011,0.8783602,0.004991337,0.0002345964,0.00003306465,0.0001669939,0.00002584154,0.001137862],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8460395,"threshold_uncertainty_score":0.995795,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science","scholarly_communication"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2765696161","doi":"10.5334/dsj-2017-048","title":"The Northern Voice: Listening to Indigenous and Northern Perspectives on Management of Data in Canada","year":2017,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Indigenous Studies and Ecology","field":"Health Professions","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Waterloo","funders":"","keywords":"Indigenous; Interoperability; Active listening; Class (philosophy); The Internet; Geography; World Wide Web; Computer science; Sociology; Ecology","retraction":null,"screen_n_in":null,"score":{"opus":0.08398072553422269,"gpt":0.4148932033926937,"spread":0.330912477858471,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.003954632,0.00007643382,0.0001436001,0.00006057784,0.01627838,0.00008110194,0.003850025,0.00001945129,0.000009181359],"category_scores_gemma":[0.00033608,0.00004859612,0.000005470411,0.00011473,0.0002358414,0.0004972207,0.004155916,0.0004007744,0.00001116165],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007870976,"about_ca_system_score_gemma":0.002968389,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.7843449,"about_ca_topic_score_gemma":0.9988107,"domain_scores_codex":[0.9981743,0.00006118305,0.0002983507,0.0003278545,0.0003892437,0.0007491359],"domain_scores_gemma":[0.9975564,0.0001248022,0.0003525574,0.001666476,0.0001816332,0.0001181016],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00004296388,0.00002667585,0.8940334,0.00002305425,0.00003112597,0.000071886,0.08198106,0.00002288261,0.00000657582,0.00007346623,0.0005402929,0.02314656],"study_design_scores_gemma":[0.0002010431,0.00005237883,0.6729949,0.00007919507,0.000007682805,0.000008783264,0.3171856,0.0001214617,3.293725e-7,0.00001929294,0.009267185,0.00006218204],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9937873,0.0002397991,0.00001429096,0.002490656,0.0005455817,0.0003409187,0.0002018254,0.00000176601,0.0023778],"genre_scores_gemma":[0.9985251,0.0007912348,0.0002978756,0.000156183,0.0001111473,0.000004419514,0.000006653747,0.000005663748,0.000101761],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2352045,"threshold_uncertainty_score":0.9850023,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2083696640","doi":"10.2481/dsj.009-026","title":"Data Management Activities of Canada's National Science Library - 2010 Update and Prospective","year":2010,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"National Research Council Canada","funders":"","keywords":"Mandate; Political science; Transformational leadership; Parliament; Research council; Library science; Engineering management; Public relations; Management; Business; Engineering; Computer science; Government (linguistics)","retraction":null,"screen_n_in":null,"score":{"opus":0.0732256112792575,"gpt":0.3498536390072697,"spread":0.2766280277280122,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.009693839,0.0001244676,0.0001253187,0.0006788873,0.001184352,0.007575611,0.02879043,0.00001659627,0.00003912962],"category_scores_gemma":[0.001194698,0.0001035004,0.000007178218,0.002103661,0.002074969,0.2241235,0.02845682,0.0005545259,0.000002362873],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007663316,"about_ca_system_score_gemma":0.006350324,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001653714,"about_ca_topic_score_gemma":0.008087299,"domain_scores_codex":[0.9943461,0.00004184804,0.0002961527,0.001084342,0.003643888,0.0005877037],"domain_scores_gemma":[0.9955748,0.0001241691,0.0003415164,0.00335637,0.0002397172,0.0003634019],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002282267,0.0001601859,0.006098174,0.00004432392,0.00005330461,0.000129926,0.00009035276,0.00001399875,0.007926772,0.8784237,0.07163564,0.03540082],"study_design_scores_gemma":[0.001209832,0.000158118,0.3781372,0.0001096049,0.00004029651,0.001075996,0.001133705,0.1477136,0.01245421,0.02147282,0.4355026,0.0009920506],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.3382461,0.001084413,0.3075303,0.1239088,0.02077748,0.004043595,0.01645581,0.0004694433,0.187484],"genre_scores_gemma":[0.6512221,0.001498996,0.3455606,0.0005804572,0.0002538911,0.000005621283,0.0001377426,0.0000132404,0.0007272992],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8569509,"threshold_uncertainty_score":0.9992828,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2610869427","doi":"10.5334/dsj-2017-024","title":"All