{"meta":{"query_hash":"2d95964af09c","filters":{"venue":"ACM Transactions on Knowledge Discovery from Data"},"cohort_total":47,"direct_labels_cover":0,"predictions_cover":47,"exported":47,"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/2d95964af09c","api":"https://metacan.xera.ac/api/v1/cohort?venue=ACM+Transactions+on+Knowledge+Discovery+from+Data"},"results":[{"id":"W1965792576","doi":"10.1145/1376815.1376816","title":"Joint cluster analysis of attribute data and relationship data","year":2008,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":50,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Data mining; Cluster analysis; Heuristic; Identification (biology); Tree (set theory); Machine learning; Mathematics; Artificial intelligence","score_opus":0.15403939983569437,"score_gpt":0.3374659296680593,"score_spread":0.1834265298323649,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1965792576","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04467297,0.00026163962,0.95105094,0.00038066527,0.000033552238,0.00015095017,0.0012420475,0.00070171605,0.0015055533],"genre_scores_gemma":[0.46446684,0.0003810547,0.5261062,0.0001314788,0.00010149668,0.00044021974,0.005582307,0.00028610983,0.0025043085],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9936633,0.0019425382,0.00032714655,0.0015502751,0.002011249,0.0005054963],"domain_scores_gemma":[0.98409903,0.008251806,0.001208199,0.0033807019,0.0025297601,0.00053050474],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051629604,0.0011624935,0.0020375242,0.0065833568,0.0018868624,0.0035180917,0.0032144655,0.0017267262,0.002199593],"category_scores_gemma":[0.022721602,0.0007094794,0.00208849,0.009076138,0.0014486263,0.004764669,0.0031392584,0.0023711787,0.0008813201],"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":[0.0010475245,0.0006635603,0.028849786,0.00059288373,0.0006617541,0.00052886334,0.0010258689,0.5278539,0.006805596,0.13175432,0.012436289,0.28777972],"study_design_scores_gemma":[0.00001945373,0.00004276196,0.0022217855,0.000018219558,0.000054486856,0.00013177203,0.0002464179,0.90747595,0.002907759,0.08348496,0.003364379,0.000032117303],"about_ca_topic_score_codex":0.0051208762,"about_ca_topic_score_gemma":0.0046851523,"teacher_disagreement_score":0.0065833568,"about_ca_system_score_codex":0.002006607,"about_ca_system_score_gemma":0.0020682653,"threshold_uncertainty_score":0.02730465},"labels":[],"label_agreement":null},{"id":"W1973522621","doi":"10.1145/2757281","title":"Refining Social Graph Connectivity via Shortcut Edge Addition","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":52,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Graph; Computation; Theoretical computer science; Algorithm; Data mining","score_opus":0.08509666336746628,"score_gpt":0.3278070009381637,"score_spread":0.24271033757069738,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1973522621","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06978099,0.00058011233,0.924515,0.0003733182,0.00008555193,0.0002985526,0.000394555,0.0014105311,0.0025614034],"genre_scores_gemma":[0.38236913,0.00046413852,0.6111705,0.0003105763,0.00013450615,0.00048138652,0.0014173695,0.0003489769,0.0033034408],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99836856,0.000517352,0.00006941529,0.0004163206,0.000511595,0.00011677252],"domain_scores_gemma":[0.98884445,0.0078043663,0.0009782662,0.0010481281,0.0009603046,0.00036449797],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0018869656,0.0015981703,0.0017701649,0.003588905,0.0012478824,0.001308111,0.0021140561,0.0021584735,0.0041772663],"category_scores_gemma":[0.01707276,0.0008754465,0.0015728509,0.0019894028,0.0013086848,0.0036042004,0.0025458834,0.0019771117,0.0006758735],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00029569134,0.00028844553,0.005943553,0.00060898234,0.00019019195,0.00066079386,0.0004896233,0.6297741,0.010993466,0.05502411,0.008154308,0.2875768],"study_design_scores_gemma":[0.000036045396,0.00008316633,0.00057621155,0.000035822817,0.00003829811,0.00016869498,0.00007473069,0.95759034,0.0025750513,0.036192596,0.002609476,0.00001958964],"about_ca_topic_score_codex":0.003753771,"about_ca_topic_score_gemma":0.0074543823,"teacher_disagreement_score":0.0041772663,"about_ca_system_score_codex":0.0009465799,"about_ca_system_score_gemma":0.0012165682,"threshold_uncertainty_score":0.013974369},"labels":[],"label_agreement":null},{"id":"W1993150471","doi":"10.1145/1870096.1870099","title":"Discovering Knowledge-Sharing Communities in Question-Answering Forums","year":2010,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Expert finding and Q&A systems","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec en Outaouais","funders":"","keywords":"Computer science; Cluster analysis; Transaction data; Question answering; Set (abstract data type); Information retrieval; Identification (biology); Focus (optics); Function (biology); Database transaction; Data mining; Artificial intelligence; Database","score_opus":0.055056583051641024,"score_gpt":0.32007919195545265,"score_spread":0.26502260890381163,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1993150471","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6938417,0.001190601,0.2983999,0.00070573646,0.000047679892,0.0006892813,0.0016545851,0.0008700353,0.002600487],"genre_scores_gemma":[0.8668677,0.00015738915,0.12892768,0.000074561154,0.000057606336,0.00028403007,0.0027246051,0.000065268934,0.00084120344],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98910856,0.005772627,0.00050998543,0.0023549765,0.001666512,0.0005872613],"domain_scores_gemma":[0.9492251,0.03596724,0.0047737225,0.004502148,0.0038780174,0.0016538311],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010915463,0.0009805014,0.0013009228,0.008974935,0.0024870133,0.003028107,0.0023627114,0.0025846236,0.001291093],"category_scores_gemma":[0.044031918,0.0007049059,0.0014108038,0.0050365604,0.0018204646,0.00865994,0.0037650934,0.0013414109,0.0006060864],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0030485492,0.0026479687,0.34295022,0.002155127,0.0007812362,0.0042126225,0.04684595,0.09399186,0.043871664,0.055795204,0.008198698,0.39550087],"study_design_scores_gemma":[0.00017804396,0.0004802692,0.072688885,0.00024170765,0.00021760172,0.0016641634,0.014884112,0.76548314,0.016366474,0.11161247,0.015985461,0.00019771676],"about_ca_topic_score_codex":0.0031678486,"about_ca_topic_score_gemma":0.0038750633,"teacher_disagreement_score":0.010915463,"about_ca_system_score_codex":0.0011903731,"about_ca_system_score_gemma":0.000998556,"threshold_uncertainty_score":0.057727158},"labels":[],"label_agreement":null},{"id":"W1993625478","doi":"10.1145/1409620.1409623","title":"On disclosure risk analysis of anonymized itemsets in the presence of prior knowledge","year":2008,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Hacker; Computer science; Heuristics; Context (archaeology); Set (abstract data type); Function (biology); Data mining; Benchmark (surveying); Information retrieval; Computer security","score_opus":0.05564791742840004,"score_gpt":0.3095288978769892,"score_spread":0.25388098044858914,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1993625478","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.15452039,0.0039426847,0.83450013,0.0030757126,0.000092318056,0.00026541532,0.00083336316,0.00039002474,0.0023800423],"genre_scores_gemma":[0.9049052,0.00279791,0.08811141,0.00043586816,0.00038445665,0.00034575138,0.0012018934,0.000106785155,0.0017106364],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98601484,0.0072542443,0.0010329813,0.0019904352,0.0028801945,0.00082724524],"domain_scores_gemma":[0.6065529,0.35905257,0.016926114,0.0103193605,0.0055236584,0.001625378],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.027872676,0.0019626906,0.003217729,0.0048852065,0.0015326871,0.0042895135,0.0030653484,0.0036769905,0.0021070142],"category_scores_gemma":[0.1670895,0.0011520276,0.002363153,0.004193527,0.004377269,0.009897526,0.0037536426,0.004901734,0.00031301874],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0004142344,0.00013647182,0.01936088,0.00042563566,0.00028217948,0.00084217644,0.00073934445,0.8536233,0.0008570471,0.08845264,0.0017061443,0.03316],"study_design_scores_gemma":[0.000012964087,0.000076210155,0.0015111928,0.00010602034,0.000048241393,0.0003529197,0.00013854931,0.92681855,0.0005380995,0.069902,0.00045493306,0.000040315765],"about_ca_topic_score_codex":0.003009452,"about_ca_topic_score_gemma":0.0015324573,"teacher_disagreement_score":0.027872676,"about_ca_system_score_codex":0.003128707,"about_ca_system_score_gemma":0.0019156365,"threshold_uncertainty_score":0.14740658},"labels":[],"label_agreement":null},{"id":"W2020791430","doi":"10.1145/1552303.1552306","title":"Link spam target detection using page farms","year":2009,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"PageRank; Computer science; Backlink; Web page; Spamdexing; HITS algorithm; Information retrieval; Ranking (information retrieval); Link analysis; Search engine; Precision and recall; Data mining; World Wide Web; Web search engine; Static web page; Web search query; Web development","score_opus":0.05779916987046218,"score_gpt":0.2945533312560526,"score_spread":0.2367541613855904,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2020791430","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.49807152,0.0021152354,0.4785665,0.0004431481,0.00015693146,0.00059395336,0.0016887859,0.011164868,0.0071991086],"genre_scores_gemma":[0.8544123,0.0005341041,0.13868503,0.00009952424,0.00021979489,0.00018895893,0.0023298839,0.00014784586,0.003382475],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99862695,0.00033502243,0.00009824903,0.00026806805,0.0005510918,0.00012072545],"domain_scores_gemma":[0.99107915,0.0035628062,0.0016767274,0.0013022624,0.0020939135,0.00028511634],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014169174,0.000686666,0.0009854912,0.009510489,0.00092214724,0.0015192702,0.00066865963,0.001224445,0.0012614357],"category_scores_gemma":[0.008350234,0.0003563109,0.0005219337,0.002613494,0.00074581464,0.003632901,0.0009666558,0.0005194026,0.0012832792],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007850026,0.0006806268,0.22185905,0.0005149396,0.00030102258,0.00096947054,0.0006764358,0.03251593,0.047906257,0.008431553,0.021713916,0.6636458],"study_design_scores_gemma":[0.00005904761,0.00038861902,0.06485593,0.000072570256,0.00014778716,0.0023290906,0.00039121456,0.8528716,0.05240298,0.016158335,0.010205301,0.000117434894],"about_ca_topic_score_codex":0.0010498395,"about_ca_topic_score_gemma":0.001656089,"teacher_disagreement_score":0.009510489,"about_ca_system_score_codex":0.00041963233,"about_ca_system_score_gemma":0.00054361834,"threshold_uncertainty_score":0.007493496},"labels":[],"label_agreement":null},{"id":"W2038873345","doi":"10.1145/1857947.1857950","title":"Centralized and Distributed Anonymization for High-Dimensional Healthcare Data","year":2010,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":133,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Tech University; Concordia University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Data anonymization; Scalability; Data sharing; Information privacy; Information sharing; Health care; Data mining; Data science; Computer security; Database; World Wide Web; Medicine","score_opus":0.06175333628821846,"score_gpt":0.3219685442326207,"score_spread":0.2602152079444023,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2038873345","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.028266521,0.00034747666,0.96820337,0.0011946021,0.000048575188,0.00010069813,0.0001739526,0.00024636753,0.0014183727],"genre_scores_gemma":[0.78286093,0.0004716853,0.21373864,0.00030040357,0.00015321765,0.00022692677,0.00062696775,0.000064083186,0.0015571134],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98440486,0.009379582,0.0007042715,0.0020574231,0.0028611277,0.00059278496],"domain_scores_gemma":[0.9695189,0.011657187,0.0025026926,0.014546048,0.0014639477,0.000311212],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011675835,0.00045389237,0.0010311459,0.0010157736,0.0022845415,0.0029009585,0.0017694077,0.0012885162,0.0005819566],"category_scores_gemma":[0.021736894,0.00047027314,0.0012058251,0.0023887793,0.0030473135,0.0051161507,0.0037655425,0.0017926353,0.0002378479],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005314783,0.00018895682,0.009086363,0.00025286854,0.0002716662,0.0007134747,0.0024119248,0.4419955,0.008131644,0.351915,0.006798383,0.1777028],"study_design_scores_gemma":[0.0000518512,0.00008195516,0.0019083879,0.000036667087,0.00005817091,0.000579536,0.00068801857,0.73609734,0.0071250815,0.24237056,0.01095743,0.000045020668],"about_ca_topic_score_codex":0.0015511675,"about_ca_topic_score_gemma":0.0016657822,"teacher_disagreement_score":0.011675835,"about_ca_system_score_codex":0.002227637,"about_ca_system_score_gemma":0.0030291856,"threshold_uncertainty_score":0.061748445},"labels":[],"label_agreement":null},{"id":"W2045929671","doi":"10.1145/1376815.1376819","title":"Semantic