{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":19,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":19,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"28c5f4f246c8","filters":{"venue":"International Journal on Digital Libraries"}},"results":[{"id":"W2044606334","doi":"10.1007/s00799-009-0051-7","title":"Long-term digital preservation: preserving authenticity and usability of 3-D data","year":2009,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":23,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"National Research Council Canada; University of Ottawa","funders":"","keywords":"Computer science; Usability; Term (time); Digital preservation; World Wide Web; Human–computer interaction","authors":[{"name":"Julie Doyle","is_ca":false},{"name":"Herna L. Viktor","is_ca":true},{"name":"Eric Paquet","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0681427584605457,"gpt":0.2631286340561562,"spread":0.1949858755956105,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00006974611,0.0001376011,0.0001455897,0.000120618,0.0001228285,0.004349909,0.0009141907,0.00001173284,0.0002332492],"category_scores_gemma":[0.0001283908,0.0001096035,0.00007000812,0.00002498378,0.0003024268,0.0152419,0.0004212782,0.0001260676,0.00001532355],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000137355,"about_ca_system_score_gemma":0.00003472149,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001494302,"about_ca_topic_score_gemma":0.000001997066,"domain_scores_codex":[0.9987005,0.00001088381,0.0003900878,0.000219046,0.0005446033,0.0001348661],"domain_scores_gemma":[0.9992108,0.000158652,0.0001891495,0.0002259474,0.0001250587,0.00009038101],"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":[0.0002312156,0.0006285863,0.008898759,0.00002169453,0.0002016597,0.00004055408,0.001144125,0.00001058703,0.000001614476,0.8390701,0.002222051,0.1475291],"study_design_scores_gemma":[0.0005634293,0.0004425953,0.3047505,0.0002212304,0.00001886701,0.00004826502,0.000188032,0.0002493554,0.00002863692,0.6259758,0.06725892,0.0002543226],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6091859,0.00006158034,0.0003476264,0.004440482,0.0004697434,0.0001257543,0.001664279,0.00005219123,0.3836524],"genre_scores_gemma":[0.9948646,0.00001380221,0.00005966458,0.0002931234,0.0005804526,0.000001012787,0.0004525284,0.00000767681,0.003727103],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3856787,"threshold_uncertainty_score":0.9985314,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1967853860","doi":"10.1007/s007990100046","title":"Managing database server performance to meet QoS requirements in electronic commerce systems","year":2002,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Mobile Agent-Based Network Management","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"IBM (Canada); Queen's University","funders":"","keywords":"Computer science; Database; Workload; Quality of service; Database server; Server; Transaction processing; Quality (philosophy); Computer network; Operating system; Database transaction","authors":[{"name":"Patrick Martin","is_ca":true},{"name":"Wendy Powley","is_ca":true},{"name":"Hoi-Ying Li","is_ca":true},{"name":"Keri Romanufa","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02833109344278201,"gpt":0.2390427269405456,"spread":0.2107116334977636,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0002550491,0.0001951114,0.0001596541,0.0004676761,0.00009579748,0.002987294,0.002065443,0.00002619868,0.00006663998],"category_scores_gemma":[0.00003101836,0.000183558,0.00005982602,0.0003693473,0.00002463812,0.006219925,0.0006970622,0.0002199157,0.0002279447],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000397997,"about_ca_system_score_gemma":0.00004033739,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005446301,"about_ca_topic_score_gemma":0.000004272449,"domain_scores_codex":[0.9977443,0.00005263658,0.0004682102,0.0003360971,0.0009401232,0.0004586153],"domain_scores_gemma":[0.999131,0.0000784149,0.0001691487,0.0003925334,0.00007410665,0.000154791],"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":[0.0001944587,0.0006368815,0.01841239,0.00005269285,0.0003562509,0.001088171,0.0006255588,0.1001462,0.00002313842,0.6334485,0.1342209,0.1107949],"study_design_scores_gemma":[0.002371974,0.0008681235,0.004386296,0.001370422,0.00001145943,0.00031541,0.0001393847,0.4934264,0.0003361009,0.005300547,0.4904734,0.001000547],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6306814,0.002045545,0.1900208,0.03372815,0.01315878,0.001650482,0.0000871642,0.0006061223,0.1280216],"genre_scores_gemma":[0.9948543,0.0001716461,0.0004877948,0.002631413,0.0003665996,0.00002489709,0.00001252764,0.00001805911,0.001432768],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6281479,"threshold_uncertainty_score":0.9980477,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1979784788","doi":"10.1007/s00799-011-0066-8","title":"Interaction and the epistemic potential of digital libraries","year":2010,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Computer science; Leverage (statistics); Digital library; Human–computer interaction; Task (project management); Chunking (psychology); World Wide Web; Data science; Artificial intelligence","authors":[{"name":"Karl V. Fast","is_ca":false},{"name":"Kamran