or Nothing: The False Promise of Anonymity","year":2017,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Juvenile Diabetes Research Foundation","funders":"","keywords":"Anonymity; Data sharing; Computer science; Identification (biology); Data anonymization; Process (computing); Internet privacy; State (computer science); Computer security; Information privacy; Medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.8656067754571136,"gpt":0.6914985035282208,"spread":0.1741082719288928,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","open_science"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.03151648,0.00005958479,0.0001650103,0.00006552825,0.001122744,0.0004520937,0.005912262,0.00007475684,0.0002378541],"category_scores_gemma":[0.1409741,0.00002751473,0.00003663931,0.0001436345,0.003676838,0.001149817,0.002111848,0.001965259,0.00002729828],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004366994,"about_ca_system_score_gemma":0.003104599,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007090837,"about_ca_topic_score_gemma":0.00009745258,"domain_scores_codex":[0.9968384,0.00004413291,0.0003605729,0.0002852578,0.002152088,0.0003195629],"domain_scores_gemma":[0.9927489,0.00196587,0.0004336303,0.003665978,0.0007987228,0.0003869356],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.005482406,0.00307168,0.4390132,0.0008889219,0.0003766928,0.002861463,0.009198436,0.00001133415,0.2227885,0.02739555,0.1074329,0.1814789],"study_design_scores_gemma":[0.005370066,0.00158044,0.7693512,0.001694214,0.0002052598,0.002718841,0.001458064,0.008160527,0.01875034,0.03115438,0.1591769,0.000379748],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8206691,0.0001147274,0.0005255033,0.1651663,0.001081701,0.0005327821,0.0001218052,0.00001480764,0.0117733],"genre_scores_gemma":[0.9883703,0.0007402597,0.007724615,0.001457489,0.0004103649,0.000001091105,0.000003229516,0.000006837964,0.001285793],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.330338,"threshold_uncertainty_score":0.9994662,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2132929844","doi":"10.2481/dsj.008-011","title":"Structured Query Translation in Peer to Peer Database Sharing Systems","year":2009,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Scope (computer science); Usability; Transparency (behavior); Data sharing; World Wide Web; Metadata; Implementation; Open data; Reuse; Data science; Database; Software engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.07014835883326626,"gpt":0.3407390784558296,"spread":0.2705907196225634,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.004816004,0.0001603143,0.0002074944,0.0005169348,0.0004300717,0.0009863587,0.003708255,0.00003412766,0.000008998238],"category_scores_gemma":[0.0005497672,0.000131899,0.00002147206,0.00155949,0.00007680603,0.01389431,0.0006362256,0.000345153,0.00002007154],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001225949,"about_ca_system_score_gemma":0.0003060758,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001047656,"about_ca_topic_score_gemma":0.0001066931,"domain_scores_codex":[0.9963745,0.00004287991,0.000507667,0.0007776316,0.001750441,0.0005468266],"domain_scores_gemma":[0.9973039,0.00003497904,0.0001625392,0.001796202,0.000362012,0.0003403475],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001252531,0.0003127203,0.006884302,0.00009912108,0.00002503199,0.001129402,0.01231096,0.03223836,0.1690877,0.4044748,0.04056547,0.332747],"study_design_scores_gemma":[0.001611095,0.0002852707,0.03220744,0.0007929841,0.00001389369,0.002956096,0.001893995,0.560372,0.002085119,0.002121904,0.3942818,0.001378425],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01965325,0.0002326729,0.9741516,0.003436356,0.0015027,0.0002364641,0.0003267931,0.00005692693,0.0004032186],"genre_scores_gemma":[0.6321974,0.00002999386,0.3662995,0.0005383096,0.0005072461,0.000004003308,0.0002119414,0.00001026644,0.0002013688],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6125442,"threshold_uncertainty_score":0.9998978,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4382141301","doi":"10.5334/dsj-2023-018","title":"Polar