text similarity using corpus-based word similarity and string similarity","year":2008,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Topic Modeling","field":"Computer Science","cited_by":485,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Similarity (geometry); Semantic similarity; String metric; Artificial intelligence; Word (group theory); Natural language processing; Longest common subsequence problem; Focus (optics); String searching algorithm; String (physics); Representation (politics); Information retrieval; Matching (statistics); Pattern matching; Mathematics; Algorithm","score_opus":0.12043376955524356,"score_gpt":0.3074205502158795,"score_spread":0.18698678066063595,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2045929671","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.051913235,0.0021264022,0.930313,0.00026806977,0.0004651728,0.0006221626,0.0023097303,0.0025874637,0.009394729],"genre_scores_gemma":[0.2841628,0.0011061836,0.7038279,0.00012819398,0.0005406998,0.0010794784,0.00547775,0.00041786322,0.0032591443],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99439275,0.0011248859,0.00070041994,0.0010071052,0.0026378462,0.0001368597],"domain_scores_gemma":[0.99211943,0.0030306864,0.0009637382,0.0011390698,0.0024947596,0.00025224863],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021779973,0.00092376,0.0014646518,0.017746313,0.0009860167,0.002976211,0.0016307629,0.0012851419,0.003514882],"category_scores_gemma":[0.018149171,0.00031206195,0.0011634456,0.016821468,0.0012409195,0.005136708,0.0020246366,0.0010016097,0.0019029668],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005017755,0.00040544325,0.012898957,0.001379963,0.00056546653,0.00041738807,0.0011314675,0.02129064,0.028761242,0.053923283,0.011165284,0.86755913],"study_design_scores_gemma":[0.00019402873,0.00092868815,0.027077055,0.00043689326,0.00055513874,0.0028144156,0.0016984984,0.67202175,0.048917342,0.18239895,0.062490948,0.0004662709],"about_ca_topic_score_codex":0.0016587181,"about_ca_topic_score_gemma":0.0016378546,"teacher_disagreement_score":0.017746313,"about_ca_system_score_codex":0.000980401,"about_ca_system_score_gemma":0.0013563224,"threshold_uncertainty_score":0.011758447},"labels":[],"label_agreement":null},{"id":"W2055028639","doi":"10.1145/1460797.1460801","title":"On domination game analysis for microeconomic data mining","year":2009,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":31,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"National University of Singapore","keywords":"Nash equilibrium; Game theory; Dimension (graph theory); Product (mathematics); Commodity; Computer science; Point (geometry); Best response; Space (punctuation); Mathematical economics; Economics; Mathematics","score_opus":0.07051834428210217,"score_gpt":0.3199586946775136,"score_spread":0.2494403503954114,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2055028639","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0058465484,0.0007105552,0.98895746,0.00054526544,0.000058306538,0.00017899729,0.00014326611,0.00009317749,0.0034664092],"genre_scores_gemma":[0.380694,0.0029500714,0.60568845,0.0008449686,0.00044677078,0.0017647544,0.00063428783,0.00022739364,0.006749227],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9949528,0.003103144,0.00023418397,0.0005499937,0.0007952742,0.00036464413],"domain_scores_gemma":[0.97059757,0.026434833,0.00088560814,0.00078049104,0.00089348963,0.00040801562],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008198864,0.002995659,0.0033913318,0.0041731503,0.001543905,0.0033031336,0.0029941038,0.0019516514,0.0052226395],"category_scores_gemma":[0.028091377,0.0012116439,0.0030329393,0.0055365195,0.004365378,0.0049340995,0.0036478324,0.0036906993,0.00071825867],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00006142603,0.000112388596,0.0013417582,0.00021851127,0.00017205601,0.00016083641,0.00018746579,0.6798005,0.00052576134,0.29495156,0.0017992316,0.020668563],"study_design_scores_gemma":[0.000018712011,0.000015105315,0.00008055411,0.000016855778,0.000012827823,0.000020609514,0.000016735774,0.8109585,0.00009537335,0.1876908,0.0010612158,0.000012799057],"about_ca_topic_score_codex":0.010837673,"about_ca_topic_score_gemma":0.0067493008,"teacher_disagreement_score":0.010837673,"about_ca_system_score_codex":0.0043159705,"about_ca_system_score_gemma":0.0033180725,"threshold_uncertainty_score":0.043360233},"labels":[],"label_agreement":null},{"id":"W2069857456","doi":"10.1145/1993077.1993080","title":"Can the Utility of Anonymized Data be Used for Privacy Breaches?","year":2011,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":84,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Office of Integrative Activities; Office of International Science and Engineering; Division of Information and Intelligent Systems; Research Grants Council, University Grants Committee","keywords":"Data publishing; Computer science; Data anonymization; Adversary; Anonymity; Information privacy; k-anonymity; Closeness; Internet privacy; Data mining; Computer security; Data science; Publishing; Mathematics","score_opus":0.21436211065159538,"score_gpt":0.3377835758204947,"score_spread":0.12342146516889929,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2069857456","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.122061335,0.009972642,0.6868676,0.12378419,0.0016676799,0.00055275235,0.0021991301,0.0015499619,0.05134471],"genre_scores_gemma":[0.91305816,0.0027387692,0.07262773,0.0049071684,0.00077356026,0.0002442447,0.00059283775,0.00029406126,0.004763343],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.93723214,0.03898875,0.0026754593,0.0051759854,0.013912735,0.002014912],"domain_scores_gemma":[0.79453385,0.07921044,0.010812132,0.10475795,0.00952527,0.0011603746],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.040782165,0.0010098014,0.0020392402,0.0035419383,0.0028916139,0.010129154,0.0040052417,0.0070991674,0.0029150043],"category_scores_gemma":[0.15962079,0.0010092927,0.0019426773,0.006119642,0.012623148,0.031296626,0.0072709923,0.006013757,0.0019101767],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009675998,0.00019605484,0.016463814,0.00065518456,0.0006397338,0.0009428842,0.0042281426,0.029731562,0.003289681,0.66960984,0.017289056,0.25598654],"study_design_scores_gemma":[0.00007523049,0.00021688035,0.0022090655,0.00048556033,0.00014833192,0.0019990513,0.0024920579,0.040454317,0.010282769,0.8801966,0.061335158,0.000105023515],"about_ca_topic_score_codex":0.0008672254,"about_ca_topic_score_gemma":0.00048694393,"teacher_disagreement_score":0.040782165,"about_ca_system_score_codex":0.0030987894,"about_ca_system_score_gemma":0.0026419493,"threshold_uncertainty_score":0.21567929},"labels":[],"label_agreement":null},{"id":"W2096281132","doi":"10.1145/1460797.1460799","title":"Mining frequent cross-graph quasi-cliques","year":2009,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":91,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Division of Information and Intelligent Systems","keywords":"Heuristics; Computer science; Clique; Data mining; Vertex (graph theory); Graph; Theoretical computer science; Combinatorics; Mathematics","score_opus":0.05441437221485232,"score_gpt":0.33723344906684105,"score_spread":0.2828190768519887,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2096281132","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.44646245,0.0010824013,0.540214,0.002123375,0.000069744725,0.00041356977,0.0055886367,0.0008243404,0.00322147],"genre_scores_gemma":[0.76084507,0.00050477544,0.22623183,0.0003565323,0.00015229362,0.00029880836,0.009530523,0.00015131265,0.0019289217],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99693906,0.00079344714,0.00016901024,0.0012256051,0.0005564271,0.00031642008],"domain_scores_gemma":[0.98242927,0.01115739,0.002299686,0.0019144521,0.0015028333,0.0006963699],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0022631083,0.00079741504,0.0016308151,0.002742835,0.001765262,0.0019662038,0.0025695893,0.0015008379,0.0022961632],"category_scores_gemma":[0.013560224,0.0011167311,0.0018599418,0.003231357,0.001189523,0.005258562,0.0021433746,0.0014420863,0.00039546977],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0013328213,0.0012165959,0.10233593,0.0019564366,0.001693472,0.0054636216,0.0021719935,0.47891745,0.020222723,0.18495709,0.028088203,0.1716437],"study_design_scores_gemma":[0.00007655209,0.00008587109,0.0063406895,0.000058563466,0.00009082631,0.0009850245,0.0005107828,0.8424877,0.0024945722,0.14205584,0.00477494,0.000038529608],"about_ca_topic_score_codex":0.0038569188,"about_ca_topic_score_gemma":0.0054404153,"teacher_disagreement_score":0.0038569188,"about_ca_system_score_codex":0.0009719478,"about_ca_system_score_gemma":0.0012210057,"threshold_uncertainty_score":0.011968613},"labels":[],"label_agreement":null},{"id":"W2180566385","doi":"10.1145/2733381","title":"Hierarchical Density Estimates for Data Clustering, Visualization, and Outlier Detection","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":857,"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 Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Cluster analysis; Computer science; Visualization; Outlier; Anomaly detection; Data mining; Hierarchy; Hierarchical clustering; Pattern recognition (psychology); CURE data clustering algorithm; Mathematics; Artificial intelligence; Correlation clustering","score_opus":0.09123092616616002,"score_gpt":0.34674055329636305,"score_spread":0.25550962713020303,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2180566385","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00021150032,0.0001250232,0.9989655,0.000040676234,0.000010669499,0.000025327725,0.000052253934,0.00039018437,0.00017881679],"genre_scores_gemma":[0.020256488,0.00047568846,0.9776962,0.000048002024,0.000072882576,0.0002803237,0.00034865446,0.00029595016,0.0005258215],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9958762,0.0017314348,0.00020506475,0.00053713284,0.001494855,0.00015526144],"domain_scores_gemma":[0.9937251,0.0031350162,0.0006781687,0.0009870902,0.001324417,0.0001501807],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0040461863,0.0019239098,0.0017803591,0.0047106766,0.0010044881,0.0031552329,0.00315768,0.001755723,0.0036112522],"category_scores_gemma":[0.01844155,0.0012389336,0.0020991785,0.005171964,0.001430842,0.0030338978,0.003300014,0.0037815683,0.002406699],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00010811778,0.00010712009,0.0014666911,0.0006635342,0.000245612,0.0001679101,0.00063950394,0.32972276,0.008275618,0.32504082,0.014506575,0.31905577],"study_design_scores_gemma":[0.000012996437,0.000028183737,0.00049562904,0.000067131194,0.00002847878,0.000092598384,0.00006836481,0.89071566,0.0028910765,0.09389562,0.011654139,0.00005008999],"about_ca_topic_score_codex":0.00835002,"about_ca_topic_score_gemma":0.0057263873,"teacher_disagreement_score":0.00835002,"about_ca_system_score_codex":0.0022598065,"about_ca_system_score_gemma":0.0025791556,"threshold_uncertainty_score":0.021398544},"labels":[],"label_agreement":null},{"id":"W2214080290","doi":"10.1145/2700409","title":"Classification with Streaming Features: An Emerging-Pattern Mining Approach","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Streaming data; Feature selection; Data mining; Benchmark (surveying); Classifier (UML); Curse of dimensionality; Feature (linguistics); Machine learning; Artificial intelligence; Feature vector; Pattern recognition (psychology)","score_opus":0.11101318228584427,"score_gpt":0.32390555529082665,"score_spread":0.21289237300498237,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2214080290","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.036984995,0.0007916342,0.95903707,0.0003847621,0.00011071055,0.0001713821,0.0006274522,0.0011087628,0.0007832348],"genre_scores_gemma":[0.48803386,0.000791578,0.50621855,0.0001977513,0.00031526896,0.00029129162,0.00220712,0.00012485086,0.0018198299],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99872905,0.0002076957,0.00014900755,0.00043129883,0.0003893489,0.00009358759],"domain_scores_gemma":[0.9963097,0.0012184865,0.00054554717,0.0007189023,0.0010585678,0.00014884931],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017681303,0.00096507336,0.0015211195,0.0029949867,0.00055632344,0.0013139897,0.002436844,0.0009118359,0.0008835172],"category_scores_gemma":[0.0061300136,0.00035565556,0.001008936,0.003925871,0.0005004139,0.0035610385,0.0010938639,0.001398631,0.0006812078],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00035923132,0.0004661958,0.025318941,0.00027782097,0.00024824182,0.00060046604,0.00028235756,0.08947644,0.0098674735,0.008495264,0.00989962,0.85470796],"study_design_scores_gemma":[0.000018925653,0.00010517798,0.0022984303,0.00002155433,0.000041313902,0.00025906437,0.00008384064,0.978508,0.0024484775,0.0128702065,0.0033229408,0.000022117385],"about_ca_topic_score_codex":0.0016668997,"about_ca_topic_score_gemma":0.0019823753,"teacher_disagreement_score":0.0029949867,"about_ca_system_score_codex":0.0004834595,"about_ca_system_score_gemma":0.0006526641,"threshold_uncertainty_score":0.009350836},"labels":[],"label_agreement":null},{"id":"W2257410811","doi":"10.1145/3365672","title":"A