{"meta":{"query_hash":"84a422687fe3","filters":{"venue":"ACM SIGKDD Explorations Newsletter"},"cohort_total":40,"direct_labels_cover":0,"predictions_cover":40,"exported":40,"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/84a422687fe3","api":"https://metacan.xera.ac/api/v1/cohort?venue=ACM+SIGKDD+Explorations+Newsletter"},"results":[{"id":"W1965269329","doi":"10.1145/1147234.1147235","title":"Introduction to the special issue on successful real-world data mining applications","year":2006,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","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 Alberta","funders":"","keywords":"Computer science; Data science; Field (mathematics); Database transaction; Transaction data; Knowledge extraction; Data discovery; Data stream mining; Data mining; World Wide Web; Database; Metadata","score_opus":0.03975452893360564,"score_gpt":0.2972189779695197,"score_spread":0.25746444903591403,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1965269329","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0010028811,0.07082165,0.020917777,0.07909087,0.6991462,0.0007269252,0.0038137913,0.002623769,0.12185616],"genre_scores_gemma":[0.003426979,0.053577326,0.01030057,0.021508593,0.6522149,0.00039569067,0.0047387723,0.0019400266,0.25189713],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9961204,0.00051231595,0.00039657496,0.0005080903,0.0022812502,0.00018130882],"domain_scores_gemma":[0.9813572,0.007568348,0.0011096868,0.001450196,0.0059007704,0.0026137382],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0040380945,0.0021474457,0.0021638903,0.005425847,0.0018374001,0.008494624,0.0019207598,0.0031195604,0.13330865],"category_scores_gemma":[0.013599206,0.00077248266,0.0013640819,0.005363408,0.001058865,0.007381915,0.0032153165,0.0038627011,0.099630095],"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":[0.000009217181,0.000022091688,0.00006741856,0.00016778421,0.000006405303,0.000028233046,0.000014487005,0.000044565277,0.000114506765,0.00060163013,0.9650856,0.033838045],"study_design_scores_gemma":[0.0000059943054,0.000031021755,0.00024642,0.0001570598,0.0000086548225,0.00013688994,0.000024440526,0.0001348304,0.00008125543,0.0015457697,0.9976192,0.000008440879],"about_ca_topic_score_codex":0.0003454486,"about_ca_topic_score_gemma":0.0010807918,"teacher_disagreement_score":0.13330865,"about_ca_system_score_codex":0.0009681179,"about_ca_system_score_gemma":0.0015584246,"threshold_uncertainty_score":0.44596183},"labels":[],"label_agreement":null},{"id":"W1967519701","doi":"10.1145/772862.772885","title":"KDD-2002 workshop report fractals and self-similarity in data mining","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"Carnegie Mellon University; University of Southern California; National Science Foundation","keywords":"Computer science; Similarity (geometry); Data mining; Conjunction (astronomy); Fractal; Data science; Self-similarity; Information retrieval; Artificial intelligence; Mathematics","score_opus":0.11892294112044009,"score_gpt":0.29976370573021005,"score_spread":0.18084076460976994,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1967519701","genre_codex":"methods","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":[0.016856601,0.08913877,0.69750744,0.055875923,0.050355807,0.0013037137,0.014668021,0.007713255,0.06658049],"genre_scores_gemma":[0.12121659,0.08215702,0.5237071,0.010171836,0.025831264,0.0019908429,0.0760776,0.00411619,0.15473153],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99158937,0.0022769845,0.000800215,0.001079015,0.0038880587,0.00036634732],"domain_scores_gemma":[0.9846577,0.0046572783,0.00036120467,0.0022230598,0.006263889,0.0018369063],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012742248,0.0012202007,0.0023062946,0.0042950152,0.0011616439,0.0067951726,0.0025146322,0.002629062,0.009911507],"category_scores_gemma":[0.023557842,0.00080808974,0.0012766983,0.0042725815,0.0008576661,0.004709425,0.0030331525,0.003625528,0.0090189325],"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":[0.00027283392,0.00025176382,0.0015000573,0.0008629645,0.0001589015,0.00024946773,0.00020425298,0.0057716356,0.0013645027,0.0208231,0.68995935,0.2785811],"study_design_scores_gemma":[0.00019414927,0.00022845154,0.0021313988,0.00047622068,0.00013743658,0.00060725916,0.00026906878,0.04328719,0.0063844896,0.050715994,0.8954658,0.00010253462],"about_ca_topic_score_codex":0.003146335,"about_ca_topic_score_gemma":0.0023301002,"teacher_disagreement_score":0.012742248,"about_ca_system_score_codex":0.001703676,"about_ca_system_score_gemma":0.0038935777,"threshold_uncertainty_score":0.067388296},"labels":[],"label_agreement":null},{"id":"W1968064480","doi":"10.1145/2031331.2031336","title":"Trajectory