{"meta":{"query_hash":"cd9713cc8e05","filters":{"venue":"Archiving Conference"},"cohort_total":13,"direct_labels_cover":0,"predictions_cover":13,"exported":13,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/cd9713cc8e05","api":"https://metacan.xera.ac/api/v1/cohort?venue=Archiving+Conference"},"results":[{"id":"W267152328","doi":"10.2352/issn.2168-3204.2010.7.1.art00003","title":"Theories, Methods and Testbeds for Curation and Preservation of Digital Art","year":2010,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"University of Toronto; University of Glasgow","keywords":"Digital preservation; Computer science; Digital curation; Testbed; Digital art; Key (lock); Data science; Data curation; Digital media; World Wide Web; Art; Computer security","score_opus":0.04394106136096881,"score_gpt":0.280852768064651,"score_spread":0.23691170670368217,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W267152328","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.3961103,0.000008207556,0.07145641,0.0009810143,0.00019386681,0.00035249858,0.00020312409,0.000042771073,0.5306518],"genre_scores_gemma":[0.99141234,0.0000048675224,0.00545111,0.000022922433,0.00006550793,0.000019741892,0.00006234948,0.000004333091,0.0029568141],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9996583,0.000009874389,0.00010593362,0.00010734858,0.000050313873,0.00006823391],"domain_scores_gemma":[0.99943364,0.00038941024,0.000047154066,0.00006362908,0.00004059204,0.00002556218],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00009333779,0.00005728174,0.00007002118,0.000028756545,0.00009032269,0.00022438401,0.000053684806,0.0000044978515,0.000020749088],"category_scores_gemma":[0.00011339685,0.00004717382,0.000014977708,0.000008057464,0.00027161144,0.00069701736,0.000051134644,0.000044151617,0.0000010139636],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000008393193,0.000018821547,0.0003925908,0.000043310247,0.000009528673,4.5583974e-8,0.0022404266,1.5324052e-7,0.0005382223,0.88917494,0.000056275156,0.10751732],"study_design_scores_gemma":[0.00010023779,0.00008559702,0.021197068,0.000021634374,0.000008961439,8.9710176e-7,0.00012061804,0.0016790368,0.00010821181,0.9323963,0.044212997,0.00006845718],"about_ca_topic_score_codex":0.000006801392,"about_ca_topic_score_gemma":0.00004445827,"teacher_disagreement_score":0.59530205,"about_ca_system_score_codex":6.889224e-7,"about_ca_system_score_gemma":0.000010959991,"threshold_uncertainty_score":0.21637401},"labels":[],"label_agreement":null},{"id":"W2971896075","doi":"10.2352/issn.2168-3204.2005.2.1.art00022","title":"A Metadata Schema Registry for the Registration and Analysis of Recordkeeping and Preservation Metadata","year":2005,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada; Monash University; University of California, Los Angeles; National Archives of Australia; National Science Foundation","keywords":"Metadata; Schema (genetic algorithms); Computer science; Metadata repository; Meta Data Services; Database; Information retrieval; World Wide Web","score_opus":0.09698456522857066,"score_gpt":0.2569613748133919,"score_spread":0.15997680958482122,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2971896075","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.47671133,0.00059580087,0.32763478,0.020334562,0.00015154979,0.0012864233,0.00201757,0.00009607495,0.1711719],"genre_scores_gemma":[0.9917409,0.00010896426,0.002783715,0.00011648174,0.00007293896,0.000024888082,0.00024507145,0.000003893338,0.0049031386],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99943584,0.000018291277,0.0001811497,0.0001780994,0.00010110414,0.00008553422],"domain_scores_gemma":[0.9990456,0.0005600326,0.00012361324,0.00020728253,0.000040639665,0.000022843002],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00018974568,0.00007129904,0.00013030165,0.00006539329,0.00018439279,0.00042249856,0.00015366718,0.000004458528,0.000015503903],"category_scores_gemma":[0.00011828031,0.000051883082,0.00004239699,0.000039858598,0.00021314275,0.001231845,0.00009162456,0.000040919735,2.574875e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000014788359,0.0000191694,0.0004191771,0.000042135558,0.00042652266,8.072182e-8,0.0014867363,0.000014941725,0.00004605152,0.9330647,0.00026894672,0.06419678],"study_design_scores_gemma":[0.00052303635,0.00016232798,0.34037447,0.00021812458,0.0026410508,0.000002403943,0.0013800038,0.14618713,0.00010568305,0.12410387,0.3839058,0.00039608747],"about_ca_topic_score_codex":0.00019652826,"about_ca_topic_score_gemma":0.0015572237,"teacher_disagreement_score":0.8089608,"about_ca_system_score_codex":0.000002745096,"about_ca_system_score_gemma":0.000014261462,"threshold_uncertainty_score":0.40741634},"labels":[],"label_agreement":null},{"id":"W4377836928","doi":"10.2352/issn.2168-3204.2006.3.1.art00028","title":"Photo