New Smooth Approximation to the Zero One Loss with a Probabilistic Interpretation","year":2019,"lang":"en","type":"preprint","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Polytechnique Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Hinge loss; Algorithm; Computer science; Probabilistic logic; Maxima and minima; Artificial intelligence; Mathematics; Benchmark (surveying); Mathematical optimization; Machine learning; Pattern recognition (psychology); Support vector machine","score_opus":0.04480686181851787,"score_gpt":0.29227264341833536,"score_spread":0.2474657815998175,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2257410811","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0058077667,0.00027103073,0.99205667,0.00034605872,0.000033714285,0.000028520426,0.0000786449,0.0003064206,0.0010712254],"genre_scores_gemma":[0.4355817,0.0011948338,0.5498866,0.0008931443,0.00041470397,0.00039327165,0.0010466427,0.00066268526,0.009926406],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99746644,0.0008675138,0.00011310386,0.00048971013,0.00087969523,0.00018353142],"domain_scores_gemma":[0.99556386,0.0023583237,0.00051476,0.0007130739,0.00067967956,0.0001703255],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004260919,0.0016138862,0.0012863508,0.0017857322,0.00049729476,0.0024210613,0.0029379255,0.0018072005,0.002255877],"category_scores_gemma":[0.014436748,0.00068851653,0.0011745951,0.0016983456,0.0021174378,0.004091696,0.0029646056,0.0037572728,0.0012846659],"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":[0.000269928,0.00020992843,0.0031469497,0.00025962866,0.00017242816,0.0003012864,0.00019010599,0.64335495,0.006884985,0.13476355,0.011032556,0.19941373],"study_design_scores_gemma":[0.000013351669,0.0000634009,0.0003336038,0.000020518932,0.000013125329,0.0000777551,0.000010637571,0.9612232,0.0009905138,0.03541619,0.0018194949,0.00001831446],"about_ca_topic_score_codex":0.0019424824,"about_ca_topic_score_gemma":0.0014815384,"teacher_disagreement_score":0.004260919,"about_ca_system_score_codex":0.0013603164,"about_ca_system_score_gemma":0.0018309165,"threshold_uncertainty_score":0.022534132},"labels":[],"label_agreement":null},{"id":"W2257550357","doi":"10.1145/2976744","title":"Scalable and Accurate Online Feature Selection for Big Data","year":2016,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":145,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"National Natural Science Foundation of China","keywords":"Feature selection; Pairwise comparison; Big data; Computer science; Scalability; Feature (linguistics); Benchmark (surveying); Curse of dimensionality; Selection (genetic algorithm); Artificial intelligence; Data mining; Dimensionality reduction; Set (abstract data type); Machine learning; Pattern recognition (psychology); Database","score_opus":0.10563085290882107,"score_gpt":0.31964763730881357,"score_spread":0.2140167843999925,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2257550357","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.026696641,0.0010495348,0.9671476,0.0003862286,0.00008806078,0.00013457613,0.0005054205,0.0033554619,0.00063641113],"genre_scores_gemma":[0.44404927,0.00053376175,0.54866105,0.0003536175,0.0003011262,0.00045786085,0.0034438216,0.000294214,0.00190533],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99806803,0.00042963427,0.000119553486,0.00048530646,0.00072079245,0.00017671539],"domain_scores_gemma":[0.99589765,0.0019991188,0.00038502357,0.0009573074,0.00059052877,0.0001704668],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019229997,0.0019029467,0.0019351295,0.0024027308,0.0010014442,0.0012643882,0.0022734725,0.0010345303,0.0018370845],"category_scores_gemma":[0.008684577,0.00064960506,0.0014017412,0.0032108189,0.0006331607,0.0030545418,0.0015338536,0.0018086745,0.0010738402],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000654714,0.00045799895,0.006358625,0.00025094787,0.00027282952,0.00041856183,0.00018571745,0.22503857,0.016939938,0.0045510354,0.017457651,0.72741336],"study_design_scores_gemma":[0.00005028029,0.00009853363,0.0011674414,0.000008767111,0.000025477457,0.000116445524,0.000056084387,0.98212147,0.0032215067,0.010835889,0.0022798304,0.000018232851],"about_ca_topic_score_codex":0.0039437097,"about_ca_topic_score_gemma":0.0054232692,"teacher_disagreement_score":0.0039437097,"about_ca_system_score_codex":0.00072211726,"about_ca_system_score_gemma":0.0013591968,"threshold_uncertainty_score":0.010169864},"labels":[],"label_agreement":null},{"id":"W2325563383","doi":"10.1145/2297456.2297460","title":"The Latent Maximum Entropy Principle","year":2012,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Statistical Mechanics and Entropy","field":"Physics and Astronomy","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Principle of maximum entropy; Mathematics; Statistics","score_opus":0.04418102761629817,"score_gpt":0.30897019030708406,"score_spread":0.2647891626907859,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2325563383","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0036355732,0.00030202887,0.99115235,0.00049726665,0.000041885334,0.0000375137,0.00011526077,0.00011235746,0.0041057244],"genre_scores_gemma":[0.39697272,0.0016738753,0.5913596,0.0008821021,0.0005597147,0.000737852,0.00064822705,0.00032451516,0.0068414565],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99567777,0.0021608884,0.0001774828,0.00057766115,0.0011569194,0.0002493636],"domain_scores_gemma":[0.9902418,0.0071079605,0.0005502017,0.0013248364,0.00055977073,0.00021542488],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005099506,0.0010447277,0.0013873219,0.0018046646,0.000966015,0.0028969888,0.0021653776,0.0016448573,0.0046051154],"category_scores_gemma":[0.01962971,0.0009034575,0.0017862185,0.0017217279,0.0037602747,0.0060237413,0.004261637,0.0031109212,0.0013166807],"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":[0.00005552924,0.000064343694,0.0012251994,0.0002583129,0.00011730745,0.00015101164,0.00029998703,0.11229876,0.0024784915,0.81557584,0.0028747662,0.064600445],"study_design_scores_gemma":[0.000016856846,0.000042282394,0.00038988845,0.000048426777,0.00001773101,0.00009627628,0.000022290204,0.27214348,0.0011311742,0.7200282,0.0060301824,0.000033162865],"about_ca_topic_score_codex":0.00091022515,"about_ca_topic_score_gemma":0.00079617364,"teacher_disagreement_score":0.005099506,"about_ca_system_score_codex":0.0011612091,"about_ca_system_score_gemma":0.0019133231,"threshold_uncertainty_score":0.026969075},"labels":[],"label_agreement":null},{"id":"W2400248106","doi":"10.1145/2724720","title":"Improving Top-N Recommendation for Cold-Start Users via Cross-Domain Information","year":2015,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":88,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"","keywords":"Cold start (automotive); Computer science; Domain (mathematical analysis); Matrix decomposition; Cluster analysis; Precision and recall; Recommender system; Data mining; Artificial intelligence; Machine learning; Information retrieval; Mathematics; Engineering","score_opus":0.06982123719770345,"score_gpt":0.32001129143094836,"score_spread":0.2501900542332449,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2400248106","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.269363,0.0135685215,0.69204926,0.0014059235,0.0008338083,0.00048009382,0.0043187025,0.0060037603,0.011976899],"genre_scores_gemma":[0.65921223,0.0016971756,0.32188535,0.0007986244,0.00041757882,0.0001782186,0.007461824,0.00036766884,0.00798136],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9969938,0.0008374988,0.00022595836,0.0011807355,0.000515231,0.00024679993],"domain_scores_gemma":[0.9892943,0.004845143,0.0006473848,0.0025153882,0.0022272335,0.00047062058],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003890548,0.0021767644,0.0030673265,0.0037882205,0.0017369401,0.0017368839,0.0024233644,0.0020571013,0.002699725],"category_scores_gemma":[0.014930118,0.0008323142,0.0018168014,0.0038364925,0.0006673021,0.004701747,0.0014288096,0.0023159569,0.0038098928],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016276342,0.0024350917,0.094062045,0.0012817826,0.0012539381,0.0007392413,0.0011554067,0.09770698,0.019495986,0.0048151314,0.06753005,0.7078967],"study_design_scores_gemma":[0.00015129364,0.00048670702,0.012007319,0.00018434491,0.00047805722,0.0011047043,0.0007408861,0.95024145,0.009766207,0.011514737,0.013114186,0.0002102466],"about_ca_topic_score_codex":0.015568234,"about_ca_topic_score_gemma":0.053301122,"teacher_disagreement_score":0.015568234,"about_ca_system_score_codex":0.0007710382,"about_ca_system_score_gemma":0.0018234948,"threshold_uncertainty_score":0.030955195},"labels":[],"label_agreement":null},{"id":"W2405494619","doi":"10.1145/2983533","title":"Structural Analysis of User Choices for Mobile App Recommendation","year":2016,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":38,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Fundamental Research Funds for the Central Universities; National Natural Science Foundation of China","keywords":"Computer science; Mobile apps; App store; Variety (cybernetics); Recommender system; World Wide Web; Margin (machine learning); Focus (optics); Taxonomy (biology); Data science; Human–computer interaction; Artificial intelligence; Machine learning","score_opus":0.05976271325554005,"score_gpt":0.3313509147680883,"score_spread":0.27158820151254826,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2405494619","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.38084263,0.0011444913,0.60967463,0.0019866724,0.00006141481,0.00042054665,0.0023287723,0.0004809297,0.003059961],"genre_scores_gemma":[0.9377237,0.00041627165,0.056957137,0.00015881906,0.000051294926,0.00027289064,0.0017685118,0.00003891182,0.0026124064],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99721634,0.001549976,0.00013035505,0.0005795545,0.00033551315,0.0001882523],"domain_scores_gemma":[0.9807716,0.01585925,0.0010237206,0.001203682,0.0007573565,0.0003843436],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0045071775,0.00081601227,0.0013128958,0.002259586,0.00080158695,0.00097532204,0.0017783899,0.0015982859,0.005125963],"category_scores_gemma":[0.021386435,0.00085518614,0.0020934131,0.0024276022,0.0010089182,0.002695148,0.0010399544,0.0019930578,0.00096575724],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010356538,0.0009398473,0.19171575,0.00077232055,0.0010538365,0.00061556156,0.0018223578,0.4648748,0.0038291663,0.13454747,0.009059825,0.18973342],"study_design_scores_gemma":[0.000015574959,0.00008021106,0.0071346946,0.000023193896,0.00003452053,0.00006717318,0.00007924354,0.96303886,0.00019828624,0.028457945,0.0008494901,0.000020752148],"about_ca_topic_score_codex":0.016506411,"about_ca_topic_score_gemma":0.03453283,"teacher_disagreement_score":0.016506411,"about_ca_system_score_codex":0.00199139,"about_ca_system_score_gemma":0.0013463259,"threshold_uncertainty_score":0.032820642},"labels":[],"label_agreement":null},{"id":"W2504431993","doi":"10.1145/2948068","title":"The Convergence Behavior of Naive Bayes on Large Sparse Datasets","year":2016,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Naive Bayes classifier; Computer science; Machine learning; Missing data; Bayes' theorem; Artificial intelligence; Convergence (economics); Data mining; Bayesian probability; Support vector machine","score_opus":0.05267241250637728,"score_gpt":0.3180910509676954,"score_spread":0.26541863846131813,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2504431993","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.09432787,0.0029717775,0.8919097,0.0034340478,0.00019211178,0.00034594798,0.0004199694,0.0011863706,0.0052121314],"genre_scores_gemma":[0.6786148,0.0015122847,0.3140092,0.0013213309,0.00032443617,0.00055166776,0.0012003768,0.00033771736,0.0021282341],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98355776,0.009208647,0.00079357455,0.002027664,0.0036121397,0.00080023776],"domain_scores_gemma":[0.8319974,0.1399123,0.0046864166,0.009698394,0.012082728,0.0016226865],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.036620706,0.0013798526,0.0024467423,0.0035789828,0.0019411396,0.0027825641,0.002111165,0.00234012,0.0017488728],"category_scores_gemma":[0.17987716,0.0008991663,0.001350531,0.002164804,0.003616876,0.006263993,0.0025259904,0.0036210935,0.0008480089],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016339301,0.00041432978,0.03283789,0.0010615701,0.0005448213,0.00059321284,0.001679861,0.50492823,0.0045009507,0.17106402,0.014142547,0.2665986],"study_design_scores_gemma":[0.00006888292,0.00015271592,0.0019556514,0.0001453485,0.000036235626,0.00028453703,0.00017620594,0.84707665,0.0014346929,0.14728282,0.0013410528,0.00004518577],"about_ca_topic_score_codex":0.004712497,"about_ca_topic_score_gemma":0.0034031738,"teacher_disagreement_score":0.036620706,"about_ca_system_score_codex":0.00243485,"about_ca_system_score_gemma":0.001975697,"threshold_uncertainty_score":0.1936711},"labels":[],"label_agreement":null},{"id":"W2602756700","doi":"10.1145/3046941","title":"An