Data Forum IV – An Ocean of Opportunities","year":2023,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University; Inuvialuit Regional Corporation; Nunavut Research Institute; University of Calgary","funders":"LifeWatch – Niclas Öberg Foundation; Belgian Federal Science Policy Office","keywords":"Interoperability; Polar; Data management; Computer science; Library science; World Wide Web; Data science; Database","retraction":null,"screen_n_in":null,"score":{"opus":0.5412294111931125,"gpt":0.4607770822650198,"spread":0.08045232892809268,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.01804052,0.0001113326,0.0001457413,0.0009064993,0.0007440143,0.00653685,0.05107421,0.00002100627,0.00003188833],"category_scores_gemma":[0.002168272,0.00009257899,0.00001543656,0.002428796,0.0006128668,0.2749197,0.02690796,0.0003238205,0.00004998841],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003045955,"about_ca_system_score_gemma":0.001033398,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009234218,"about_ca_topic_score_gemma":0.00003134763,"domain_scores_codex":[0.9956493,0.0001625812,0.0003922172,0.0008428774,0.002225379,0.0007276489],"domain_scores_gemma":[0.9908743,0.0001528801,0.000351234,0.007964422,0.0002291937,0.0004279974],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002371671,0.0004156588,0.01217221,0.0000623186,0.00007394336,0.0009971892,0.0005527071,0.00009248458,0.003244279,0.2372533,0.5734704,0.1716418],"study_design_scores_gemma":[0.0003182186,0.0002762746,0.01168761,0.00005305255,0.00001586987,0.0003168738,0.004246083,0.4034061,0.0002224879,0.002594304,0.5765502,0.0003128719],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1347816,0.0007322101,0.7074738,0.1175034,0.006402428,0.001127292,0.0112522,0.0008592298,0.01986786],"genre_scores_gemma":[0.7515243,0.01730854,0.2196027,0.002106031,0.0009529655,0.000002815833,0.005868339,0.00005537063,0.002578837],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6167427,"threshold_uncertainty_score":0.9944944,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"not_applicable","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4401463404","doi":"10.5334/dsj-2024-042","title":"Decentralised Semantics: A Semantic Engine User Perspective","year":2024,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Canada First Research Excellence Fund","keywords":"Computer science; Perspective (graphical); Semantics (computer science); Information retrieval; Semantic computing; World Wide Web; Semantic Web; Programming language; Artificial intelligence","retraction":null,"screen_n_in":null,"score":{"opus":0.04095387098060591,"gpt":0.3302790571581251,"spread":0.2893251861775192,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":[],"category_scores_codex":[0.002198078,0.0001684487,0.0001826763,0.0004687434,0.0004819443,0.003496691,0.006216246,0.00004078367,0.00005308841],"category_scores_gemma":[0.0005429689,0.0001230149,0.0000592048,0.001701935,0.000404392,0.007117949,0.001383622,0.0004090825,0.0001804293],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001717681,"about_ca_system_score_gemma":0.0009449071,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004103635,"about_ca_topic_score_gemma":0.00001863771,"domain_scores_codex":[0.997281,0.00004127291,0.0002847162,0.0007868195,0.0009294889,0.000676701],"domain_scores_gemma":[0.9979569,0.0001273054,0.00006870571,0.001341311,0.0002451685,0.0002606128],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001721267,0.0002724361,0.0031862,0.0001035343,0.0001947909,0.004238712,0.01036661,0.0002968273,0.01349351,0.7974194,0.09557097,0.07483985],"study_design_scores_gemma":[0.0005402567,0.0001686429,0.01083949,0.0004047213,0.00007521008,0.006940324,0.002752134,0.8822439,0.00386915,0.0259994,0.06540945,0.0007572723],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03415705,0.004386669,0.9296539,0.02267856,0.006394634,0.00019549,0.00003436736,0.0005507027,0.001948636],"genre_scores_gemma":[0.9200616,0.000579875,0.07830273,0.000416782,0.0004568434,0.000001440897,0.000004440924,0.00001186639,0.0001643617],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8859046,"threshold_uncertainty_score":0.9991606,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4380538161","doi":"10.5334/dsj-2023-015","title":"Legal