Sedig","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01085154341543618,"gpt":0.2548423658934141,"spread":0.243990822477978,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.000102701,0.000112432,0.0001324659,0.0001658802,0.0000996727,0.007191111,0.001026988,0.00004050407,0.00003764864],"category_scores_gemma":[0.0003088804,0.00007317701,0.00009684473,0.0001166066,0.0003592378,0.009233959,0.0003935152,0.0002815003,0.00002028811],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000007371395,"about_ca_system_score_gemma":0.0000835369,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":8.658567e-7,"about_ca_topic_score_gemma":3.160659e-7,"domain_scores_codex":[0.9989052,0.00002001646,0.0003602083,0.000145108,0.0004641084,0.0001053522],"domain_scores_gemma":[0.9990763,0.0002371518,0.0002785522,0.000175601,0.0001540947,0.00007832314],"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":[0.0001272597,0.00006773167,0.0009653723,0.00000264821,0.00007978284,0.00002053133,0.0002258909,0.00002944367,0.00007988394,0.9731293,0.001396538,0.02387559],"study_design_scores_gemma":[0.003923608,0.0002948299,0.003501055,0.000165971,0.00002684749,0.002458642,0.0006673682,0.03686197,0.004597304,0.8052359,0.1417334,0.0005331559],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3609672,0.0003465649,0.4619003,0.07727416,0.01478937,0.0003642464,0.0005797011,0.000380355,0.08339809],"genre_scores_gemma":[0.997897,0.00002434186,0.0006599714,0.0004359286,0.0003023313,9.438169e-7,0.00003237827,0.000007809926,0.0006393093],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6369298,"threshold_uncertainty_score":0.9938395,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2068371547","doi":"10.1007/s007990100040","title":"Exploring agent support at the user interface in e-commerce applications","year":2002,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Auction Theory and Applications","field":"Decision Sciences","cited_by":9,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Human–computer interaction; User interface; World Wide Web; E-commerce; Interface (matter); Operating system","authors":[{"name":"Patrick Desharnais","is_ca":true},{"name":"Jin‐Ray Lu","is_ca":true},{"name":"T. Radhakrishnan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.250586491250903,"gpt":0.3609713437338586,"spread":0.1103848524829556,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.0004582087,0.0001330389,0.000131676,0.0002800122,0.0002995761,0.002081323,0.00150536,0.00003177872,0.004509022],"category_scores_gemma":[0.0004269301,0.00008482901,0.0001366344,0.0004101665,0.0002012122,0.003349398,0.0003272096,0.0003073971,0.004740132],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001086573,"about_ca_system_score_gemma":0.00002186162,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":8.355973e-7,"about_ca_topic_score_gemma":0.000002935204,"domain_scores_codex":[0.9977756,0.00006934017,0.0006590954,0.0002672424,0.001050067,0.0001786708],"domain_scores_gemma":[0.9980849,0.001008081,0.0002770192,0.0003554774,0.0001633646,0.0001111386],"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":[0.0002027594,0.0006641787,0.01324122,0.000001303543,0.0001142849,0.00004893183,0.00310648,0.00398725,0.0001016923,0.4493405,0.1253596,0.4038318],"study_design_scores_gemma":[0.0002007602,0.00003210115,0.003011743,0.00001144636,0.00000208839,0.0001497149,0.000847968,0.0001121632,0.0007959183,0.05557779,0.9391518,0.0001065056],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5285889,0.0003987939,0.05982799,0.09963508,0.003139381,0.000710126,0.0003416249,0.0001768769,0.3071812],"genre_scores_gemma":[0.9644759,0.00007875704,0.00008113559,0.001133014,0.000342419,0.00009152215,0.000007911387,0.00001305738,0.03377625],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8137922,"threshold_uncertainty_score":0.9989546,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2071860408","doi":"10.1007/s007990100045","title":"Performance stress vectors and capacity planning for e-commerce applications","year":2002,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Software System Performance and Reliability","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University; IBM (Canada)","funders":"","keywords":"Computer science; Software deployment; Revenue; Capacity planning; Reliability (semiconductor); E-commerce; World Wide Web; Business; Software engineering; Finance; Operating system","authors":[{"name":"Marin Litoiu","is_ca":true},{"name":"Diwakar Krishnamurthy","is_ca":true},{"name":"Jerome Rolia","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02998629412239586,"gpt":0.2410800638063831,"spread":0.2110937696839872,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0000790636,0.0001095414,0.0001059129,0.0001121717,0.0002149576,0.001328466,0.000656708,0.00004228295,0.000009798368],"category_scores_gemma":[0.00005264962,0.00008680661,0.00005800988,0.00008538726,0.0000771882,0.003200925,0.0001074331,0.0001478107,0.00002032455],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003500345,"about_ca_system_score_gemma":0.00001823175,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":5.190751e-7,"about_ca_topic_score_gemma":6.999328e-8,"domain_scores_codex":[0.999132,0.000009601575,0.0002362479,0.0001810093,0.0002888232,0.0001522621],"domain_scores_gemma":[0.9993067,0.000205319,0.0001211003,0.000151902,0.0001243338,0.00009065854],"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":[0.00009212632,0.0004395992,0.4927911,0.0001018047,0.0002302706,0.00001505464,0.003035298,0.001503519,0.00002248522,0.06763105,0.006079039,0.4280587],"study_design_scores_gemma":[0.003228283,0.001331209,0.2247247,0.0009354751,0.00002281501,0.001297632,0.0003903926,0.1699737,0.003567835,0.02704986,0.5660163,0.001461798],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8494243,0.0003561293,0.1360013,0.003067362,0.00143332,0.0003147778,0.000101897,0.0002104969,0.00909045],"genre_scores_gemma":[0.9964846,0.00003708857,0.002569854,0.0002563079,0.0003373127,0.00002800881,0.00000578656,0.000007040498,0.0002740485],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5599373,"threshold_uncertainty_score":0.9997082,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4385837278","doi":"10.1007/s00799-023-00378-x","title":"Graduate