Influence Propagation View of PageRank","year":2017,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Opinion Dynamics and Social Influence","field":"Physics and Astronomy","cited_by":82,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"National Natural Science Foundation of China; National Science Fund for Distinguished Young Scholars; Youth Innovation Promotion Association; National Science Foundation","keywords":"PageRank; Computer science; Ranking (information retrieval); Computation; Constraint (computer-aided design); Node (physics); Theoretical computer science; Data mining; Information retrieval; Algorithm; Mathematics","score_opus":0.041134871033084615,"score_gpt":0.345878011987555,"score_spread":0.3047431409544704,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2602756700","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00864299,0.001885635,0.9496561,0.0026086117,0.00020543324,0.00007108715,0.0002864632,0.00039498156,0.036248703],"genre_scores_gemma":[0.77564645,0.0066287057,0.17936242,0.0009647612,0.002081324,0.0003380799,0.000645915,0.00025856652,0.034073804],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99847704,0.00042691783,0.00005133149,0.00031701705,0.00062376726,0.000103871906],"domain_scores_gemma":[0.9968976,0.0016849298,0.00033450176,0.00034276865,0.0005911713,0.00014892152],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013728223,0.0009984666,0.00093397073,0.0031153415,0.00089989346,0.0031667901,0.0017615218,0.0014000254,0.0051739565],"category_scores_gemma":[0.007133948,0.00047016333,0.0008065512,0.0025500862,0.0021541528,0.0060232566,0.0010336476,0.0021684293,0.0014498257],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000027695027,0.000032733828,0.00078881084,0.0001680523,0.000054743516,0.00011502632,0.00022426246,0.06631402,0.0011739898,0.8932958,0.0053898217,0.032415025],"study_design_scores_gemma":[0.000019878278,0.000041974734,0.00040224157,0.00003806987,0.000036888836,0.00014613001,0.000055906803,0.38785726,0.0006470433,0.5943128,0.01641701,0.000024754609],"about_ca_topic_score_codex":0.004847981,"about_ca_topic_score_gemma":0.0037353155,"teacher_disagreement_score":0.0051739565,"about_ca_system_score_codex":0.0016661176,"about_ca_system_score_gemma":0.00090212206,"threshold_uncertainty_score":0.017308593},"labels":[],"label_agreement":null},{"id":"W2724995883","doi":"10.1145/3080574","title":"Mining Community Structures in Multidimensional Networks","year":2017,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":41,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec à Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Relevance (law); Identification (biology); Set (abstract data type); Data mining; Node (physics); Function (biology); Machine learning; Feature (linguistics); Artificial intelligence; Data science","score_opus":0.06927095770474913,"score_gpt":0.3432355618617533,"score_spread":0.2739646041570042,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2724995883","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06779605,0.0014112645,0.9278592,0.000583895,0.000039294282,0.00014145758,0.0005157563,0.00026699627,0.0013860762],"genre_scores_gemma":[0.5175289,0.0011303708,0.47838667,0.00017375968,0.00014769917,0.00030620763,0.0012444227,0.000065020206,0.0010170513],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9969614,0.0011529543,0.0001613263,0.0008433151,0.0006877701,0.0001932527],"domain_scores_gemma":[0.9880393,0.0074688485,0.0022243361,0.00093639025,0.0009880707,0.00034302942],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024748563,0.0009803575,0.0010790372,0.0073756427,0.0016633953,0.0019370458,0.0018078131,0.0020060493,0.0007324074],"category_scores_gemma":[0.015047089,0.0006916371,0.0011063109,0.0054631075,0.0014361133,0.0038423177,0.0027403033,0.0015226136,0.00034285535],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002030073,0.00037526563,0.049677305,0.0012890898,0.0005847695,0.0011486635,0.0018947544,0.54746133,0.01442554,0.10869073,0.00666435,0.26758513],"study_design_scores_gemma":[0.0000156501,0.000041513173,0.0039365045,0.00007491137,0.00004964223,0.00027512803,0.00034330806,0.8966676,0.0025343457,0.09237883,0.0036470604,0.000035523302],"about_ca_topic_score_codex":0.0021716447,"about_ca_topic_score_gemma":0.0025158094,"teacher_disagreement_score":0.0073756427,"about_ca_system_score_codex":0.0010060014,"about_ca_system_score_gemma":0.0006864919,"threshold_uncertainty_score":0.013088465},"labels":[],"label_agreement":null},{"id":"W2786524829","doi":"10.1145/3047010","title":"ATR-Vis","year":2018,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Video Analysis and Summarization","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Dalhousie University","funders":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação de Amparo à Pesquisa do Estado de São Paulo; Natural Sciences and Engineering Research Council of Canada; International Development Research Centre","keywords":"Computer science; Variety (cybernetics); Terminology; Task (project management); Information retrieval; Set (abstract data type); Visualization; Data science; World Wide Web; Data mining; Artificial intelligence","score_opus":0.048978138937205005,"score_gpt":0.29976558278340987,"score_spread":0.2507874438462049,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2786524829","genre_codex":"software","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.024106322,0.0031187835,0.18371814,0.002259042,0.0013055295,0.0015112882,0.21235305,0.40921232,0.16241547],"genre_scores_gemma":[0.12667927,0.0020236326,0.28467333,0.0017997978,0.00058439764,0.0024725539,0.41769165,0.038615685,0.1254597],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99860483,0.000285358,0.00009289219,0.00036113488,0.0005100364,0.00014583697],"domain_scores_gemma":[0.9975546,0.0005458175,0.00016951728,0.00087985815,0.00063395506,0.00021637946],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0015939303,0.001548579,0.0006746308,0.002560374,0.000855409,0.0034870198,0.0015539624,0.0015911721,0.06988068],"category_scores_gemma":[0.0064744703,0.0006004754,0.0012048119,0.002037151,0.00041067527,0.003505655,0.0030861294,0.0013797127,0.07513471],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010653834,0.00021359915,0.003542906,0.001479408,0.000114877876,0.00026240668,0.0013158866,0.0028801067,0.018335607,0.012075163,0.6799309,0.27878377],"study_design_scores_gemma":[0.00016604009,0.00021747581,0.0041589793,0.00021748742,0.000043646494,0.00047581564,0.0006853417,0.025430506,0.012758746,0.015012151,0.9407105,0.00012330734],"about_ca_topic_score_codex":0.00560389,"about_ca_topic_score_gemma":0.009078515,"teacher_disagreement_score":0.93011934,"about_ca_system_score_codex":0.00071202783,"about_ca_system_score_gemma":0.0010743681,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W2791982474","doi":"10.1145/3023363","title":"Visual Analysis of Brain Networks Using Sparse Regression Models","year":2018,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Visualization and Analytics","field":"Computer Science","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":"Army Research Office; National Institute of Biomedical Imaging and Bioengineering; DoD Alzheimer's Disease Neuroimaging Initiative; Defense Threat Reduction Agency; Canadian Institutes of Health Research; National Institute on Aging; National Institutes of Health; National Natural Science Foundation of China","keywords":"Visual analytics; Computer science; Discriminative model; Visualization; Artificial intelligence; Machine learning; Intersection (aeronautics); Human–computer interaction","score_opus":0.09753420314028476,"score_gpt":0.3740269390566549,"score_spread":0.27649273591637014,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2791982474","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.023424849,0.00014821572,0.9728295,0.00043853454,0.000016825335,0.00004137179,0.00021897179,0.0019529404,0.000928808],"genre_scores_gemma":[0.42772907,0.00061358063,0.568591,0.00016794608,0.0000659884,0.00017219005,0.00066615676,0.000639444,0.0013546677],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994254,0.00029857422,0.000019308301,0.00010080393,0.00012256944,0.00003334184],"domain_scores_gemma":[0.99803835,0.0013162958,0.00019161207,0.00019782614,0.00018083015,0.00007499838],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016096898,0.0010958523,0.0005614434,0.001945667,0.00031346877,0.0016762264,0.00088977517,0.000683139,0.0036771984],"category_scores_gemma":[0.0071847485,0.00043914327,0.0010103817,0.0008117484,0.00056002045,0.0017171132,0.0019379568,0.001229253,0.000584753],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000479549,0.0001406491,0.0028309964,0.0003866765,0.00016258548,0.00038702218,0.0009692682,0.6431458,0.050222035,0.029067373,0.007708958,0.264499],"study_design_scores_gemma":[0.000015937367,0.000034763896,0.00051665783,0.000018743864,0.000011689285,0.00008115799,0.00007380589,0.9744108,0.004492682,0.018909952,0.0014190995,0.000014662879],"about_ca_topic_score_codex":0.0021282972,"about_ca_topic_score_gemma":0.003166077,"teacher_disagreement_score":0.0036771984,"about_ca_system_score_codex":0.0005750563,"about_ca_system_score_gemma":0.0005131713,"threshold_uncertainty_score":0.012301445},"labels":[],"label_agreement":null},{"id":"W2895492593","doi":"10.1145/3237191","title":"Enumerating Trillion Subgraphs On Distributed Systems","year":2018,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Graph Theory and Algorithms","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Kootenay Association for Science & Technology","funders":"Chongqing Research Program of Basic Research and Frontier Technology","keywords":"Enumeration; Computer science; Scalability; Theoretical computer science; Graph; Algorithm; Combinatorics; Mathematics; Database","score_opus":0.04566446021945211,"score_gpt":0.28904636755667584,"score_spread":0.24338190733722373,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2895492593","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.15992759,0.00058508565,0.8274297,0.00085604575,0.0000841735,0.00023648415,0.0009616364,0.00542403,0.004495275],"genre_scores_gemma":[0.5527877,0.0004165529,0.43873712,0.0002528307,0.00005218238,0.00032071004,0.002627997,0.00044657395,0.0043583172],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987999,0.00024819485,0.00008482573,0.00034366548,0.00032424222,0.00019927254],"domain_scores_gemma":[0.9967264,0.0012942333,0.00030049623,0.0009715089,0.00050034176,0.00020700326],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006562368,0.0006812023,0.00085031166,0.001318931,0.0010093171,0.0011916743,0.0017618337,0.00065183954,0.002588041],"category_scores_gemma":[0.0050647995,0.00044413222,0.00072363624,0.0019992085,0.0008612851,0.004051041,0.0020218897,0.0008265829,0.000791176],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00055957964,0.0002502761,0.008487561,0.00069074356,0.00016017524,0.00056164846,0.0011707646,0.5629785,0.02795366,0.069433816,0.01769184,0.31006145],"study_design_scores_gemma":[0.000053274856,0.000069014364,0.00079641177,0.000019855694,0.000026829035,0.00017930863,0.00031053164,0.8937838,0.007976204,0.09074271,0.006024022,0.00001806494],"about_ca_topic_score_codex":0.00465644,"about_ca_topic_score_gemma":0.00631414,"teacher_disagreement_score":0.00465644,"about_ca_system_score_codex":0.001003861,"about_ca_system_score_gemma":0.0010884948,"threshold_uncertainty_score":0.0092586875},"labels":[],"label_agreement":null},{"id":"W3011191765","doi":"10.1145/3380928","title":"Continuous Influence Maximization","year":2020,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Maximization; Computer science; Benchmark (surveying); Mathematical optimization; Product (mathematics); Function (biology); Mathematics","score_opus":0.041078501605940246,"score_gpt":0.291683042845804,"score_spread":0.25060454123986375,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3011191765","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0077217226,0.0004294049,0.98496765,0.00029185376,0.000039409013,0.00007586874,0.00014784504,0.00016531842,0.0061608898],"genre_scores_gemma":[0.6202383,0.0015794293,0.36470354,0.0003776563,0.0003317239,0.0005556305,0.00068852096,0.00027865358,0.011246514],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99807537,0.000865783,0.00007791579,0.00048369856,0.000347292,0.00014988442],"domain_scores_gemma":[0.99437225,0.004555423,0.00024744077,0.0003298759,0.00034331216,0.0001517984],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025860586,0.0014630452,0.0016513782,0.0009892838,0.0006375557,0.0018562133,0.0018492249,0.0013911327,0.0051892865],"category_scores_gemma":[0.011717546,0.0006758786,0.0012648887,0.0014119591,0.001684666,0.0022749763,0.0014410409,0.001601709,0.0010673855],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011005736,0.00007610811,0.0015317011,0.00036858852,0.00012832366,0.00019686794,0.00021626148,0.68780327,0.0019814454,0.22267818,0.0052661393,0.07964313],"study_design_scores_gemma":[0.000015481497,0.000027309365,0.00021934789,0.000017219725,0.000020334252,0.000053923817,0.000018386574,0.9270104,0.0007059812,0.06926904,0.0026314869,0.000011165206],"about_ca_topic_score_codex":0.0037503361,"about_ca_topic_score_gemma":0.0032995169,"teacher_disagreement_score":0.0051892865,"about_ca_system_score_codex":0.0016614738,"about_ca_system_score_gemma":0.001273557,"threshold_uncertainty_score":0.017359912},"labels":[],"label_agreement":null},{"id":"W3027864066","doi":"10.1145/3377939","title":"Story Forest","year":2020,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":60,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Timeline; Computer science; Event (particle physics); Information retrieval; The Internet; Novelty; World Wide Web; Cluster analysis; Set (abstract data type); News aggregator; Graph; Data science; Artificial intelligence; History","score_opus":0.06664237805374586,"score_gpt":0.31143453917885444,"score_spread":0.2447921611251086,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3027864066","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01626917,0.0039834026,0.5543593,0.002554548,0.0015691004,0.0021011687,0.1660955,0.089188926,0.16387884],"genre_scores_gemma":[0.09396937,0.002320024,0.48769617,0.0013040793,0.00039541433,0.0013803239,0.33396837,0.005827719,0.07313858],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99916327,0.00012693045,0.000059221944,0.00031697543,0.00023451386,0.00009914594],"domain_scores_gemma":[0.9990376,0.00032152433,0.000065522654,0.00025934642,0.00023483818,0.0000811186],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006922385,0.0017553822,0.0006985309,0.0031074923,0.0013317042,0.0026058871,0.0022977763,0.0013025184,0.07456609],"category_scores_gemma":[0.0038586634,0.00062553026,0.0018523503,0.0025705355,0.0003868566,0.004175462,0.0022834626,0.0014784132,0.03810922],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00035075351,0.0001809252,0.0028700111,0.00090498273,0.00013052174,0.00045538624,0.00024213098,0.012356058,0.002915271,0.02594774,0.4491531,0.50449306],"study_design_scores_gemma":[0.00013533577,0.000116324656,0.00206778,0.00023571418,0.00010193977,0.0010177935,0.00044214897,0.17169416,0.0063846027,0.08241154,0.73532397,0.000068710135],"about_ca_topic_score_codex":0.004977882,"about_ca_topic_score_gemma":0.0142069375,"teacher_disagreement_score":0.07456609,"about_ca_system_score_codex":0.00067785446,"about_ca_system_score_gemma":0.0013831878,"threshold_uncertainty_score":0.24944842},"labels":[],"label_agreement":null},{"id":"W3039137796","doi":"10.1145/3394053","title":"Internal Evaluation of Unsupervised Outlier Detection","year":2020,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":33,"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 Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Outlier; Anomaly detection; Computer science; Cluster analysis; Data mining; Artificial intelligence; Domain (mathematical analysis); Pattern recognition (psychology); Binary number; Mathematics","score_opus":0.1100711224031969,"score_gpt":0.32755916439882915,"score_spread":0.21748804199563226,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3039137796","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.22348912,0.0018340844,0.7572403,0.00060202956,0.0004099645,0.0005633871,0.0017460694,0.00428494,0.009830108],"genre_scores_gemma":[0.72637564,0.00035313668,0.2666107,0.00021044165,0.00015444211,0.0003283015,0.003687615,0.00060446805,0.0016751882],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9721696,0.008948387,0.003028749,0.003417876,0.011514845,0.00092058204],"domain_scores_gemma":[0.92990345,0.030285979,0.009880595,0.008466583,0.01983213,0.0016312662],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.025752824,0.002216955,0.0022693833,0.0069388095,0.0013422513,0.0043837577,0.0022822905,0.002177985,0.0013203135],"category_scores_gemma":[0.09601629,0.00038174455,0.0014362391,0.0040000654,0.0019678562,0.0037940391,0.003600796,0.0016943484,0.00080510677],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0018104624,0.0009420441,0.116428226,0.0015851918,0.0015303219,0.00027381652,0.0011277588,0.22662845,0.023388946,0.02658168,0.0135216545,0.5861814],"study_design_scores_gemma":[0.00011745789,0.0015819336,0.031221693,0.00029469692,0.00028877382,0.00047772806,0.0009715053,0.88071394,0.05042863,0.026392743,0.007284181,0.0002267736],"about_ca_topic_score_codex":0.001395493,"about_ca_topic_score_gemma":0.001960982,"teacher_disagreement_score":0.025752824,"about_ca_system_score_codex":0.0015169097,"about_ca_system_score_gemma":0.0021615648,"threshold_uncertainty_score":0.1361956},"labels":[],"label_agreement":null},{"id":"W3044604502","doi":"10.1145/3480243","title":"Time-Aware Graph Embedding: A Temporal Smoothness and Task-Oriented Approach","year":2021,"lang":"en","type":"preprint","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Graph Neural Networks","field":"Computer Science","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":"Simon Fraser University","funders":"","keywords":"Embedding; Computer science; Granularity; Theoretical computer science; Graph embedding; Graph; Smoothing; Artificial intelligence; Machine learning","score_opus":0.038241671914906405,"score_gpt":0.2946700430901839,"score_spread":0.2564283711752775,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3044604502","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012272867,0.00029941532,0.9859934,0.00018845533,0.000032403044,0.00004306302,0.00013002196,0.0005526141,0.0004877389],"genre_scores_gemma":[0.46358025,0.000868818,0.5289403,0.00026141937,0.00019716563,0.0002121404,0.0016109612,0.0003725995,0.0039562974],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99895597,0.00025442563,0.00007038875,0.00040705362,0.00023727246,0.00007498383],"domain_scores_gemma":[0.99708384,0.0012929261,0.0003649219,0.0006855264,0.00042015134,0.00015258849],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012793861,0.001256676,0.0011496893,0.0019089617,0.0004874457,0.0010085703,0.0022253066,0.0014588346,0.0013754691],"category_scores_gemma":[0.0061427276,0.00051068637,0.0011076302,0.0026215306,0.000873297,0.0039623603,0.002074467,0.0020947272,0.00046968582],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003223035,0.0003286632,0.0035123534,0.00030249616,0.00017521167,0.00019481343,0.00041673606,0.43508935,0.013710253,0.035510514,0.0056475555,0.50478965],"study_design_scores_gemma":[0.000012056995,0.000055889817,0.00044089725,0.0000099143335,0.0000296135,0.0000557201,0.000036930556,0.97500086,0.0014535381,0.021338977,0.0015486577,0.000016985228],"about_ca_topic_score_codex":0.0042020283,"about_ca_topic_score_gemma":0.006217894,"teacher_disagreement_score":0.0042020283,"about_ca_system_score_codex":0.00070941885,"about_ca_system_score_gemma":0.0010698799,"threshold_uncertainty_score":0.008355141},"labels":[],"label_agreement":null},{"id":"W3119526997","doi":"10.1145/3428077","title":"Context-Based Evaluation of Dimensionality Reduction Algorithms—Experiments and Statistical Significance Analysis","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"IBM (Canada); University of Alberta","funders":"","keywords":"Dimensionality reduction; Computer science; Data mining; Curse of dimensionality; Algorithm; Redundancy (engineering); Context (archaeology); Generalizability theory; Parametric statistics; Machine learning; Reduction (mathematics); Artificial intelligence; Mathematics; Statistics","score_opus":0.10026200597180086,"score_gpt":0.3783655596835932,"score_spread":0.27810355371179235,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3119526997","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.48105308,0.010566731,0.49179637,0.0010280301,0.00092398736,0.0031424754,0.0032426026,0.0022629434,0.0059837564],"genre_scores_gemma":[0.62534034,0.0014857595,0.36517912,0.00039612432,0.00029481048,0.0030957975,0.0033506497,0.00028408066,0.0005732671],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9557993,0.028911203,0.004028373,0.0039207404,0.006690117,0.00065024156],"domain_scores_gemma":[0.880228,0.08487849,0.0063193473,0.015542696,0.012070875,0.0009605265],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.029753583,0.0019666755,0.0012087304,0.0034269502,0.00120524,0.0018859548,0.0015803485,0.001726404,0.0011354246],"category_scores_gemma":[0.12699865,0.00037139925,0.0017848383,0.003211533,0.002067736,0.0027539977,0.0023221823,0.0017453001,0.0004276221],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0054697706,0.0051478054,0.06422506,0.0073135,0.004243555,0.0005568605,0.0021038267,0.27871054,0.035479415,0.03219227,0.016555851,0.5480015],"study_design_scores_gemma":[0.0011130575,0.011800224,0.046745036,0.0011502225,0.0015710675,0.0011375225,0.0018019337,0.7949512,0.0742805,0.043782156,0.02111079,0.00055621256],"about_ca_topic_score_codex":0.0012637608,"about_ca_topic_score_gemma":0.0012244647,"teacher_disagreement_score":0.029753583,"about_ca_system_score_codex":0.001007082,"about_ca_system_score_gemma":0.0014962109,"threshold_uncertainty_score":0.15735388},"labels":[],"label_agreement":null},{"id":"W3120491054","doi":"10.1145/3424672","title":"Knowledge Graph Embedding for Link Prediction","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":476,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Variety (cybernetics); Embedding; Link (geometry); Graph; Scale (ratio); Knowledge graph; Data science; Machine learning; Data mining; Artificial intelligence; Theoretical computer science","score_opus":0.06056504407758513,"score_gpt":0.32793144493453946,"score_spread":0.26736640085695434,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3120491054","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.033087514,0.006256744,0.92954856,0.001097339,0.0003032034,0.00029864215,0.0102200955,0.0094676595,0.009720197],"genre_scores_gemma":[0.43948847,0.0037001541,0.52077687,0.00043675833,0.00029389968,0.0002922476,0.028225949,0.00059184455,0.0061937296],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9982754,0.0003791427,0.00011174353,0.00057624216,0.0005585456,0.00009890614],"domain_scores_gemma":[0.9929838,0.004046838,0.0005855557,0.0014443012,0.00081991346,0.000119553806],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011663268,0.0016168355,0.0010665932,0.005101556,0.00061952684,0.0016779238,0.001806926,0.0015226934,0.0053251158],"category_scores_gemma":[0.0131077375,0.0004635409,0.00089898077,0.0059369365,0.00054822443,0.004817743,0.001373057,0.0019737799,0.0031823518],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017730065,0.00028220462,0.0072045033,0.00077032344,0.0002234612,0.00019861257,0.00016733073,0.17629872,0.002172216,0.021393703,0.03528995,0.75582165],"study_design_scores_gemma":[0.000019075254,0.000070087175,0.0015427099,0.00014200281,0.00007975237,0.00023126285,0.0001117621,0.9130898,0.0038193653,0.06280353,0.01806149,0.000029183151],"about_ca_topic_score_codex":0.006439603,"about_ca_topic_score_gemma":0.009317739,"teacher_disagreement_score":0.006439603,"about_ca_system_score_codex":0.00083456526,"about_ca_system_score_gemma":0.0012975304,"threshold_uncertainty_score":0.017814279},"labels":[],"label_agreement":null},{"id":"W3163912335","doi":"10.1145/3448021","title":"Similarity Embedding Networks for Robust Human Activity Recognition","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Context-Aware Activity Recognition Systems","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Artificial intelligence; Computer science; Embedding; Pattern recognition (psychology); Similarity (geometry); Pairwise comparison; Activity recognition; Cluster analysis; Convolutional neural network; Deep learning; Machine learning; Generalization; Data