CD PCD to Tiff Migration Case Study","year":2006,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Archaeological Research and Protection","field":"Earth and Planetary Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Shared Services Canada; Canadian Museum of Nature","funders":"","keywords":"Globe; Computer science; Computer graphics (images); Visual arts; Art","score_opus":0.037560535554304234,"score_gpt":0.2682472271803024,"score_spread":0.2306866916259982,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377836928","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9793114,0.000016532937,0.010346656,0.0003578026,0.000061471845,0.0004949779,0.000020025807,0.00006120854,0.0093299365],"genre_scores_gemma":[0.99736446,0.000003739615,0.0017768012,0.0000912997,0.00008456472,0.000010530458,0.000020193242,0.0000020089428,0.0006464239],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9988109,0.00018782214,0.00015136313,0.00027977113,0.00024986677,0.00032026536],"domain_scores_gemma":[0.9993536,0.00025188297,0.000029102996,0.0001648759,0.00004171822,0.00015882326],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00034070254,0.00009961855,0.000099897814,0.000093907984,0.00027932232,0.00007365257,0.00016873852,0.000022576713,0.0013167971],"category_scores_gemma":[0.00016020135,0.00007747103,0.000026157792,0.00018612275,0.00006339835,0.00011830252,0.000037638038,0.00017010779,0.0003045134],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000794731,0.000046887923,0.6659588,0.00000933218,0.000007794835,0.0011061826,0.001605154,0.0008431718,0.0008218207,0.00014553173,0.0004670814,0.3289087],"study_design_scores_gemma":[0.00015886118,0.0008951891,0.97944486,0.000013087472,0.000004094549,0.00022832619,0.00069682876,0.008043881,0.00022514637,0.008353468,0.0017686015,0.00016765967],"about_ca_topic_score_codex":0.100601114,"about_ca_topic_score_gemma":0.22964579,"teacher_disagreement_score":0.32874107,"about_ca_system_score_codex":0.0000035333644,"about_ca_system_score_gemma":0.000045987028,"threshold_uncertainty_score":0.9995961},"labels":[],"label_agreement":null},{"id":"W4377836996","doi":"10.2352/issn.2168-3204.2006.3.1.art00055","title":"A Report from the Open Content Alliance Project","year":2006,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Alliance; Digitization; Context (archaeology); Content (measure theory); Political science; Public relations; World Wide Web; Knowledge management; Computer science; Geography; Telecommunications; Mathematics","score_opus":0.22493981145317163,"score_gpt":0.2843865348193642,"score_spread":0.05944672336619258,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377836996","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.33451778,0.00007631722,0.00007254633,0.0008064893,0.00030689896,0.0003829516,0.000115843715,0.00006363039,0.66365755],"genre_scores_gemma":[0.92152274,0.0000065321005,0.00013321228,0.0005403,0.0005082669,0.000057958154,0.00010017551,0.000014439393,0.07711638],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99891126,0.000058989368,0.00028473625,0.00029228462,0.00019678542,0.0002559649],"domain_scores_gemma":[0.9990471,0.00025832918,0.00013267172,0.0004327895,0.000101888705,0.00002723518],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0002313506,0.00014646685,0.00017111121,0.00001437374,0.000492428,0.0030101845,0.00093579653,0.000014220324,0.0008986664],"category_scores_gemma":[0.00008204598,0.00009621324,0.00005869058,0.000016840042,0.0003102808,0.0006008438,0.00034655054,0.00018301976,0.000066930406],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000017071243,0.00005377524,0.0029349052,0.0000049