student search strategies within academic digital libraries","year":2023,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Information Retrieval and Search Behavior","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Regina","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Information seeking; Session (web analytics); Digital library; Matching (statistics); Information retrieval; Online search; Search analytics; Variety (cybernetics); Search engine; Process (computing); Information seeking behavior; Recall; Search problem; Information needs; World Wide Web; Psychology; Web search query; Artificial intelligence","authors":[{"name":"Orland Hoeber","is_ca":true},{"name":"Dale Storie","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07505277667741765,"gpt":0.3368057641022226,"spread":0.2617529874248049,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","insufficient_payload"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0003041144,0.0002462724,0.0001970075,0.0006964473,0.0002889837,0.02269675,0.002741743,0.00009507767,0.00004285012],"category_scores_gemma":[0.0001761691,0.0001935412,0.0001664796,0.0006991132,0.0002223033,0.0309267,0.001034285,0.0008960208,0.001238715],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007194121,"about_ca_system_score_gemma":0.0005840628,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001078849,"about_ca_topic_score_gemma":1.439953e-7,"domain_scores_codex":[0.9961711,0.00003786611,0.0006734694,0.0002763914,0.00238191,0.0004593174],"domain_scores_gemma":[0.9986382,0.0002817675,0.0002087495,0.0002335161,0.0003548976,0.0002829146],"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":[0.0001690768,0.0001416529,0.00654132,0.00001037554,0.0001773025,0.0005570668,0.008173196,0.001750571,0.00004926693,0.9284168,0.006561525,0.04745188],"study_design_scores_gemma":[0.004601247,0.002498236,0.1936562,0.0007506364,0.00002365435,0.002159923,0.02657745,0.02342325,0.01562849,0.6463476,0.08152844,0.002804785],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9383658,0.00005812509,0.01281844,0.01130166,0.003434918,0.0002376409,0.0002164845,0.0009335562,0.03263338],"genre_scores_gemma":[0.9951658,0.00004584422,0.0003912707,0.0004416014,0.0004310682,0.000007837431,0.00009357188,0.00001994315,0.003403026],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2820691,"threshold_uncertainty_score":0.999539,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2077507893","doi":"10.1007/s00799-011-0077-5","title":"Artexte metadata conversion to EPrints: adaptation of digital repository software to visual and media arts documentation","year":2010,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Research Data Management Practices","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Metadata; World Wide Web; Software; Documentation; Personalization; Metadata repository; Information retrieval; Multimedia; Database; Operating system","authors":[{"name":"Tomasz Neugebauer","is_ca":true},{"name":"Corina MacDonald","is_ca":false},{"name":"Felicity Tayler","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03974999555386578,"gpt":0.3221794515468179,"spread":0.2824294559929521,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0002158414,0.0001200302,0.0001164532,0.0004404865,0.0000892095,0.01743001,0.001330014,0.00003073352,0.00001711046],"category_scores_gemma":[0.001776622,0.0001076537,0.00004190974,0.0001862066,0.00006700876,0.09976316,0.001041613,0.0002113504,0.0000510802],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000291018,"about_ca_system_score_gemma":0.00008528204,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006221786,"about_ca_topic_score_gemma":0.000003023281,"domain_scores_codex":[0.9980845,0.00002740145,0.0003554664,0.0003093514,0.001070589,0.0001527011],"domain_scores_gemma":[0.9984887,0.0004891513,0.0002362942,0.0002660108,0.000273792,0.0002460573],"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":[0.000988862,0.0006267562,0.02571356,0.00003110788,0.0004945046,0.0004635011,0.002046624,0.0005213433,0.003864179,0.4777212,0.005756828,0.4817716],"study_design_scores_gemma":[0.004318823,0.003262133,0.2361197,0.0006332683,0.00006715242,0.001287003,0.003311766,0.007119055,0.06787184,0.06030171,0.6138734,0.001834184],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.477851,0.00001766765,0.5130031,0.005159753,0.001640184,0.0002062049,0.000108458,0.00006142902,0.001952247],"genre_scores_gemma":[0.9797177,0.00001915263,0.0189108,0.0003483427,0.0001753745,0.000005384622,0.0000623696,0.000009467239,0.000751448],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6081165,"threshold_uncertainty_score":0.98359,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3209614841","doi":"10.1007/s00799-021-00311-0","title":"Improving