mining; Mathematics","score_opus":0.1595380997247741,"score_gpt":0.34323993684266196,"score_spread":0.18370183711788787,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3163912335","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01856285,0.0006486047,0.9782406,0.00016335702,0.000056562567,0.00003702268,0.0001871744,0.0011401554,0.00096374855],"genre_scores_gemma":[0.80949944,0.0008149525,0.18284464,0.00023962243,0.00011315863,0.0001692256,0.0013645133,0.00020021183,0.0047541787],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991352,0.00022667921,0.000057894693,0.00032548237,0.00018170083,0.00007298773],"domain_scores_gemma":[0.99914074,0.0003126073,0.00015736475,0.00018899895,0.00015953314,0.000040843734],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00094928266,0.0009024217,0.0009630568,0.00080234616,0.00021094742,0.00077127636,0.0012256352,0.00088008284,0.0017916544],"category_scores_gemma":[0.004010358,0.0004057517,0.00061702565,0.0010159164,0.00066439575,0.002010965,0.0014274693,0.0015849688,0.00081897667],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002435842,0.00017740742,0.0016352293,0.00014909726,0.000119732314,0.000083099716,0.000103350154,0.42253304,0.010330642,0.015066339,0.0042276224,0.54533076],"study_design_scores_gemma":[0.0000040886316,0.0000307113,0.00033670425,0.000006128843,0.0000058769597,0.000023235738,0.000009236749,0.988605,0.0020191702,0.008323506,0.00062844553,0.000007926209],"about_ca_topic_score_codex":0.0022457468,"about_ca_topic_score_gemma":0.0026852486,"teacher_disagreement_score":0.0022457468,"about_ca_system_score_codex":0.0007221204,"about_ca_system_score_gemma":0.00047457803,"threshold_uncertainty_score":0.0059936643},"labels":[],"label_agreement":null},{"id":"W3209720970","doi":"10.1145/3473910","title":"Semi-Supervised Ensemble Learning for Dealing with Inaccurate and Incomplete Supervision","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"IBM (Canada); University of Alberta","funders":"","keywords":"Computer science; Machine learning; Artificial intelligence; Classifier (UML); Ensemble learning; Supervised learning; Noise (video); Semi-supervised learning; Labeled data; Data mining; Artificial neural network","score_opus":0.059776607221555915,"score_gpt":0.3027910564366063,"score_spread":0.2430144492150504,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3209720970","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012962757,0.0003550395,0.9853951,0.00009248387,0.000031003634,0.00004045676,0.000050838942,0.0006173758,0.00045498414],"genre_scores_gemma":[0.49293673,0.00046382417,0.5031748,0.00021892761,0.00023529105,0.00022517113,0.00093192677,0.00016477078,0.0016487296],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.996216,0.0014343736,0.00021343476,0.00093566935,0.0010201473,0.00018044145],"domain_scores_gemma":[0.991839,0.003455444,0.000776606,0.0021548085,0.0015604105,0.00021378021],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006365681,0.0015230704,0.0025569396,0.0019026941,0.0010193475,0.0012056988,0.00304222,0.0018699985,0.0008660409],"category_scores_gemma":[0.012101019,0.0006573084,0.0012362448,0.00159379,0.001086938,0.0029535748,0.0021961953,0.0029151042,0.0006696826],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00033216883,0.00034177722,0.006472449,0.00028924987,0.00038267163,0.00022805888,0.0005591214,0.389867,0.009666024,0.011268855,0.0051970575,0.5753955],"study_design_scores_gemma":[0.000006389278,0.000045888875,0.00036828997,0.000014147249,0.000022209253,0.000047194077,0.000025633439,0.990003,0.0020591482,0.006590726,0.0008043593,0.000012993988],"about_ca_topic_score_codex":0.0018281727,"about_ca_topic_score_gemma":0.0032925883,"teacher_disagreement_score":0.006365681,"about_ca_system_score_codex":0.00061383215,"about_ca_system_score_gemma":0.001248289,"threshold_uncertainty_score":0.03366536},"labels":[],"label_agreement":null},{"id":"W3212326024","doi":"10.1145/3487046","title":"Scalable Mining of High-Utility Sequential Patterns With Three-Tier MapReduce Model","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Brandon University","funders":"Narodowe Centrum Badań i Rozwoju; Høgskulen på Vestlandet","keywords":"Computer science; Scalability; Correctness; Data mining; Completeness (order theory); Computation; Scale (ratio); Database; Algorithm","score_opus":0.05944832735846505,"score_gpt":0.28622481848680353,"score_spread":0.22677649112833848,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3212326024","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.040699642,0.00026293215,0.954759,0.00036084384,0.00004706087,0.00016915494,0.00046283123,0.001503985,0.0017346542],"genre_scores_gemma":[0.6603029,0.00048574808,0.33424774,0.00015905687,0.00006949661,0.00031837748,0.0011394785,0.0001226553,0.003154552],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9989473,0.0001765561,0.000078411555,0.00028560398,0.00034465434,0.00016754138],"domain_scores_gemma":[0.99912006,0.00022245687,0.000065005406,0.0002340859,0.00026712168,0.00009122945],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011341069,0.0006174631,0.0009325662,0.00078743784,0.0007444914,0.0012244397,0.002463921,0.0004940712,0.0011360202],"category_scores_gemma":[0.0021451805,0.00044881503,0.0014130144,0.0014526948,0.00048290362,0.0020525418,0.0012652676,0.00064970844,0.00033594563],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005822737,0.00042695412,0.0110723255,0.0004835542,0.00036171288,0.00096212665,0.00051524915,0.7235528,0.01425434,0.05944934,0.011375285,0.17696403],"study_design_scores_gemma":[0.000015714528,0.000033633234,0.00035969142,0.0000032051134,0.0000207153,0.000102937214,0.00004794797,0.98102826,0.0011664644,0.016091475,0.0011210613,0.000008944599],"about_ca_topic_score_codex":0.008288474,"about_ca_topic_score_gemma":0.010885433,"teacher_disagreement_score":0.008288474,"about_ca_system_score_codex":0.00070872216,"about_ca_system_score_gemma":0.0024300606,"threshold_uncertainty_score":0.016480446},"labels":[],"label_agreement":null},{"id":"W4205969642","doi":"10.1145/3495011","title":"sGrapp: Butterfly Approximation in Streaming Graphs","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":23,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Bipartite graph; Theoretical computer science; Streaming algorithm; Graph; Mathematics; Upper and lower bounds","score_opus":0.034768323639447415,"score_gpt":0.2913620502884173,"score_spread":0.25659372664896984,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4205969642","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.053211987,0.0005491522,0.93649936,0.0005823655,0.00012928963,0.000156027,0.0007024371,0.0057037296,0.0024655438],"genre_scores_gemma":[0.30436617,0.00040629157,0.6886351,0.00030299925,0.00013008523,0.00025687294,0.002013742,0.0006824942,0.0032063082],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992397,0.000110899884,0.000044790686,0.00025103753,0.00023184363,0.00012175696],"domain_scores_gemma":[0.9970964,0.0014700047,0.0002607719,0.00067775196,0.0003178523,0.00017716455],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011501384,0.0010966663,0.0011821318,0.0016488928,0.0008120152,0.0015596672,0.0028932241,0.0012214478,0.003907339],"category_scores_gemma":[0.00900408,0.0005149547,0.0007662442,0.0021970293,0.0007535898,0.0035189297,0.0018277246,0.0016921503,0.0012364438],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000641951,0.00024356278,0.005858849,0.00039974568,0.00015372495,0.00024972827,0.0003869247,0.5152089,0.01097521,0.050785374,0.025842885,0.38925314],"study_design_scores_gemma":[0.000023534103,0.00003133776,0.00024493554,0.000008209091,0.000008652174,0.00005206506,0.000037414786,0.98014563,0.0013308568,0.016592273,0.0015191173,0.0000059337776],"about_ca_topic_score_codex":0.010488982,"about_ca_topic_score_gemma":0.01342499,"teacher_disagreement_score":0.010488982,"about_ca_system_score_codex":0.0013312459,"about_ca_system_score_gemma":0.0022498544,"threshold_uncertainty_score":0.020855844},"labels":[],"label_agreement":null},{"id":"W4220966911","doi":"10.1145/3502727","title":"Stochastic Variational Optimization of a Hierarchical Dirichlet Process Latent Beta-Liouville Topic Model","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Latent Dirichlet allocation; Dirichlet process; Hierarchical Dirichlet process; Topic model; Computer science; Measure (data warehouse); Dirichlet distribution; Inference; Prior probability; Parametric statistics; Bayesian inference; Posterior probability; Mathematics; Bayesian probability; Applied mathematics; Artificial intelligence; Data mining; Statistics","score_opus":0.04890874881870901,"score_gpt":0.30447905946343445,"score_spread":0.25557031064472546,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4220966911","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007446159,0.00034628896,0.98768914,0.0005571695,0.00004737103,0.000047507845,0.00019767508,0.00012457001,0.0035439923],"genre_scores_gemma":[0.4961058,0.0012770621,0.4688516,0.0007188032,0.000316067,0.001006607,0.001578246,0.00060118595,0.029544594],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99885523,0.00054778665,0.00003966273,0.00024790934,0.00019026457,0.000119160286],"domain_scores_gemma":[0.99804,0.001472193,0.00010711102,0.000079577054,0.00020420078,0.00009703454],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0030035076,0.0010842212,0.00188884,0.00095591496,0.0005818132,0.0019953449,0.002457764,0.0024088086,0.0043637906],"category_scores_gemma":[0.006316235,0.0010629796,0.0015117755,0.0013717479,0.0017411785,0.0021903245,0.0021797763,0.002525476,0.0007825039],"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":[0.000036371,0.000027852966,0.00047571026,0.00009379716,0.000038165548,0.00008849273,0.00009638261,0.8604539,0.0006214658,0.12619315,0.0021694673,0.009705183],"study_design_scores_gemma":[0.000005873683,0.000004224346,0.000042421234,0.0000058525757,0.0000032643682,0.0000072411913,0.0000082441875,0.978869,0.000059366128,0.020447718,0.0005417023,0.0000051158427],"about_ca_topic_score_codex":0.009584678,"about_ca_topic_score_gemma":0.010556928,"teacher_disagreement_score":0.009584678,"about_ca_system_score_codex":0.0021226408,"about_ca_system_score_gemma":0.002415505,"threshold_uncertainty_score":0.01905775},"labels":[],"label_agreement":null},{"id":"W4225986943","doi":"10.1145/3502726","title":"On the Robustness of Metric Learning: An Adversarial Perspective","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"National Science Foundation","keywords":"Robustness (evolution); Pairwise comparison; Computer science; Metric (unit); Adversarial system; Artificial intelligence; Machine learning","score_opus":0.04692128029731266,"score_gpt":0.309908607807081,"score_spread":0.26298732750976833,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4225986943","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.016197387,0.0018098165,0.97773963,0.0010898353,0.00009837188,0.000048300444,0.000075489166,0.00024295271,0.0026981903],"genre_scores_gemma":[0.84618974,0.0036811016,0.14535701,0.00075116206,0.0005203259,0.0002490075,0.00037934544,0.00032906752,0.0025433693],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99106795,0.0038805353,0.00054438936,0.001660833,0.0022897187,0.00055657333],"domain_scores_gemma":[0.94816285,0.039645437,0.0036427695,0.0052188453,0.0024557456,0.00087437895],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008724856,0.0028673292,0.0018399268,0.0025226718,0.0010174714,0.0025402808,0.0023949267,0.0030511154,0.0015051443],"category_scores_gemma":[0.06184284,0.0008873204,0.0018856513,0.00169526,0.0058805165,0.0074311118,0.0054895505,0.006412732,0.00047158124],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016596234,0.00006000792,0.002101506,0.00020237306,0.00019115544,0.0001921633,0.0001684388,0.8275742,0.0033899096,0.12821996,0.0012549276,0.03647931],"study_design_scores_gemma":[0.000007899924,0.00012770835,0.00031283998,0.000047481073,0.000024443118,0.000121478755,0.000026293155,0.9165372,0.0019447431,0.079937354,0.0008845998,0.000027965758],"about_ca_topic_score_codex":0.0017328515,"about_ca_topic_score_gemma":0.00072330737,"teacher_disagreement_score":0.008724856,"about_ca_system_score_codex":0.002261467,"about_ca_system_score_gemma":0.0012186301,"threshold_uncertainty_score":0.046141982},"labels":[],"label_agreement":null},{"id":"W4281975883","doi":"10.1145/3805034","title":"Streaming Hierarchical Clustering Based on Point-Set Kernel","year":2022,"lang":"en","type":"preprint","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","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":"Artificial Intelligence in Medicine (Canada)","funders":"State Key Laboratory of Novel Software Technology; Nanjing University; National Natural Science Foundation of China","keywords":"Hierarchical clustering; Cluster analysis; Computer science; Data mining; Single-linkage clustering; Brown clustering; Tree (set theory); Consensus clustering; Correlation clustering; Hierarchical clustering of networks; CURE data clustering algorithm; Scalability; Kernel (algebra); Tree structure; Constrained clustering; Artificial intelligence; Mathematics; Algorithm; Database; Binary