anonymity in publishing personal mobility data","year":2011,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":68,"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":"Computer science; Data publishing; Identifier; Anonymity; Global Positioning System; TRACE (psycholinguistics); Location-based service; Computer security; Object (grammar); World Wide Web; GSM; Mobile device; Location data; Personally identifiable information; k-anonymity; Internet privacy; Data science; Publishing; Computer network; Telecommunications","score_opus":0.2671506687550872,"score_gpt":0.34191459208467995,"score_spread":0.07476392332959275,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1968064480","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":[0.0414396,0.0027669019,0.9369996,0.0039232983,0.0004612191,0.0002825004,0.0009068308,0.0012751489,0.011944869],"genre_scores_gemma":[0.7040697,0.0055705407,0.27556422,0.00084931497,0.0012112539,0.0005882494,0.0016960705,0.0003486509,0.010102013],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.97104746,0.013636236,0.0032502695,0.0023148,0.008446764,0.0013045708],"domain_scores_gemma":[0.89482456,0.0371167,0.0078095626,0.05078919,0.008062039,0.0013979825],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.018219326,0.00064904656,0.0018330093,0.0032217675,0.0038894096,0.008955074,0.0036970782,0.0023589188,0.0022282624],"category_scores_gemma":[0.06870109,0.0009008547,0.001336283,0.009061621,0.0039360207,0.020514237,0.008629141,0.0032200224,0.0015558848],"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":[0.0010220194,0.0001999134,0.011635993,0.0008084534,0.00021325042,0.0008957203,0.0049317596,0.04349284,0.0062875017,0.6039342,0.010921743,0.31565666],"study_design_scores_gemma":[0.0001293124,0.0004723945,0.0029199712,0.00048530253,0.00028756593,0.003940495,0.0038513455,0.24166448,0.049481813,0.55631596,0.14019208,0.00025936222],"about_ca_topic_score_codex":0.0013289725,"about_ca_topic_score_gemma":0.0007051249,"teacher_disagreement_score":0.018219326,"about_ca_system_score_codex":0.0023928143,"about_ca_system_score_gemma":0.0035126368,"threshold_uncertainty_score":0.09635413},"labels":[],"label_agreement":null},{"id":"W1972132022","doi":"10.1145/1046456.1046468","title":"Subspace clustering for high dimensional categorical data","year":2004,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":73,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Cluster analysis; Computer science; Clustering high-dimensional data; Categorical variable; Correlation clustering; CURE data clustering algorithm; Data mining; Subspace topology; Canopy clustering algorithm; Data stream clustering; Focus (optics); Algorithm; Pattern recognition (psychology); Artificial intelligence; Machine learning","score_opus":0.11230554248111278,"score_gpt":0.34489126193553477,"score_spread":0.23258571945442197,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1972132022","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":[0.0029771768,0.00018914425,0.99621105,0.00008878906,0.000013166352,0.000021611699,0.000062117135,0.00019685042,0.0002400615],"genre_scores_gemma":[0.08437975,0.00045641002,0.9132391,0.000067568246,0.000060282993,0.00015289032,0.00065029063,0.00010090363,0.0008928911],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9958442,0.0018400035,0.00022451645,0.0005922827,0.0013391597,0.00015977501],"domain_scores_gemma":[0.9945497,0.0025171295,0.00042908706,0.0009376222,0.0013997749,0.00016661272],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036670242,0.0007325431,0.0013952,0.0031338092,0.0016159528,0.0015248206,0.0013621288,0.001086037,0.0013659481],"category_scores