data quality in large-scale repositories through conflict resolution","year":2021,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Technische Universität Wien; Vienna Science and Technology Fund; Technische Universität Wien Bibliothek; Ontario Research Foundation","keywords":"Metadata; Computer science; Data quality; Correctness; Scalability; Profiling (computer programming); Data mining; Data science; Conflict resolution; Quality (philosophy); Data collection; Database; Information retrieval; World Wide Web","authors":[{"name":"Artur Kulmukhametov","is_ca":false},{"name":"Andreas Rauber","is_ca":false},{"name":"Christoph Becker","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2067003591168642,"gpt":0.4259471446766152,"spread":0.219246785559751,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.002161196,0.0001378097,0.0002443509,0.000174254,0.0001661236,0.007993273,0.002237656,0.00006186411,0.0002141402],"category_scores_gemma":[0.007950504,0.0001119978,0.00009556126,0.0003529169,0.000113494,0.0120503,0.001936166,0.0003088012,0.000120963],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007694055,"about_ca_system_score_gemma":0.0002308036,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003336521,"about_ca_topic_score_gemma":0.00009093128,"domain_scores_codex":[0.9956418,0.000251455,0.001135609,0.0005413893,0.002186815,0.000242884],"domain_scores_gemma":[0.9971104,0.0009951978,0.0004939765,0.0008815677,0.0004321226,0.00008679413],"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":[0.0008726743,0.001160201,0.0394472,0.00001517691,0.0002323454,0.001468677,0.003464002,0.0003116721,0.0001522557,0.7396085,0.1297334,0.08353385],"study_design_scores_gemma":[0.0006784792,0.00004429807,0.01555593,0.00006167451,0.000004462988,0.0001023477,0.00438292,0.000689899,0.0006395709,0.07895861,0.8986949,0.0001869363],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.3167973,0.002794235,0.2636025,0.09106276,0.02494401,0.0005076742,0.01057187,0.0003076143,0.2894121],"genre_scores_gemma":[0.9784673,0.0001369116,0.005389538,0.003684867,0.001434701,0.000004193125,0.001104832,0.00002029587,0.009757346],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7689615,"threshold_uncertainty_score":0.9930365,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3210786751","doi":"10.1007/s00799-021-00314-x","title":"Correspondence as the primary measure of information quality for web archives: a human-centered grounded theory study","year":2021,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"World Wide Web; Computer science; Quality (philosophy); Web development; Grounded theory; The Internet; Field (mathematics); Relevance (law); Web standards; Sociology; Qualitative research; Political science","authors":[{"name":"Brenda Reyes Ayala","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0373775472066888,"gpt":0.29931203503727,"spread":0.2619344878305812,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0005185673,0.0001112405,0.0001701104,0.0001700941,0.0001751845,0.002722311,0.001492221,0.00002274833,0.00002225747],"category_scores_gemma":[0.0008370125,0.00007681866,0.0001786483,0.0001759823,0.00009625501,0.004877408,0.0003569325,0.0001679777,0.00001358032],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002727028,"about_ca_system_score_gemma":0.0003773609,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000002019904,"about_ca_topic_score_gemma":0.000001557226,"domain_scores_codex":[0.99825,0.0001926291,0.000520236,0.0001513097,0.0007651558,0.0001206426],"domain_scores_gemma":[0.9980052,0.0009350393,0.0003989196,0.0003097454,0.0002961908,0.00005490754],"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":[0.0007967317,0.000811028,0.004338584,0.00001584515,0.0007807082,0.00004904318,0.007656739,0.00007806229,0.0005238907,0.8650504,0.001255474,0.1186435],"study_design_scores_gemma":[0.007763891,0.002176464,0.1241335,0.000743958,0.0001233021,0.0006938967,0.02133165,0.003723646,0.004734811,0.76186,0.07172811,0.0009867463],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5638287,0.0002032458,0.3877119,0.005137333,0.001467403,0.0003528507,0.0005107963,0.0001135158,0.04067422],"genre_scores_gemma":[0.997144,0.000005198635,0.001140354,0.0007881331,0.0001025278,0.000007603129,0.00008114711,0.000004568153,0.0007264137],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4333153,"threshold_uncertainty_score":0.998313,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2736532806","doi":"10.1007/s00799-017-0225-7","title":"The context of multiple in-text references and their signification","year":2017,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Topic Modeling","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Computer science; Rhetorical question; Context (archaeology); Section (typography); Noun phrase; Point (geometry); Natural language processing; Information retrieval; Linguistics; Artificial intelligence; Noun; Mathematics; History","authors":[{"name":"Marc Bertin","is_ca":true},{"name":"Iana