204587,0.000022553235,0.00006720554,0.0059027495,6.363749e-7,0.000040258394,0.9693268,0.017154887,0.004474267],"study_design_scores_gemma":[0.00028589842,0.000056962243,0.046569314,0.00020190715,0.000018939487,0.0000146976145,0.0032282402,0.00005047808,0.00004900786,0.14565279,0.80357224,0.00029951107],"about_ca_topic_score_codex":0.03947101,"about_ca_topic_score_gemma":0.03676636,"teacher_disagreement_score":0.82367396,"about_ca_system_score_codex":0.000014722907,"about_ca_system_score_gemma":0.0001073145,"threshold_uncertainty_score":0.99802476},"labels":[],"label_agreement":null},{"id":"W4377837011","doi":"10.2352/issn.2168-3204.2006.3.1.art00035","title":"Technical Issues in Digitization of Large Online Collections of Phonograph Records","year":2006,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Music and Audio Processing","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Centre for Interdisciplinary Research in Music Media and Technology","funders":"Centre for Interdisciplinary Research in Music Media and Technology","keywords":"Digitization; Metadata; Computer science; Phonograph; World Wide Web; Multimedia; Information retrieval; Track (disk drive); Computer graphics (images); Computer vision; Engineering; Operating system","score_opus":0.014713757817075524,"score_gpt":0.2707972862311153,"score_spread":0.2560835284140398,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837011","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.15256797,0.00004790823,0.83647066,0.00033807653,0.000057075544,0.00006853789,0.000009988366,0.000054020213,0.0103857955],"genre_scores_gemma":[0.94972,0.000025397552,0.049917124,0.000029532204,0.000018157863,0.0000034031884,0.0000064999635,0.000003039903,0.00027678913],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.9992575,0.000034412285,0.0002789352,0.00016727486,0.00012177356,0.00014007914],"domain_scores_gemma":[0.99949974,0.000094047755,0.0001234344,0.00019251625,0.00007308517,0.000017167482],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00013135505,0.00006553107,0.00014996312,0.00022181586,0.000053582124,0.000035705132,0.00031520432,0.000024318168,0.0000094571915],"category_scores_gemma":[0.000066964014,0.000062638464,0.000031480213,0.00082762516,0.000066808825,0.00019080473,0.000118382384,0.00008205749,4.401846e-7],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000027823999,0.0025467698,0.20246875,0.00042687796,0.000020032025,0.000011932121,0.004684161,0.00082536676,0.13798472,0.4513571,0.0052287905,0.19441767],"study_design_scores_gemma":[0.0012451159,0.00030173687,0.5941798,0.0017822956,0.0000147702585,0.000021269083,0.00019812156,0.16715369,0.0585854,0.17095536,0.0049187075,0.00064370665],"about_ca_topic_score_codex":0.00023484863,"about_ca_topic_score_gemma":0.0010347385,"teacher_disagreement_score":0.7971521,"about_ca_system_score_codex":0.0000075433136,"about_ca_system_score_gemma":0.00010041509,"threshold_uncertainty_score":0.25543222},"labels":[],"label_agreement":null},{"id":"W4377837238","doi":"10.2352/issn.2168-3204.2015.12.1.art00026","title":"Using SharePoint workflows and InfoPath forms to manage a large-scale digitization project: A case study of the Kissinger Papers Project","year":2015,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Library and Archives Canada","funders":"","keywords":"Digitization; Workflow; Vendor; Computer science; Tracking (education); Engineering management; Service (business); Process (computing); Project management; Scale (ratio); World Wide Web; Software engineering; Library science; Engineering; Systems engineering; Database; Business; Telecommunications; Sociology; Geography","score_opus":0.11297496918773835,"score_gpt":0.28043145758639465,"score_spread":0.1674564883986563,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837238","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8511064,0.0000023238645,0.000363826,0.00005504143,0.00008632405,0.00083075726,0.000049602684,0.000031405598,0.1474743],"genre_scores_gemma":[0.9984478,7.538036e-7,0.00028994944,0.000108275744,0.00005405957,0.000048427202,0.000007950142,0.000011639013,0.0010311647],"study_design_codex":"qualitative","study_design_gemma":"qualitative","domain_scores_codex":[0.999099,0.000043363