tree","score_opus":0.07734526681048454,"score_gpt":0.3535982704819712,"score_spread":0.27625300367148664,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4281975883","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.050538886,0.00021713287,0.9464798,0.00010702428,0.000040184048,0.00010922219,0.00019568244,0.001368327,0.0009438687],"genre_scores_gemma":[0.58594936,0.00014389881,0.4111653,0.000053462216,0.000038884056,0.00009514342,0.0008726843,0.0001427719,0.0015384282],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9982589,0.00034042078,0.00010976511,0.00037048926,0.0007787325,0.00014179459],"domain_scores_gemma":[0.995756,0.00096117775,0.00041913008,0.00091892376,0.0017053123,0.00023951083],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001455392,0.00052677875,0.0013158765,0.0027596767,0.0008972445,0.0013143646,0.0021902649,0.0007687507,0.0013163214],"category_scores_gemma":[0.006048637,0.0003819089,0.00083369366,0.0033975872,0.00070325227,0.0019222613,0.0017318833,0.0009941638,0.0006816812],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003153714,0.00017599946,0.0050422433,0.00016345392,0.00013648116,0.00016745464,0.00035411984,0.73250055,0.012197332,0.023124266,0.0045285816,0.22129416],"study_design_scores_gemma":[0.000004346829,0.000012988229,0.00030588167,0.000002004946,0.000003995419,0.000017980254,0.000014963267,0.9952342,0.0010524103,0.0030114013,0.00033286016,0.0000070317733],"about_ca_topic_score_codex":0.0141627025,"about_ca_topic_score_gemma":0.010351281,"teacher_disagreement_score":0.0141627025,"about_ca_system_score_codex":0.0018896104,"about_ca_system_score_gemma":0.0015597786,"threshold_uncertainty_score":0.028160512},"labels":[],"label_agreement":null},{"id":"W4306407742","doi":"10.1145/3563947","title":"Lifelong Online Learning from Accumulated Knowledge","year":2022,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University; Université Laval","funders":"Universidade de Macau","keywords":"Computer science; Lifelong learning; Task (project management); Artificial intelligence; Machine learning; Process (computing); Transfer of learning; Knowledge base; Online learning; Multi-task learning; Knowledge transfer; Knowledge management","score_opus":0.09332883032205538,"score_gpt":0.3227769710322922,"score_spread":0.2294481407102368,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4306407742","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020230548,0.00068559643,0.97586185,0.0005259564,0.00006025943,0.000102862374,0.00014180638,0.0008167634,0.0015742776],"genre_scores_gemma":[0.7585849,0.00065918773,0.23234119,0.0006914213,0.00024267976,0.0005127579,0.0011598711,0.00023049761,0.0055774683],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984699,0.00043395252,0.00009279959,0.0005567474,0.00028090426,0.00016560145],"domain_scores_gemma":[0.99138653,0.0059186616,0.0004128315,0.0013052697,0.00060073845,0.0003760037],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0030193846,0.0015066599,0.002116941,0.00093558006,0.00082892075,0.0017011985,0.0041653616,0.0020846077,0.0034306378],"category_scores_gemma":[0.012944406,0.0008364461,0.00075516483,0.0011146325,0.0018799495,0.0060915975,0.00402462,0.0031700446,0.0010688835],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00039200333,0.00065480324,0.0029299192,0.00046115162,0.00014703491,0.00031268463,0.0003725291,0.61096257,0.002869973,0.04257281,0.0069767185,0.33134776],"study_design_scores_gemma":[0.000019259924,0.00007267649,0.00018314201,0.000020719435,0.000013479875,0.00004522245,0.000027089476,0.94782794,0.0009966219,0.049893968,0.00088672875,0.00001312704],"about_ca_topic_score_codex":0.0022852547,"about_ca_topic_score_gemma":0.0034786589,"teacher_disagreement_score":0.0041653616,"about_ca_system_score_codex":0.0012845931,"about_ca_system_score_gemma":0.0018601656,"threshold_uncertainty_score":0.015968204},"labels":[],"label_agreement":null},{"id":"W4313437244","doi":"10.1145/3578586","title":"Summarizing User-item Matrix By Group Utility Maximization","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Greedy algorithm; Computer science; Selection (genetic algorithm); Population; Similarity (geometry); Group (periodic table); Maximization; Data mining; Preference; Cluster analysis; Information retrieval; Theoretical computer science; Artificial intelligence; Algorithm; Mathematical optimization; Mathematics; Statistics","score_opus":0.06137206547769925,"score_gpt":0.3151175045849592,"score_spread":0.25374543910725994,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4313437244","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01090069,0.00038105025,0.98655415,0.00031435402,0.000025701163,0.00020570612,0.00038644657,0.00036292794,0.00086895813],"genre_scores_gemma":[0.34217995,0.0007814487,0.6479152,0.00042351306,0.00017076916,0.0008971153,0.002666049,0.0002670297,0.004698923],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9957985,0.0021965115,0.00020678098,0.0008570653,0.00065081014,0.00029031097],"domain_scores_gemma":[0.9946431,0.0033351954,0.000382829,0.00064331235,0.0007567223,0.00023886403],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005186846,0.0021305275,0.002749487,0.0018175072,0.0007964453,0.0019820584,0.0027024958,0.0015661153,0.0027085813],"category_scores_gemma":[0.013006728,0.0009199542,0.0013874922,0.004304815,0.0009159183,0.0041545313,0.0021280786,0.0016681523,0.0011625156],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000371552,0.0003498326,0.004762113,0.00051223906,0.00041326237,0.0002564578,0.00060037465,0.7154962,0.0038186,0.044010393,0.015127037,0.21428202],"study_design_scores_gemma":[0.000025402142,0.00008724707,0.0003798124,0.000013711479,0.000049703445,0.000071714196,0.000074504416,0.96250725,0.0007816516,0.03449769,0.001491454,0.000019774983],"about_ca_topic_score_codex":0.0047648433,"about_ca_topic_score_gemma":0.0064277006,"teacher_disagreement_score":0.005186846,"about_ca_system_score_codex":0.0016181652,"about_ca_system_score_gemma":0.0018477777,"threshold_uncertainty_score":0.027430952},"labels":[],"label_agreement":null},{"id":"W4317209824","doi":"10.1145/3580367","title":"Differentially Private Release of Heterogeneous Network for Managing Healthcare Data","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; Concordia University","funders":"","keywords":"Flexibility (engineering); Computer science; Health care; Scalability; Variety (cybernetics); Data sharing; Data science; Enhanced Data Rates for GSM Evolution; Big data; Information exchange; Private information retrieval; Information sharing; Data mining; Computer security; World Wide Web; Database; Artificial intelligence","score_opus":0.09979903213826967,"score_gpt":0.33325412466265963,"score_spread":0.23345509252438995,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4317209824","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.049000107,0.00035124016,0.9465158,0.000813713,0.00009285005,0.00020904136,0.00042935228,0.0006376378,0.0019503621],"genre_scores_gemma":[0.83040917,0.00031836928,0.16461273,0.0003551002,0.00008484987,0.00019860534,0.0009307917,0.00007416061,0.003016237],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.996935,0.0012794556,0.00017163111,0.00064845034,0.000713003,0.0002525245],"domain_scores_gemma":[0.9947119,0.0017107006,0.0005585894,0.002522683,0.00030497366,0.00019120834],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036533093,0.0005139298,0.0008069944,0.0007425875,0.0012530223,0.0015324156,0.001999087,0.0011245973,0.0013949702],"category_scores_gemma":[0.009331152,0.00029698512,0.00070103107,0.0012471582,0.0012356414,0.0051188166,0.0037439032,0.0012533547,0.00036690425],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0015721207,0.00026676754,0.0055121644,0.00025871582,0.0001837162,0.0010884186,0.0010631847,0.49060217,0.02363507,0.19321302,0.013932005,0.26867265],"study_design_scores_gemma":[0.00004769675,0.00010616891,0.00050504564,0.000013064935,0.000031606905,0.00035783486,0.00013338287,0.93156946,0.0082237525,0.053033203,0.0059550283,0.000023769937],"about_ca_topic_score_codex":0.0009887405,"about_ca_topic_score_gemma":0.0010882993,"teacher_disagreement_score":0.0036533093,"about_ca_system_score_codex":0.0014556007,"about_ca_system_score_gemma":0.0011623839,"threshold_uncertainty_score":0.019320726},"labels":[],"label_agreement":null},{"id":"W4319349118","doi":"10.1145/3563040","title":"DeltaShield: Information Theory for Human- Trafficking Detection","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"National Science Foundation","keywords":"Computer science; Interpretability; Scalability; Spotting; Domain (mathematical analysis); Law enforcement; Laptop; Information retrieval; Data mining; Matching (statistics); Artificial intelligence; Database","score_opus":0.059850506837205636,"score_gpt":0.3110097719310146,"score_spread":0.2511592650938089,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4319349118","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005061099,0.002201581,0.9803991,0.0013209201,0.00017055482,0.0001679774,0.0019206641,0.0039911564,0.0047669047],"genre_scores_gemma":[0.32553342,0.003430679,0.6563412,0.0014861055,0.0009824674,0.0009423281,0.004571973,0.00083968195,0.005872179],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99573517,0.0016619229,0.00027141688,0.00070427184,0.0014301607,0.00019697793],"domain_scores_gemma":[0.97346073,0.021378148,0.00076347776,0.0023659852,0.0017444551,0.00028720428],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0063668457,0.0013359103,0.0020693757,0.007617709,0.0012058119,0.0037434443,0.0028125902,0.0021489402,0.006852128],"category_scores_gemma":[0.036940385,0.001087188,0.0017077918,0.0052634957,0.0023747063,0.007302968,0.002858155,0.0038637891,0.002458626],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00040914642,0.00026349374,0.00489338,0.0010746797,0.00033546067,0.0003338617,0.00039931806,0.21824618,0.0020545237,0.2529109,0.06048967,0.45858935],"study_design_scores_gemma":[0.00003286169,0.00004377522,0.0005168497,0.00008309272,0.00003284299,0.00010807115,0.000055282682,0.71591353,0.0012250674,0.2732481,0.008703804,0.000036658013],"about_ca_topic_score_codex":0.0071982713,"about_ca_topic_score_gemma":0.0058467477,"teacher_disagreement_score":0.007617709,"about_ca_system_score_codex":0.0034612466,"about_ca_system_score_gemma":0.003025845,"threshold_uncertainty_score":0.0336715},"labels":[],"label_agreement":null},{"id":"W4366087542","doi":"10.1145/3592857","title":"Modeling Regime Shifts in Multiple Time Series","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Time Series Analysis and Forecasting","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec à Montréal; Université de Sherbrooke","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Series (stratigraphy); Computer science; Time series; Dependency (UML); Representation (politics); Grid; Machine learning; Artificial intelligence; Data mining; Mathematics","score_opus":0.06498033503273945,"score_gpt":0.2730715737043471,"score_spread":0.20809123867160764,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4366087542","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.29315904,0.0007485248,0.7016759,0.0010454175,0.00009444123,0.000057815123,0.00036294077,0.00036817035,0.0024878595],"genre_scores_gemma":[0.95940274,0.0004564543,0.03826475,0.00008101848,0.00008177425,0.00006972969,0.00020328652,0.000032258136,0.0014079493],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9993482,0.00016236852,0.000040277777,0.00024701195,0.00012214702,0.00008002389],"domain_scores_gemma":[0.9967709,0.0020705229,0.0006998152,0.00022611218,0.00012222255,0.00011040508],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021558614,0.00053533464,0.00067890924,0.0012613619,0.00036127132,0.0009547403,0.0014666268,0.0011224,0.0011325614],"category_scores_gemma":[0.007990829,0.00033770388,0.001002479,0.0011742617,0.0007979368,0.0020829847,0.0012484655,0.0016089948,0.00016357182],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000054102893,0.000077408185,0.019154947,0.00009176744,0.00016927104,0.00028875616,0.00024733794,0.90328974,0.002350923,0.043310042,0.0004946891,0.030470984],"study_design_scores_gemma":[0.0000021833923,0.000014402451,0.0013012137,0.0000065312556,0.0000115481635,0.000020377607,0.00001680062,0.9816559,0.0001730279,0.016451778,0.0003405492,0.000005666931],"about_ca_topic_score_codex":0.0041674334,"about_ca_topic_score_gemma":0.0029466094,"teacher_disagreement_score":0.0041674334,"about_ca_system_score_codex":0.00076427485,"about_ca_system_score_gemma":0.00048376631,"threshold_uncertainty_score":0.011401415},"labels":[],"label_agreement":null},{"id":"W4386497976","doi":"10.1145/3622940","title":"Discovering Interesting Patterns from Hypergraphs","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":14,"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 Manitoba","funders":"Natural Sciences and Engineering Research Council of Canada; University of Manitoba","keywords":"Hypergraph; Data mining; Computer science; Theoretical computer