Atanassova","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03906275880157609,"gpt":0.2561606087268078,"spread":0.2170978499252317,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0001294125,0.00006146564,0.00007145556,0.00005866604,0.0001813128,0.003021538,0.001315605,0.00002120191,0.000001735207],"category_scores_gemma":[0.0003827436,0.00003654212,0.00002930208,0.00001736321,0.0001269055,0.002669324,0.0002206289,0.000106995,0.000003111428],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001234391,"about_ca_system_score_gemma":0.0000480891,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007651266,"about_ca_topic_score_gemma":0.00001006179,"domain_scores_codex":[0.9993831,0.00001576455,0.0002190368,0.000105973,0.000198769,0.00007733772],"domain_scores_gemma":[0.9990605,0.0003456198,0.0002519625,0.0002169268,0.00009327858,0.00003167771],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00003928439,0.0000296101,0.02547096,0.00000115887,0.00002484107,0.000006412543,0.0008499676,0.00002449498,0.00008447456,0.4372105,0.00008622133,0.5361721],"study_design_scores_gemma":[0.001882586,0.0002969183,0.2374182,0.0003694635,0.000002395544,0.0001574237,0.002070772,0.0680549,0.01604399,0.6180184,0.05528239,0.0004024856],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9224297,0.000474426,0.03947238,0.02013343,0.001476514,0.0001062528,0.00002116723,0.00002831714,0.01585779],"genre_scores_gemma":[0.9991921,0.00007889278,0.0004130003,0.00008695315,0.00008319916,0.000001510408,8.189825e-7,0.000002481368,0.0001410631],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5357696,"threshold_uncertainty_score":0.9980134,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2791424851","doi":"10.1007/s00799-018-0232-3","title":"Capisco: low-cost concept-based access to digital libraries","year":2018,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"Andrew W. Mellon Foundation","keywords":"Computer science; Information retrieval; Metadata; Search engine indexing; Digital library; Semantics (computer science); Context (archaeology); Markup language; Semantic search; World Wide Web; Semantic Web; XML; Programming language","authors":[{"name":"Annika Hinze","is_ca":false},{"name":"David Bainbridge","is_ca":false},{"name":"Sally Jo Cunningham","is_ca":false},{"name":"Craig Taube-Schock","is_ca":false},{"name":"Rangi Matamua","is_ca":false},{"name":"J. Stephen Downie","is_ca":false},{"name":"Edie Rasmussen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0277142637395759,"gpt":0.2855934993284919,"spread":0.257879235588916,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00007031053,0.0002844182,0.0002492624,0.0003947266,0.0002279809,0.02793289,0.004044386,0.00008136731,0.0001464828],"category_scores_gemma":[0.0005567392,0.0002281063,0.0001660212,0.0003278953,0.0004094612,0.01793217,0.0009417321,0.0002459578,0.0005240318],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006334503,"about_ca_system_score_gemma":0.0004233965,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003267046,"about_ca_topic_score_gemma":0.000001968104,"domain_scores_codex":[0.9977401,0.00002493427,0.0004711742,0.0004205197,0.000922987,0.0004202884],"domain_scores_gemma":[0.9983484,0.0003935064,0.0002175247,0.0003777495,0.0003561563,0.0003066351],"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":[0.0007428515,0.0006558739,0.02838514,0.00001199244,0.000370929,0.001103464,0.001993361,0.0005944482,0.00005495007,0.5286593,0.1347583,0.3026694],"study_design_scores_gemma":[0.003483522,0.001897764,0.04269564,0.000820164,0.00001673961,0.0008214517,0.0004831504,0.00474112,0.07539937,0.2172417,0.6503556,0.002043746],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2011467,0.0002885683,0.5960921,0.06768377,0.01414791,0.0006053306,0.0004861371,0.001195146,0.1183543],"genre_scores_gemma":[0.9876791,0.000002582862,0.00306316,0.007125071,0.001246543,0.000009177673,0.00003000929,0.00002240326,0.0008220015],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7865323,"threshold_uncertainty_score":0.9958035,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2065329070","doi":"10.1007/s007990100041","title":"An approach for high-yield requirements capture for e-commerce and its application","year":2002,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Service-Oriented Architecture and Web Services","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"IBM (Canada); University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Synchronization (alternating current); Concurrency; The Internet; Quality (philosophy); Software; Software engineering; Software quality; Non-functional requirement; Distributed computing; Software development; World Wide Web; Operating system; Software construction; Computer network","authors":[{"name":"Kassem Saleh","is_ca":false},{"name":"Robert