Regulation of State Electronic Services: Relevant Issues and Ways of Improvement","year":2023,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Corruption and Economic Development","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Government (linguistics); Business; Politics; Public relations; Public sector; Quality (philosophy); State (computer science); Order (exchange); Service (business); Public administration; The Republic; Political science; Marketing; Computer science; Law; Finance","retraction":null,"screen_n_in":null,"score":{"opus":0.05167142070269107,"gpt":0.3291735194967818,"spread":0.2775020987940907,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003956559,0.00003871186,0.00008365617,0.0001522931,0.0003331833,0.0001137984,0.0005497718,0.00001479626,0.00008793757],"category_scores_gemma":[0.00005645088,0.00003425407,0.00001019925,0.000431873,0.0002825914,0.001272763,0.0002055121,0.00007813505,0.00001010442],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000988856,"about_ca_system_score_gemma":0.0006314468,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007557455,"about_ca_topic_score_gemma":0.0007673367,"domain_scores_codex":[0.9989079,0.0000173624,0.0002624774,0.0001542613,0.0003955268,0.0002624416],"domain_scores_gemma":[0.9994328,0.00001900965,0.0002048375,0.0001734121,0.00009912356,0.00007078946],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007934333,0.000109066,0.01775118,0.0001322055,0.00005770689,0.000009594992,0.05337938,0.0003022764,0.1644015,0.06989039,0.00730984,0.6865776],"study_design_scores_gemma":[0.0008146807,0.0003012698,0.1552546,0.0001586733,0.00002233898,0.00002291075,0.02568352,0.0184495,0.006073407,0.01516942,0.7776952,0.0003544269],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9932594,0.00004156477,0.0001199151,0.005257801,0.0002292367,0.00008061459,0.0000223573,0.00001337492,0.0009757307],"genre_scores_gemma":[0.9974339,0.001476603,0.0004691178,0.0001721828,0.00005838235,5.66561e-7,0.00001202417,0.000002218717,0.0003749949],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7703854,"threshold_uncertainty_score":0.256261,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1446508819","doi":"","title":"The Canadian Enhanced Polar Outflow Probe (e-POP) Mission: Current Status and Planned Observations and Data Distribution:Current Status and Planned Observations and Data Distribution","year":2009,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Methane Hydrates and Related Phenomena","field":"Environmental Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Current (fluid); Outflow; Distribution (mathematics); Polar; Environmental science; Meteorology; Geography; Physics; Engineering; Electrical engineering; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.08234495579074003,"gpt":0.3189444292911663,"spread":0.2365994735004263,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.002909425,0.0002401307,0.0001969257,0.00005520909,0.004041933,0.001315776,0.001678363,0.00006906412,0.0000346295],"category_scores_gemma":[0.001005492,0.0001658977,0.000009498141,0.000658163,0.0009925384,0.003989012,0.001883913,0.0005933863,0.000002818113],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003547755,"about_ca_system_score_gemma":0.0008029844,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003746766,"about_ca_topic_score_gemma":0.02214091,"domain_scores_codex":[0.9966586,0.0001111248,0.0004671864,0.0009808751,0.0009040389,0.0008781733],"domain_scores_gemma":[0.9968095,0.0001109728,0.0002668582,0.001491536,0.0000842508,0.001236829],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.0003172159,0.000603352,0.3788021,0.00009330009,0.0001041427,0.00003521326,0.003033282,0.000337788,0.005341636,0.003980754,0.1405325,0.4668187],"study_design_scores_gemma":[0.0005975133,0.0001440354,0.6063023,0.0001084774,0.0000679036,0.00007140024,0.0003738041,0.03341579,0.00002689107,0.001354277,0.3572185,0.0003190837],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8589219,0.01680348,0.003621425,0.01453638,0.001103641,0.0009552955,0.1039559,0.00003865036,0.00006330491],"genre_scores_gemma":[0.9356751,0.02516319,0.001417019,0.0001415714,0.0001550183,0.000003496366,0.03740413,0.00001151166,0.00002899449],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4664996,"threshold_uncertainty_score":0.9997209,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}