776,0.00022015997,0.00022484439,0.00022340786,0.00018920981],"domain_scores_gemma":[0.99954027,0.00004133237,0.00009306505,0.00019236111,0.00007512652,0.0000578408],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012475417,0.00012961449,0.00013759256,0.000121207784,0.00024918007,0.00034516832,0.00014655862,0.000005334122,0.00000741116],"category_scores_gemma":[0.000051455354,0.00008505441,0.00003447662,0.00008077121,0.0001014392,0.0003156285,0.00034770946,0.00007216516,0.0000016118589],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000079751,0.0011090171,0.006036262,0.00022272511,0.0001174965,0.00023563628,0.80415875,0.0002066062,0.00003430875,0.13788778,0.00024704612,0.049664624],"study_design_scores_gemma":[0.005153864,0.0038969214,0.079160795,0.0032941285,0.00048015488,0.0008029923,0.67252666,0.053059373,0.000052013176,0.07892452,0.10031908,0.0023295241],"about_ca_topic_score_codex":0.0008199902,"about_ca_topic_score_gemma":0.002893801,"teacher_disagreement_score":0.14734137,"about_ca_system_score_codex":0.000015715294,"about_ca_system_score_gemma":0.00006191409,"threshold_uncertainty_score":0.3468418},"labels":[],"label_agreement":null},{"id":"W4377837264","doi":"10.2352/issn.2168-3204.2011.8.1.art00002","title":"Long-term Access to Archival Materials That are Not in a Traditional Archives","year":2011,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital Imaging in Medicine","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Dialog box; Intellectual property; Library science; Discipline; Atlas (anatomy); Special collections; Term (time); Sociology; Visual arts; World Wide Web; Computer science; Political science; Art; Law; Social science","score_opus":0.22320168860555195,"score_gpt":0.3317088660068019,"score_spread":0.10850717740124993,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837264","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9409167,0.0000062691192,0.003714629,0.0015189276,0.00025473276,0.0005003924,0.000053570115,0.000100761135,0.052934002],"genre_scores_gemma":[0.9940758,0.000016031954,0.0032899543,0.0018011488,0.00016390535,0.00007767198,0.00006180292,0.000035498033,0.00047822983],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9982093,0.00008214983,0.00035874028,0.00048608545,0.00037654425,0.00048716206],"domain_scores_gemma":[0.998822,0.0002825876,0.00010810954,0.0004345038,0.000028644105,0.00032417834],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00018385409,0.00024904803,0.00046307791,0.0004275095,0.000045087658,0.00011508452,0.00052308134,0.000024142242,0.0003950501],"category_scores_gemma":[0.00033108317,0.00022126442,0.00006113796,0.00014082406,0.0002971028,0.00031476314,0.0002717288,0.0002765111,0.00007834861],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008408585,0.0006453521,0.8851751,0.00039664662,0.00006237516,0.0009797015,0.0061892234,0.0000011686327,0.0607439,0.004699958,0.0002614021,0.04000431],"study_design_scores_gemma":[0.000762143,0.00018352343,0.96521765,0.0022298493,0.00002438851,0.0001751359,0.000050983926,0.00003600757,0.026671167,0.0044040815,0.000024182295,0.00022090891],"about_ca_topic_score_codex":0.00006560465,"about_ca_topic_score_gemma":0.00004579824,"teacher_disagreement_score":0.08004252,"about_ca_system_score_codex":0.000024251824,"about_ca_system_score_gemma":0.00014774043,"threshold_uncertainty_score":0.9022901},"labels":[],"label_agreement":null},{"id":"W4377837265","doi":"10.2352/issn.2168-3204.2012.9.1.art00014","title":"Building On OAIS: Processing Born-Digital Records Using Archivematica","year":2012,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Dashboard; World Wide Web; Open source; Computer science; Architecture; Web application; Database; Digital Archives; Geography; Operating