L. Probert","is_ca":true},{"name":"W. Li","is_ca":true},{"name":"W. Fong","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02781493700183751,"gpt":0.251771793651979,"spread":0.2239568566501415,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.00007196295,0.0001457254,0.0001187203,0.0001402041,0.0001628281,0.001587412,0.001073548,0.0000594701,0.00000673086],"category_scores_gemma":[0.00002325207,0.0001197961,0.00005995327,0.0000899872,0.00001957806,0.003517209,0.0001048398,0.0001169113,0.000004665357],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001815979,"about_ca_system_score_gemma":0.0000152005,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000002403804,"about_ca_topic_score_gemma":9.030913e-7,"domain_scores_codex":[0.998956,0.00001085264,0.0002370033,0.0002918634,0.0003242364,0.000180026],"domain_scores_gemma":[0.9992009,0.0001605085,0.0001587419,0.0001713301,0.0001868393,0.0001216324],"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":[0.0001988645,0.0004796313,0.0006827766,0.00003953825,0.0001691268,0.000005184084,0.001798385,0.0005580515,0.0006260505,0.8775874,0.001573333,0.1162817],"study_design_scores_gemma":[0.004928551,0.001814557,0.004050238,0.0002140343,0.00005017652,0.0004549073,0.0007310407,0.5412496,0.01085435,0.2575682,0.1767114,0.001372929],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02768522,0.0003434026,0.9587549,0.008019393,0.0006538652,0.0004488477,0.0001558399,0.0001167412,0.003821771],"genre_scores_gemma":[0.9693673,0.00001639208,0.02620708,0.003582786,0.0004366481,0.00005677062,0.00009086427,0.00001523692,0.0002268989],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9416821,"threshold_uncertainty_score":0.999449,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2766971772","doi":"10.1007/s00799-017-0229-3","title":"Encoding music performance data in Humdrum and MEI","year":2017,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Music and Audio Processing","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Fonds de Recherche du Québec-Société et Culture; National Endowment for the Humanities","keywords":"Computer science; Information retrieval; World Wide Web","authors":[{"name":"Johanna Devaney","is_ca":false},{"name":"Hubert Léveillé Gauvin","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06493170774482294,"gpt":0.2826062163914074,"spread":0.2176745086465844,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0001465616,0.00008599884,0.00008834492,0.00011858,0.0002789194,0.008664988,0.002689307,0.00002389567,0.00001256475],"category_scores_gemma":[0.0001800016,0.00007090842,0.00001673856,0.00003196115,0.00008947776,0.01595761,0.001310972,0.0001838238,0.00001178008],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001677068,"about_ca_system_score_gemma":0.00008007319,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001406859,"about_ca_topic_score_gemma":0.000001430847,"domain_scores_codex":[0.9991289,0.000006857113,0.0001856829,0.0002062488,0.0003352675,0.0001369958],"domain_scores_gemma":[0.9992869,0.00005377598,0.0001932077,0.0003710729,0.00003409028,0.00006097076],"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":[0.0000344057,0.00006099449,0.05556912,0.000009367471,0.00003281418,0.0002443217,0.0006296738,0.00002039923,0.00004149578,0.05582891,0.002406929,0.8851216],"study_design_scores_gemma":[0.003146307,0.0003185012,0.5468321,0.002040584,0.000008384295,0.001894935,0.0002381062,0.07417774,0.003572486,0.09945913,0.2671371,0.001174623],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8575066,0.0004049247,0.01926315,0.01524206,0.002630536,0.00006047816,0.00002945709,0.0000676391,0.1047952],"genre_scores_gemma":[0.996713,0.0000713086,0.001859894,0.000771299,0.0003087627,5.063442e-7,0.000004595833,0.000004814397,0.0002658305],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.883947,"threshold_uncertainty_score":0.9978057,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2036382113","doi":"10.1007/s007990000036","title":"On Z39.50 wrapping and description logics","year":2000,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; University of Toronto","funders":"","keywords":"Computer science; Information retrieval; Metadata; Query language; Database; Semantics (computer science); World Wide Web; Protocol (science); Programming language","authors":[{"name":"Yannis Velegrakis","is_ca":true},{"name":"Vassilis Christophides","is_ca":false},{"name":"Panos Constantopoulos","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02274182857905333,"gpt":0.2290437351356132,"spread":0.2063019065565599,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.00006548978,0.0001111546,0.00009714744,0.0001384254,0.0001063808,0.003603138,0.0006307255,0.00004173699,0.00008550917],"category_scores_gemma":[0.00009385565,0.00008543963,0.00005422173,0.00006563203,0.00007259744,0.003434202,0.00008712005,0.0001658317,0.0001051267],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002626154,"about_ca_system_score_gemma":0.00003549084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001259374,"about_ca_topic_score_gemma":2.893774e-7,"domain_scores_codex":[0.9990795,0.00001763393,0.000196558,0.0001754809,0.0003857749,0.000145056],"domain_scores_gemma":[0.999532,0.0001616932,0.00007129491,0.0001119702,0.00005011852,0.00007291044],"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":[0.0001194368,0.0001021499,0.002000576,0.000001796883,0.00005914604,0.0001901121,0.0003056946,0.000254387,0.00002623071,0.6053128,0.00373651,0.3878911],"study_design_scores_gemma":[0.00091895,0.0006817463,0.03750528,0.0002134838,0.000004740639,0.001193111,0.0001242519,0.004564992,0.0007089101,0.7781886,0.1754876,0.0004082779],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.801142,0.0005475094,0.06144558,0.01232722,0.002120689,0.00008562478,0.00001488795,0.0002761494,0.1220403],"genre_scores_gemma":[0.9942351,0.0001233527,0.002422298,0.001732817,0.0002142972,0.000001130762,0.000002787919,0.000005406026,0.001262786],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3874828,"threshold_uncertainty_score":0.9974312,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2010700167","doi":"10.1007/s00799-006-0003-4","title":"Generating