science; Mathematics; Combinatorics","score_opus":0.07976721731493004,"score_gpt":0.31524831086129085,"score_spread":0.23548109354636082,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4386497976","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06337195,0.0010163671,0.9281379,0.0008289096,0.00004022895,0.00028877356,0.0018645771,0.0013126454,0.0031385696],"genre_scores_gemma":[0.31827945,0.0013394235,0.67260045,0.00031008042,0.00008197915,0.00039442757,0.004742163,0.00015578468,0.0020962572],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9981,0.0004964424,0.000148362,0.00057370245,0.0005540704,0.00012744822],"domain_scores_gemma":[0.9946021,0.0031544256,0.0005899136,0.0007431779,0.0007184056,0.00019209775],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010250942,0.00070978806,0.00065338885,0.0062330896,0.0009177239,0.002275039,0.0010383886,0.0008503133,0.0014547858],"category_scores_gemma":[0.0077720564,0.00060046,0.0011989224,0.0054555945,0.0008695932,0.003730134,0.0016165783,0.0011097953,0.0004656785],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00034786962,0.00034314528,0.037735295,0.0012007288,0.00062530476,0.003206151,0.0017516288,0.092497356,0.022596018,0.16654871,0.015421298,0.6577266],"study_design_scores_gemma":[0.00004659355,0.00008376808,0.005189066,0.00016044367,0.00021022832,0.0017137142,0.0009051104,0.5177238,0.008780741,0.44170013,0.023428211,0.000058203408],"about_ca_topic_score_codex":0.0022533746,"about_ca_topic_score_gemma":0.0046951775,"teacher_disagreement_score":0.0062330896,"about_ca_system_score_codex":0.0007589065,"about_ca_system_score_gemma":0.001293607,"threshold_uncertainty_score":0.0055062175},"labels":[],"label_agreement":null},{"id":"W4388421529","doi":"10.1145/3631609","title":"Laplacian Change Point Detection for Single and Multi-view Dynamic Graphs","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Mila - Quebec Artificial Intelligence Institute; Canadian Institute for Advanced Research; McGill University","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Computer science; Anomaly detection; Laplacian matrix; Snapshot (computer storage); Laplace operator; Embedding; Theoretical computer science; Graph; Algorithm; Data mining; Artificial intelligence; Mathematics","score_opus":0.09321852727549758,"score_gpt":0.335970799174395,"score_spread":0.24275227189889742,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4388421529","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.12550078,0.00077379454,0.8680271,0.00044566634,0.0001354472,0.00006566054,0.00065184024,0.0024467672,0.0019529397],"genre_scores_gemma":[0.84409297,0.0004435389,0.15152678,0.00019853338,0.000098956145,0.000054255095,0.0015627019,0.00021755605,0.0018046845],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992094,0.0001261431,0.000031393272,0.00029783515,0.000237987,0.00009720829],"domain_scores_gemma":[0.99791104,0.0008959554,0.00027387682,0.00034513147,0.0004184591,0.00015555613],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00081828795,0.0009655742,0.00091316,0.0029463698,0.00056648575,0.0011532826,0.0010855632,0.00090443593,0.0010846737],"category_scores_gemma":[0.0059996364,0.00033513442,0.00079009635,0.0018697833,0.00070214097,0.0019419915,0.001176037,0.0013633748,0.00059730606],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00047894154,0.00024088963,0.02462607,0.0002770113,0.00026833202,0.0009825889,0.0005321047,0.37024018,0.031176433,0.01847031,0.015462797,0.5372444],"study_design_scores_gemma":[0.000006271336,0.000028408887,0.0025764466,0.0000064777696,0.000013394642,0.00019536496,0.00010521345,0.983192,0.0027515248,0.009818552,0.001286707,0.000019632957],"about_ca_topic_score_codex":0.006202641,"about_ca_topic_score_gemma":0.008459907,"teacher_disagreement_score":0.006202641,"about_ca_system_score_codex":0.00084448187,"about_ca_system_score_gemma":0.00045747464,"threshold_uncertainty_score":0.012333095},"labels":[],"label_agreement":null},{"id":"W4391026313","doi":"10.1145/3641108","title":"Interpreting Deep Forest through Feature Contribution and MDI Feature Importance","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"China Postdoctoral Science Foundation; National Natural Science Foundation of China","keywords":"Interpretability; Categorical variable; Feature (linguistics); Random forest; Computer science; Tree (set theory); Artificial intelligence; Variety (cybernetics); Feature model; Data mining; Layer (electronics); Machine learning; Pattern recognition (psychology); Mathematics; Software","score_opus":0.031627270126612386,"score_gpt":0.3094050055243785,"score_spread":0.2777777353977661,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4391026313","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027306866,0.00026398178,0.97056097,0.00028701502,0.000031166226,0.00003579173,0.000112601294,0.0004712312,0.0009303527],"genre_scores_gemma":[0.69046575,0.00031829067,0.3067762,0.00020452945,0.0000904221,0.00011375497,0.0004979827,0.00018299469,0.0013500835],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99921,0.00024743535,0.000045722296,0.00019679032,0.00021057724,0.000089356996],"domain_scores_gemma":[0.9953935,0.002869468,0.0004247228,0.0005264149,0.00063427526,0.00015164394],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029198301,0.0008387495,0.00066014664,0.0020157862,0.0005450215,0.0013223466,0.0014594621,0.0009185467,0.0013636189],"category_scores_gemma":[0.012676605,0.00040309632,0.00060598704,0.0011769908,0.0012356556,0.0031384695,0.0018757471,0.0019158535,0.00027656788],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00023407998,0.00015659613,0.022145178,0.00025185235,0.00011284511,0.0003884978,0.0004257333,0.47741553,0.00865249,0.11481346,0.0049860235,0.37041774],"study_design_scores_gemma":[0.000007325226,0.000015815926,0.0010801809,0.000016052034,0.000013095255,0.000050504932,0.00002356094,0.9366219,0.0015986207,0.059867226,0.0006939699,0.000011774961],"about_ca_topic_score_codex":0.0035964653,"about_ca_topic_score_gemma":0.005124371,"teacher_disagreement_score":0.0035964653,"about_ca_system_score_codex":0.0010212254,"about_ca_system_score_gemma":0.00097113044,"threshold_uncertainty_score":0.015441656},"labels":[],"label_agreement":null},{"id":"W4399815117","doi":"10.1145/3673761","title":"Learning Individual Treatment Effects under Heterogeneous Interference in Networks","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"National Natural Science Foundation of China","keywords":"Weighting; Benchmark (surveying); Interference (communication); Computer science; Generalization; Regression; Spillover effect; Observational study; Range (aeronautics); Confounding; Sample (material); Process (computing); Machine learning; Artificial intelligence; Mathematics; Statistics; Medicine","score_opus":0.16936547919897604,"score_gpt":0.4090027000751428,"score_spread":0.23963722087616673,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399815117","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.029533558,0.0013491531,0.9661303,0.0012075711,0.000062416846,0.00020365388,0.00030034207,0.00033730967,0.0008756701],"genre_scores_gemma":[0.6409831,0.0018330509,0.3485887,0.001561308,0.0005474329,0.0013555092,0.0016021045,0.00016215623,0.0033666994],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.98580813,0.009180245,0.00048051385,0.0033215159,0.00082160183,0.0003879888],"domain_scores_gemma":[0.9393761,0.053543508,0.0029587066,0.0026431985,0.0010545137,0.00042399843],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02238196,0.001759375,0.003431172,0.0018728229,0.0010307226,0.0019294353,0.0041154777,0.0032085474,0.001976611],"category_scores_gemma":[0.066441864,0.001208616,0.0022521613,0.0022202018,0.002514911,0.0040796045,0.0030493161,0.004821913,0.00047225534],"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":[0.0006417006,0.00036883002,0.017248115,0.00073219603,0.0008809552,0.00025370275,0.0007145982,0.6464299,0.0010192061,0.050567094,0.0037410273,0.27740264],"study_design_scores_gemma":[0.00011366133,0.0000831752,0.0017729942,0.00007393608,0.00019814532,0.000050984887,0.00005771133,0.89086366,0.0005343805,0.104644045,0.0015845497,0.000022837366],"about_ca_topic_score_codex":0.0070678606,"about_ca_topic_score_gemma":0.00652333,"teacher_disagreement_score":0.02238196,"about_ca_system_score_codex":0.0025390114,"about_ca_system_score_gemma":0.0020513574,"threshold_uncertainty_score":0.118368566},"labels":[],"label_agreement":null},{"id":"W4408221131","doi":"10.1145/3721983","title":"SED2AM: Solving Multi-Trip Time-Dependent Vehicle Routing Problem Using Deep Reinforcement Learning","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Smart Parking Systems Research","field":"Engineering","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"National Research Council Canada; University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Reinforcement learning; Vehicle routing problem; Computer science; Routing (electronic design automation); Artificial intelligence; Reinforcement; Mathematical optimization; Machine learning; Mathematics; Engineering; Computer network","score_opus":0.04577542143170239,"score_gpt":0.30683840944866225,"score_spread":0.2610629880169599,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408221131","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.05463678,0.00087960245,0.93499494,0.00082797837,0.00015100178,0.00012404435,0.00045273433,0.0036751267,0.0042577717],"genre_scores_gemma":[0.7367107,0.00036080752,0.25540432,0.0006337798,0.00009176603,0.0002560044,0.0012935909,0.00026713908,0.0049818843],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996766,0.000099915116,0.000016503227,0.000100390105,0.00005257871,0.000054138945],"domain_scores_gemma":[0.9991961,0.0005151466,0.00006754343,0.00006887693,0.00009734188,0.000055000055],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00088493637,0.0010968957,0.0010370645,0.00049639493,0.00029190976,0.0007217231,0.0020346544,0.0012047235,0.0025009152],"category_scores_gemma":[0.0024939433,0.0005586143,0.0007831141,0.0005563481,0.000545259,0.0011540223,0.0011597312,0.002027834,0.00051097263],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000061657156,0.00012227002,0.0013983712,0.00009550885,0.00006473283,0.00006967853,0.000040725066,0.89826214,0.0009258761,0.00421136,0.00390756,0.0908401],"study_design_scores_gemma":[0.0000063650396,0.000011807811,0.0000381491,0.000002623151,0.0000033452784,0.000004432881,0.0000032781293,0.99815077,0.00017431795,0.0013843491,0.00021899532,0.0000015414594],"about_ca_topic_score_codex":0.012782683,"about_ca_topic_score_gemma":0.018830827,"teacher_disagreement_score":0.012782683,"about_ca_system_score_codex":0.0011169066,"about_ca_system_score_gemma":0.002046055,"threshold_uncertainty_score":0.025416553},"labels":[],"label_agreement":null},{"id":"W4413592351","doi":"10.1145/3762667","title":"ORIC V2: Improved Feature Interaction Detection Model through Online Random Interaction Chains for Click-Through Rate Prediction","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Knowledge Discovery from Data","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Petro-Canada","funders":"National Natural Science Foundation of China","keywords":"Feature (linguistics); Random forest; Computer science; Pattern recognition (psychology); Artificial intelligence; Data mining","score_opus":0.0650397683175733,"score_gpt":0.3459675967337768,"score_spread":0.2809278284162035,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413592351","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.045692097,0.0010214625,0.94513786,0.0004412373,0.00013563207,0.00028749698,0.0012616939,0.003607277,0.0024153565],"genre_scores_gemma":[0.7348792,0.0007705444,0.24983041,0.0005475893,0.0002578443,0.0006070639,0.004353062,0.00028020705,0.008474082],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9985366,0.00031285654,0.0000912771,0.0004837599,0.00039268518,0.000182817],"domain_scores_gemma":[0.99533725,0.0025664216,0.00033848142,0.0005896179,0.00095170527,0.00021645884],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029791102,0.0014766031,0.0021702969,0.0019634594,0.00059471175,0.0015150135,0.00409313,0.001660962,0.0027967535],"category_scores_gemma":[0.008065027,0.00086846086,0.001390595,0.001962289,0.00064935465,0.0020509202,0.0014071875,0.0026651237,0.0011835081],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005871164,0.00046114775,0.014827109,0.00019485354,0.0002654794,0.00024523627,0.00017471024,0.76668876,0.0035040607,0.0108440025,0.008750154,0.19345732],"study_design_scores_gemma":[0.00000874421,0.00002117412,0.00021472185,0.0000033747124,0.000008839348,0.000014895768,0.0000025138359,0.9983448,0.00017676642,0.0009481709,0.00024874628,0.0000072993475],"about_ca_topic_score_codex":0.027339257,"about_ca_topic_score_gemma":0.029151553,"teacher_disagreement_score":0.027339257,"about_ca_system_score_codex":0.0012634203,"about_ca_system_score_gemma":0.0021363718,"threshold_uncertainty_score":0.05436021},"labels":[],"label_agreement":null}]}