_gemma":[0.011253672,0.00035343794,0.0013862362,0.0048139407,0.0013641313,0.0019349727,0.0018699398,0.0015861645,0.000713491],"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":[0.00015771468,0.00009320082,0.0037724364,0.0005072683,0.00017550368,0.00023284531,0.0008529113,0.3639867,0.009032694,0.17664865,0.0070194677,0.43752062],"study_design_scores_gemma":[0.000009494041,0.000051233143,0.0009633853,0.00002273448,0.00001552045,0.00015240135,0.00015880106,0.88820785,0.0031883654,0.101581015,0.0056013856,0.00004775618],"about_ca_topic_score_codex":0.0037131135,"about_ca_topic_score_gemma":0.0043203845,"teacher_disagreement_score":0.0037131135,"about_ca_system_score_codex":0.0011755482,"about_ca_system_score_gemma":0.0017140458,"threshold_uncertainty_score":0.019393325},"labels":[],"label_agreement":null},{"id":"W1977843657","doi":"10.1145/1540276.1540279","title":"A brief survey on anonymization techniques for privacy preserving publishing of social network data","year":2008,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":419,"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":"Data publishing; Computer science; Publishing; Social network (sociolinguistics); Information privacy; Data science; k-anonymity; Data anonymization; Graph; Cluster analysis; Internet privacy; Privacy protection; World Wide Web; Data mining; Social media; Artificial intelligence; Theoretical computer science","score_opus":0.1723603973409118,"score_gpt":0.3270708739991915,"score_spread":0.1547104766582797,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1977843657","genre_codex":"methods","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0063107517,0.09704939,0.87755364,0.0030329435,0.0006341385,0.0006919824,0.0005540813,0.00087068655,0.013302378],"genre_scores_gemma":[0.10142682,0.22054967,0.66250414,0.001674603,0.0028738873,0.0013621231,0.002047841,0.00044265355,0.0071182535],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98286533,0.0075976374,0.0016815481,0.0015403766,0.00576733,0.0005477707],"domain_scores_gemma":[0.97832257,0.013282135,0.0011990333,0.0051101525,0.0019178063,0.00016838186],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009362618,0.0011268074,0.0021312402,0.0060283,0.0021827351,0.0045131077,0.0027611244,0.0021826325,0.0032349788],"category_scores_gemma":[0.019593405,0.00095821463,0.0028559284,0.011674646,0.0020453318,0.00962874,0.0025885885,0.0029266125,0.0016468524],"study_design_candidate":"not_applicable","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":[0.00016626657,0.00021863126,0.0020408558,0.0047551817,0.0003407922,0.00042196372,0.0012156883,0.016164662,0.0045339875,0.1776831,0.020482048,0.7719767],"study_design_scores_gemma":[0.00012735784,0.00060075504,0.0038852915,0.0031690479,0.00061943656,0.009467627,0.0015013828,0.10102561,0.03207237,0.21282972,0.63431185,0.00038963245],"about_ca_topic_score_codex":0.0010806358,"about_ca_topic_score_gemma":0.0007225778,"teacher_disagreement_score":0.009362618,"about_ca_system_score_codex":0.0014534752,"about_ca_system_score_gemma":0.0019724446,"threshold_uncertainty_score":0.04951489},"labels":[],"label_agreement":null},{"id":"W1978091735","doi":"10.1145/1809400.1809419","title":"Summary of the first ACM SIGKDD workshop on knowledge discovery from uncertain data (U'09)","year":2010,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","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":"Simon Fraser University","funders":"","keywords":"Computer science; Data science; Knowledge extraction; Information retrieval; Data mining","score_opus":0.07323755533522192,"score_gpt":0.30366766048651916,"score_spread":0.23043010515129725,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1978091735","genre_codex":"editorial","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.006606672,0.24069871,0.11929248,0.088022,0.30168122,0.0019243228,0.052067343,0.0080090845,0.18169816],"genre_scores_gemma":[0.02243642,0.16239826,0.04778581,0.010394713,0.05506768,0.0009781147,0.10775742,0.003716509,0.5894651],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9980003,0.0004329083,0.00016233599,0.00031334077,0.000924507,0.00016643816],"domain_scores_gemma":[0.988488,0.0019237124,0