customized yet factually consistent information: a constraint satisfaction approach","year":2006,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Constraint satisfaction problem; Personalization; Consistency (knowledge bases); Set (abstract data type); Constraint (computer-aided design); Constraint satisfaction; Snippet; Information retrieval; World Wide Web; Artificial intelligence; Mathematics","authors":[{"name":"Syed Sibte Raza Abidi","is_ca":true},{"name":"Yong Han Chong","is_ca":false},{"name":"Yan Zeng","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01486148997786254,"gpt":0.2381679105484165,"spread":0.2233064205705539,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0001190189,0.0001467166,0.000140239,0.000258361,0.0001541734,0.007524288,0.0005753325,0.00004755145,0.00005597011],"category_scores_gemma":[0.0001562855,0.0001242291,0.00010421,0.0001431234,0.00008193222,0.008426933,0.0001501519,0.0001699277,0.0000945981],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007495075,"about_ca_system_score_gemma":0.0001934308,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005548641,"about_ca_topic_score_gemma":7.365114e-7,"domain_scores_codex":[0.9984273,0.00003107435,0.0005445656,0.0001429922,0.0006973445,0.0001566705],"domain_scores_gemma":[0.9990579,0.0001008603,0.0003270335,0.0001424147,0.0002815942,0.00009020318],"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":[0.00003539392,0.00009930292,0.001374614,0.000003961214,0.00009216651,0.00002645066,0.000243502,0.006657616,0.0000231849,0.9247591,0.01587608,0.05080866],"study_design_scores_gemma":[0.005564094,0.0002812751,0.006179216,0.0001735228,0.00002472049,0.002238054,0.0008013629,0.5062244,0.0009791225,0.04964786,0.4267364,0.001149988],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007905493,0.00002416143,0.9268884,0.001766451,0.001153416,0.00009148949,0.0001466945,0.0001720463,0.06185183],"genre_scores_gemma":[0.971347,0.000009714867,0.02531234,0.002206109,0.0003105993,0.000003763052,0.0003229131,0.000007895515,0.0004796901],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9634415,"threshold_uncertainty_score":0.993506,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2014016891","doi":"10.1007/s007990100043","title":"A fast method for determining the origins of documents based on LZW compression","year":2002,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Copying; Computer science; Publication; Revenue; Information retrieval; World Wide Web; Database; Data mining; Advertising","authors":[{"name":"Gordon B. Agnew","is_ca":true},{"name":"A. Sivanandan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02826485079003627,"gpt":0.2944132124102884,"spread":0.2661483616202521,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0001175155,0.0001286215,0.0001310918,0.0001461221,0.0001597595,0.001518878,0.00145497,0.00003344387,0.00004319664],"category_scores_gemma":[0.0001238408,0.0000768583,0.0001216357,0.00008824513,0.00004296291,0.002380647,0.000231125,0.0001580674,0.00001276609],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004880636,"about_ca_system_score_gemma":0.00003275527,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001275379,"about_ca_topic_score_gemma":7.640048e-8,"domain_scores_codex":[0.998674,0.00003895342,0.0003014235,0.0001858218,0.0006534206,0.000146428],"domain_scores_gemma":[0.9985855,0.0006827439,0.0002678015,0.0002350163,0.0001570003,0.00007197692],"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":[0.0002379316,0.0006361495,0.001393974,0.000007973117,0.00009509313,0.00003341714,0.0004329003,0.004762416,0.00007640733,0.1623317,0.01216707,0.817825],"study_design_scores_gemma":[0.00146234,0.0005478435,0.001644735,0.0002817679,0.000004970177,0.00003229581,0.00001874148,0.7738176,0.002102931,0.01012544,0.2097846,0.000176681],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003920864,0.00003241906,0.9871644,0.002947726,0.001287188,0.0001287198,0.0001464169,0.00004126539,0.004330977],"genre_scores_gemma":[0.8948642,0.000009056806,0.1029063,0.001509431,0.0003537242,0.0000102928,0.00002687206,0.00001486563,0.000305256],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8909433,"threshold_uncertainty_score":0.9995176,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4416811906","doi":"10.1007/s00799-025-00436-6","title":"Going beyond digital libraries: a literature review of phygital user experience research methods","year":2025,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Innovative Human-Technology Interaction","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Saskatchewan","funders":"","keywords":"User experience design; Context (archaeology); User Research; Inclusion (mineral); User interface; Key (lock); Digital library; Usability","authors":[{"name":"Jessica Waggoner","is_ca":false},{"name":"Shannon Lucky","is_ca":true},{"name":"Stacey Redick","is_ca":false},{"name":"Amanda Rizki","is_ca":false},{"name":"Jen-chien