system; Archaeology","score_opus":0.07774912262464444,"score_gpt":0.2572676066714764,"score_spread":0.17951848404683196,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837265","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.26070702,0.000013561091,0.010516552,0.00017899436,0.00028923104,0.00015916066,0.000059612306,0.00013652137,0.72793937],"genre_scores_gemma":[0.9940424,0.0000026182292,0.0034747836,0.00022626783,0.00062799355,0.00001534318,0.000023199116,0.000025884356,0.0015615259],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99870116,0.000029899798,0.00025506975,0.00024353094,0.00027374423,0.0004965914],"domain_scores_gemma":[0.9993338,0.00019721586,0.000108160704,0.00017508688,0.000031064905,0.0001546601],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001069688,0.0002103614,0.000179136,0.00016189543,0.00042270334,0.0007847722,0.00026114885,0.0000074098025,0.0001491661],"category_scores_gemma":[0.000053278705,0.00018030322,0.00008386684,0.000035483372,0.00027918935,0.0011609903,0.00014866679,0.00016901713,0.00015597347],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000013954379,0.00017853148,0.00026569646,0.00009218981,0.000027687625,0.0000031268723,0.0061436063,0.000015759872,0.000058556892,0.84140396,0.00014830953,0.15164863],"study_design_scores_gemma":[0.0006216247,0.00044883328,0.033151984,0.0030979807,0.00011132131,0.000043113047,0.002266778,0.021356395,0.00019628277,0.7716176,0.1654312,0.001656901],"about_ca_topic_score_codex":0.000024387586,"about_ca_topic_score_gemma":0.000006314975,"teacher_disagreement_score":0.7333354,"about_ca_system_score_codex":0.000022158392,"about_ca_system_score_gemma":0.00004203579,"threshold_uncertainty_score":0.7567576},"labels":[],"label_agreement":null},{"id":"W4377837435","doi":"10.2352/issn.2168-3204.2015.12.1.art00009","title":"Wayfinding in Complex Architectural Collections: A Collaborative Approach between Creator and Curator","year":2015,"lang":"en","type":"article","venue":"Archiving Conference","topic":"3D Surveying and Cultural Heritage","field":"Earth and Planetary Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Archivist; Digitization; Globe; Metadata; Workflow; Architecture; Special collections; World Wide Web; Library science; Work (physics); Digital collections; Art history; Sociology; Computer science; Visual arts; Art; Engineering; Database; Telecommunications","score_opus":0.1030543309634732,"score_gpt":0.260231459346091,"score_spread":0.15717712838261777,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837435","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.96863776,0.000121761885,0.00043259983,0.00012776851,0.00004614536,0.00020133327,0.00010810783,0.00005232435,0.030272199],"genre_scores_gemma":[0.9947881,0.0000094421675,0.0046815677,0.000032188767,0.00005136941,0.000004471415,0.000115065384,0.0000031612183,0.00031463723],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99879426,0.00027521743,0.00018558952,0.00028327532,0.00017141509,0.00029026082],"domain_scores_gemma":[0.9992929,0.00029014662,0.00005290703,0.00010379296,0.000051110805,0.00020912797],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00035340313,0.00014335991,0.00022260229,0.000113333655,0.0002290261,0.00018062656,0.00015491698,0.00003495682,0.0000672508],"category_scores_gemma":[0.00019187233,0.000114049784,0.000018262112,0.00050980767,0.00015083255,0.00016881221,0.000024994728,0.00020287298,0.00001710012],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000026258313,0.000008924639,0.97810155,0.000014671554,0.000011372926,0.000005117233,0.005401792,0.0001372185,0.00006712105,0.0001438282,0.00019889897,0.015883226],"study_design_scores_gemma":[0.0003863796,0.0001248515,0.9850017,0.000040820323,0.0000053056706,0.00001970147,0.0026447924,0.009047266,0.000032295873,0.001964228,0.00051911426,0.00021357446],"about_ca_topic_score_codex":0.0020055026,"about_ca_topic_score_gemma":0.0032407762,"teacher_disagreement_score":0.02995756,"about_ca_system_score_codex":0.000008487623,"about_ca_system_score_gemma":0.00013322872,"threshold_uncertainty_score":0.4650815},"labels":[],"label_agreement":null},{"id":"W4377837477","doi":"10.2352/issn.2168-3204.2012.9.1.art00012","title":"A New Beginning: Collaborative Partnerships for Digitizing and Indexing Military Records of the Directorate of History and Heritage Archives","year":2012,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Image Processing and 3D Reconstruction","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Department of National Defence","funders":"","keywords":"Search engine indexing; History; Archival science; Library science; Visual arts; World Wide Web; Computer