.00026183593,0.0008106786,0.006671367,0.0018444248],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051959474,0.001423151,0.0017567496,0.00496995,0.0014150417,0.0055371965,0.001864083,0.0013744837,0.09000779],"category_scores_gemma":[0.008758026,0.00060887216,0.0012360546,0.005481289,0.0005320875,0.003903991,0.0021497633,0.0021303773,0.05761694],"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":[0.000103163795,0.00004323981,0.00013881185,0.00028874728,0.000015544509,0.000032124914,0.000022972943,0.00023726551,0.00025589255,0.00060574524,0.9040582,0.09419821],"study_design_scores_gemma":[0.000029615949,0.00007976595,0.0009775781,0.0003188518,0.000041656273,0.00008250884,0.000064265856,0.000820344,0.00089826906,0.002064766,0.9945912,0.000031192532],"about_ca_topic_score_codex":0.011372737,"about_ca_topic_score_gemma":0.022772761,"teacher_disagreement_score":0.09000779,"about_ca_system_score_codex":0.0018894091,"about_ca_system_score_gemma":0.0030574736,"threshold_uncertainty_score":0.30110604},"labels":[],"label_agreement":null},{"id":"W1983506100","doi":"10.1145/568574.568581","title":"Exploiting succinct constraints using FP-trees","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":76,"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":"Computer science; Pruning; Focus (optics); Succinctness; Set (abstract data type); Constraint (computer-aided design); Tree (set theory); Data mining; Search tree; Theoretical computer science; Algorithm; Search algorithm; Programming language; Mathematics","score_opus":0.10287589201954778,"score_gpt":0.27695920445099,"score_spread":0.1740833124314422,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1983506100","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":[0.00737023,0.0003073197,0.989903,0.0002476,0.000029363111,0.00013334987,0.0003866487,0.0008555032,0.0007670429],"genre_scores_gemma":[0.08521721,0.0004970855,0.9115506,0.00016320622,0.000059341175,0.00029267807,0.0015436935,0.00012857986,0.0005476437],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99404573,0.001790674,0.00070102006,0.00069201563,0.0024791036,0.000291416],"domain_scores_gemma":[0.9775859,0.016517643,0.00151171,0.0017826247,0.0023685659,0.00023344714],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0046906765,0.0016737212,0.001810835,0.0043124147,0.0012742863,0.0028013894,0.0024691583,0.0016757279,0.0023379715],"category_scores_gemma":[0.031805936,0.00094399066,0.0015978158,0.005396755,0.0010135904,0.0067943516,0.002222359,0.0018919448,0.0010084267],"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":[0.000500514,0.00021639184,0.0050023748,0.00072677236,0.00021960548,0.0012306005,0.00047235077,0.1828321,0.006213338,0.07130087,0.009889776,0.7213953],"study_design_scores_gemma":[0.000070732865,0.00009748056,0.00033253443,0.00012705462,0.00006486352,0.0007152213,0.00010772121,0.8250488,0.007830206,0.15744247,0.0081114825,0.000051313415],"about_ca_topic_score_codex":0.0035741904,"about_ca_topic_score_gemma":0.0035394735,"teacher_disagreement_score":0.0046906765,"about_ca_system_score_codex":0.0008563449,"about_ca_system_score_gemma":0.0024009713,"threshold_uncertainty_score":0.024806976},"labels":[],"label_agreement":null},{"id":"W1984606279","doi":"10.1145/380995.381002","title":"Mining frequent patterns by pattern-growth","year":2000,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":275,"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":"Citation; Computer science; Data science; Library science","score_opus":0.021176622480086907,"score_gpt":0.24134360734380278,"score_spread":0.2201669848637159,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1984606279","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":[0.025043061,0.0016388288,0.9662091,0.00079220696,0.000093214425,0.0005261056,0.0015344524,0.0019213783,0.0022416462],"genre_scores_gemma":[0.13158353,0.0015726336,0.86049795,0.00018635526,0.00009400722,0.0005125562,0.004078292,0.00015621894,0.0013184953],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9958681,0.0009225828,0.00042128892,0.00073894486,0.0018359311,0.00021314566],"domain_scores_gemma":[0.98811394,0.006898399,0.0011147555,0.0013944535,0.0022594305,0.00021910683],