Yu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04147550841286916,"gpt":0.4228192382402177,"spread":0.3813437298273485,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0005515682,0.000275154,0.0003666708,0.001399091,0.0002324492,0.007061881,0.003176037,0.0001527007,0.00006611085],"category_scores_gemma":[0.002867069,0.000232103,0.0002194689,0.001874412,0.0005631464,0.02052352,0.001353365,0.001219518,0.00004882976],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000149398,"about_ca_system_score_gemma":0.0004024842,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":3.087561e-7,"about_ca_topic_score_gemma":7.334325e-8,"domain_scores_codex":[0.9969671,0.0001576965,0.0009016521,0.0004854716,0.001104688,0.0003833852],"domain_scores_gemma":[0.9962312,0.001086151,0.0004246454,0.0005215362,0.001652412,0.00008405586],"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":[0.0001048199,0.0002071123,0.0007560321,0.0001701095,0.0002012168,0.0001922421,0.000891896,0.000002431131,0.0002537428,0.8987205,0.01668456,0.08181535],"study_design_scores_gemma":[0.0006805283,0.0005088807,0.0006626613,0.02184065,0.000007844499,0.000833771,0.0004536927,0.0003587539,0.03473406,0.3706497,0.5687766,0.0004928517],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02964695,0.03638637,0.4607968,0.04058452,0.01031577,0.0009428379,0.0003136405,0.0007396508,0.4202735],"genre_scores_gemma":[0.8860857,0.003968997,0.0813866,0.009058742,0.0009003633,0.0001017625,0.000150077,0.00007615436,0.01827154],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8564388,"threshold_uncertainty_score":0.9939689,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2055732723","doi":"10.1007/s00799-015-0148-0","title":"Introduction to the focussed issue on Semantic Digital Archives","year":2015,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; World Wide Web; Digital library; Information retrieval; Linguistics","authors":[{"name":"Thomas Risse","is_ca":false},{"name":"Livia Predoiu","is_ca":false},{"name":"Annett Mitschick","is_ca":false},{"name":"Andreas Nürnberger","is_ca":false},{"name":"Seamus Ross","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01999165762934203,"gpt":0.2450897246190097,"spread":0.2250980669896677,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0001069468,0.0001635426,0.0001349671,0.0002392352,0.0001116815,0.006459421,0.001755492,0.00003033992,0.00001709394],"category_scores_gemma":[0.0007565203,0.00009987817,0.00009821699,0.0001664556,0.00009117076,0.003484308,0.0003457053,0.0002299583,0.0008114097],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003838493,"about_ca_system_score_gemma":0.0001135732,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001146832,"about_ca_topic_score_gemma":6.180913e-7,"domain_scores_codex":[0.9984514,0.0000307178,0.0002741706,0.0002713265,0.0007528566,0.0002195332],"domain_scores_gemma":[0.9990069,0.0002715667,0.000118833,0.0003134357,0.0001119626,0.0001772835],"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":[0.0004525842,0.000294358,0.001460645,0.000002445606,0.0001787477,0.0002337349,0.002603727,0.00204221,0.00002012648,0.3446411,0.2248985,0.4231718],"study_design_scores_gemma":[0.0004447105,0.0006430647,0.002477732,0.00004814082,0.000003087059,0.0004702329,0.0003861462,0.001259881,0.001301243,0.08895599,0.9037904,0.0002194089],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.06865642,0.0002202718,0.1773266,0.6242267,0.01579211,0.0003205506,0.00005148368,0.0004759954,0.1129298],"genre_scores_gemma":[0.9859107,0.000007701541,0.001406054,0.002718581,0.005626449,0.000004997722,0.000009712615,0.00001289006,0.004302925],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9172543,"threshold_uncertainty_score":0.9999666,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3028150595","doi":"10.1007/s00799-020-00286-4","title":"Introduction to the focused issue on the 2017 ACM/IEEE-CS Joint Conference on Digital Libraries JCDL 2017","year":2020,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Digital library; Joint (building); World Wide Web; Engineering","authors":[{"name":"Catherine C. Marshall","is_ca":false},{"name":"Ian Milligan","is_ca":true},{"name":"Adam Jatowt","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07667170856385166,"gpt":0.2631198388417867,"spread":0.1864481302779351,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0001633444,0.0003278354,0.000267658,0.0001609252,0.0004174747,0.01489484,0.004217892,0.00006955893,0.0001199054],"category_scores_gemma":[0.002237569,0.0001820243,0.0001946504,0.0002318502,0.0002535972,0.005578353,0.000781797,0.0006072893,0.001445989],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000498122,"about_ca_system_score_gemma":0.0002153691,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000002253817,"about_ca_topic_score_gemma":0.000001032116,"domain_scores_codex":[0.9974427,0.00007172373,0.0004806268,0.00050803,0.001147321,0.0003496193],"domain_scores_gemma":[0.9978415,0.000682405,0.0003141124,0.0007185628,0.0002232682,0.0002201861],"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":[0.0003781779,0.0001292595,0.0001758363,0.000003407445,0.0001565762,0.0001609975,0.001746987,0.0004594353,0.00004857379,0.4082461,0.5468412,0.04165341],"study_design_scores_gemma":[0.0006607893,0.0016815,0.003434906,0.0001813833,0.00001114822,0.000330519,0.0007775144,0.00258032,0.01024276,0.1020335,0.877489,0.0005766358],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.02564009,0.0001094727,0.01790234,0.9179133,0.005745176,0.0003116129,0.0001189517,0.0002881179,0.031971],"genre_scores_gemma":[0.9735314,0.00004620523,0.0008620436,0.01695773,0.006615066,0.0000148807,0.00001777912,0.00002395014,0.001930935],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9478914,"threshold_uncertainty_score":0.9993315,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}