science; Archaeology; Art","score_opus":0.04187789935894849,"score_gpt":0.23958949197118062,"score_spread":0.19771159261223215,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837477","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.67121935,0.004217363,0.32113394,0.0003333312,0.00044687546,0.00021700229,0.000006219316,0.000044026707,0.0023818745],"genre_scores_gemma":[0.91920936,0.000043434047,0.0804755,0.000017074262,0.000032981367,0.000005890769,4.1001698e-7,0.0000048312713,0.00021051161],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9993003,0.00009433969,0.00017545643,0.00017289947,0.000080616715,0.00017636197],"domain_scores_gemma":[0.9991712,0.00039473144,0.000149967,0.00016768901,0.000046737536,0.00006966598],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00020304736,0.00009390878,0.00016076227,0.00006433231,0.000117565025,0.000029567118,0.0002070525,0.000019951738,7.6529454e-7],"category_scores_gemma":[0.00016687461,0.00007491122,0.000026136518,0.00009827882,0.0003663098,0.00044512283,0.00015955571,0.0000942089,4.808715e-8],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000044957924,0.000036957234,0.11900373,0.0005033414,0.000047000853,4.2940138e-7,0.08231869,0.000004221177,0.04215261,0.03524489,0.00013499097,0.72050816],"study_design_scores_gemma":[0.0031067843,0.00078434945,0.6065519,0.0061204806,0.00016326705,0.0002092698,0.0076447083,0.1060051,0.14839888,0.104599796,0.014551949,0.0018635568],"about_ca_topic_score_codex":0.000039838236,"about_ca_topic_score_gemma":0.000025560399,"teacher_disagreement_score":0.7186446,"about_ca_system_score_codex":0.000010442668,"about_ca_system_score_gemma":0.00034258634,"threshold_uncertainty_score":0.3054791},"labels":[],"label_agreement":null},{"id":"W4377837496","doi":"10.2352/issn.2168-3204.2012.9.1.art00052","title":"Zooming In on the Big Picture: Digitization of Nitrate Panoramic Film","year":2012,"lang":"en","type":"article","venue":"Archiving Conference","topic":"3D Surveying and Cultural Heritage","field":"Earth and Planetary Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Digitization; Nitrate; Computer graphics (images); Negative; Computer science; Engineering; Art; Visual arts; Chemistry; Telecommunications","score_opus":0.049504445373273004,"score_gpt":0.2118561824976519,"score_spread":0.1623517371243789,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377837496","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.97902274,0.000048231097,0.0002115005,0.00008402109,0.00012948293,0.000064294625,0.000016507809,0.000014842718,0.020408385],"genre_scores_gemma":[0.9995934,0.000016883303,0.00007721084,0.00010058374,0.0000392042,6.876041e-7,0.00002777383,0.0000016192599,0.00014265132],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99933046,0.00013803657,0.00012369563,0.00009000037,0.00011138524,0.00020642385],"domain_scores_gemma":[0.9994096,0.000378801,0.00004839226,0.000111636116,0.000012953448,0.000038608698],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00028702873,0.00007260852,0.000082743936,0.000030708678,0.00007108926,0.000033197193,0.00014835993,0.000019579084,0.0002109306],"category_scores_gemma":[0.000113251146,0.00004340192,0.000020599564,0.00012350224,0.00005362541,0.00010277676,0.00000784167,0.0001211152,0.000047525176],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000079578585,0.000010085953,0.95444083,0.00000949058,0.0000028640004,5.9667855e-7,0.0016648683,0.00023366707,0.00071267283,0.0009545918,0.000014715674,0.04194768],"study_design_scores_gemma":[0.000046542886,0.000025584362,0.9925405,0.00007229542,0.0000016457977,0.0000012838501,0.00029469663,0.005732273,0.00037754216,0.00074693596,0.000090953195,0.00006977717],"about_ca_topic_score_codex":0.00046086972,"about_ca_topic_score_gemma":0.0013542371,"teacher_disagreement_score":0.041877903,"about_ca_system_score_codex":0.0000014864204,"about_ca_system_score_gemma":0.0000226105,"threshold_uncertainty_score":0.23095421},"labels":[],"label_agreement":null},{"id":"W4377927260","doi":"10.2352/issn.2168-3204.2005.2.1.art00037","title":"Joint Compression and Restoration of Documents with Bleed-through","year":2005,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Handwritten Text Recognition Techniques","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"University of Ottawa","keywords":"Computer science; Bleed; Compression (physics); Joint (building); Segmentation; Inpainting; Artificial intelligence; Computer vision; Image (mathematics); Engineering; Medicine","score_opus":0.02762871088630275,"score_gpt":0.26340661916365643,"score_spread":0.23577790827735368,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377927260","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.16816789,0.000025006117,0.82459176,0.0007147567,0.000017270924,0.00013051857,0.0000012913287,0.00013280423,0.0062186974],"genre_scores_gemma":[0.7804348,0.000044659715,0.21935262,0.00007582755,0.000012324787,0.000009751488,0.000001786264,0.0000036001566,0.000064622625],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9992291,0.00006391975,0.00018524603,0.00019883875,0.00019274709,0.00013015508],"domain_scores_gemma":[0.99941856,0.000058438305,0.00012087351,0.0002732751,0.000084821906,0.000044059878],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012927022,0.00009032362,0.00012831241,0.00006857439,0.00006996821,0.00008294941,0.00025251476,0.000021715296,0.000014170283],"category_scores_gemma":[0.000020625976,0.0000737153,0.000013936336,0.00009886266,0.000087522414,0.0007616013,0.00018507245,0.00009172766,0.0000068299823],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000026096326,0.00016685024,0.0038806503,0.00009771501,0.000030378495,0.0000060210095,0.008864161,0.00002725292,0.11594183,0.1495287,0.00069020555,0.72074014],"study_design_scores_gemma":[0.001471281,0.0010013122,0.07664423,0.0020881684,0.000026618189,0.00010403348,0.0001497026,0.061808288,0.76981056,0.080546945,0.0055440017,0.00080486515],"about_ca_topic_score_codex":0.000056644265,"about_ca_topic_score_gemma":0.00002957695,"teacher_disagreement_score":0.7199353,"about_ca_system_score_codex":0.000012133488,"about_ca_system_score_gemma":0.000048790687,"threshold_uncertainty_score":0.3006023},"labels":[],"label_agreement":null},{"id":"W4377927295","doi":"10.2352/issn.2168-3204.2004.1.1.art00047","title":"The Digital Signature Dilemma: To Preserve or Not to Preserve","year":2004,"lang":"en","type":"article","venue":"Archiving Conference","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Library and Archives Canada","funders":"","keywords":"Digital signature; Dilemma; Electronic signature; Internet privacy; Computer security; Odds; Identity (music); Cryptography; Value (mathematics); Political science; Process (computing); Computer science; Law; Hash function; Mathematics","score_opus":0.047329641670127744,"score_gpt":0.2407983750669612,"score_spread":0.19346873339683346,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377927295","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.25452316,0.0000067283477,0.0013558991,0.016739888,0.00043764553,0.0008637261,0.00053821923,0.00018121002,0.72535354],"genre_scores_gemma":[0.9663604,0.0000033871515,0.00017241585,0.00100165,0.00030311767,0.0000832398,0.00003540838,0.000018028315,0.032022335],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9987023,0.000019538806,0.0002102134,0.0003279558,0.00034262863,0.00039737724],"domain_scores_gemma":[0.99901253,0.00034910816,0.000042196592,0.00033022562,0.00006643696,0.00019951517],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.00007580227,0.000187937,0.00013440195,0.00007101574,0.0005284645,0.0017146729,0.00076544064,0.000008464539,0.00013014257],"category_scores_gemma":[0.00019071507,0.00011376417,0.000079603684,0.000053703523,0.00017929052,0.0004663863,0.0005553104,0.00016420189,0.00044398688],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00014989803,0.00009160264,0.000014037126,0.00002247203,0.00004650641,0.000020213722,0.008728198,0.0002929753,0.000022230843,0.95895785,0.0038891868,0.027764821],"study_design_scores_gemma":[0.00031975782,0.00059767574,0.023298115,0.00026883776,0.000014138157,0.00000387957,0.00069361896,0.000072672694,0.00010242743,0.21765035,0.756557,0.0004215018],"about_ca_topic_score_codex":0.00014616991,"about_ca_topic_score_gemma":0.0012644671,"teacher_disagreement_score":0.75266784,"about_ca_system_score_codex":0.00002006328,"about_ca_system_score_gemma":0.00007042434,"threshold_uncertainty_score":0.99932164},"labels":[],"label_agreement":null}]}