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003352601,0.0010965745,0.0013067718,0.006404358,0.0008679016,0.0022410366,0.001899173,0.0010796053,0.0012786326],"category_scores_gemma":[0.018150112,0.00073991134,0.0018015167,0.007458903,0.0009778626,0.004002722,0.002000979,0.0016957959,0.0014638598],"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":[0.0002769898,0.00032199966,0.029028017,0.0011110074,0.00034141622,0.0015167544,0.00076330645,0.0817604,0.013803352,0.041906968,0.014282743,0.814887],"study_design_scores_gemma":[0.00005029027,0.0001738079,0.004046618,0.0001739515,0.000097859105,0.0017077613,0.00042963584,0.8662623,0.010301918,0.09585565,0.020836938,0.00006329585],"about_ca_topic_score_codex":0.0017568693,"about_ca_topic_score_gemma":0.00199644,"teacher_disagreement_score":0.006404358,"about_ca_system_score_codex":0.0006866153,"about_ca_system_score_gemma":0.0011202593,"threshold_uncertainty_score":0.017730415},"labels":[],"label_agreement":null},{"id":"W1993415883","doi":"10.1145/1809400.1809407","title":"FpVAT","year":2010,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":68,"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 Manitoba","funders":"","keywords":"Computer science; Process (computing); Knowledge extraction; Representation (politics); Data mining; Simple (philosophy); Visualization; Information retrieval; Visual analytics; Machine learning; Artificial intelligence","score_opus":0.03376557419937707,"score_gpt":0.29607554267871694,"score_spread":0.26230996847933985,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1993415883","genre_codex":"methods","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":[0.0054216413,0.0010570665,0.62600404,0.0012168635,0.0010258115,0.0005327042,0.017219016,0.2744164,0.073106475],"genre_scores_gemma":[0.10673681,0.0018258172,0.70534974,0.0015999341,0.0004968328,0.0020322457,0.060736876,0.051761888,0.06945979],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9982285,0.00026584158,0.0001456486,0.00035897686,0.0007840805,0.00021701494],"domain_scores_gemma":[0.99673307,0.0012000138,0.00016317658,0.000835768,0.0008630971,0.00020490715],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017968187,0.0014880224,0.00084176,0.0026379002,0.0009174957,0.0042328835,0.0038643747,0.0018037801,0.074329376],"category_scores_gemma":[0.011804146,0.0008645819,0.0019271281,0.0020361508,0.0006863873,0.005178381,0.0038953018,0.002279335,0.041151423],"study_design_candidate":"not_applicable","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":[0.000739422,0.00013441965,0.0017153106,0.0007669899,0.000105141065,0.00060591585,0.0005423429,0.004749477,0.005261746,0.041614182,0.45699567,0.4867694],"study_design_scores_gemma":[0.00020570686,0.00014084195,0.0010606463,0.000259526,0.00006447329,0.00168928,0.0001557069,0.05405937,0.0116186645,0.07385423,0.85676026,0.00013122881],"about_ca_topic_score_codex":0.002652686,"about_ca_topic_score_gemma":0.0019668627,"teacher_disagreement_score":0.074329376,"about_ca_system_score_codex":0.00078187144,"about_ca_system_score_gemma":0.001272172,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W1997102766","doi":"10.1145/1882471.1882478","title":"A brief survey on sequence classification","year":2010,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":552,"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; Sequence (biology); Feature selection; Artificial intelligence; Feature vector; Task (project management); Feature (linguistics); Machine learning; Pattern recognition (psychology); One-class classification; Support vector machine; Data mining","score_opus":0.11534281954831148,"score_gpt":0.3164656778829131,"score_spread":0.20112285833460158,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1997102766","genre_codex":"methods","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007639334,0.4490373,0.4988672,0.0047662444,0.0045320923,0.00042991972,0.0024528478,0.0024000013,0.029875135],"genre_scores_gemma":[0.0650301,0.57276493,0.31006396,0.0029975777,0.01037024,0.00084077806,0.0122805955,0.00058977836,0.025062056],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99731183,0.0005059847,0.00030691308,0.0004640861,0.0012614591,0.00014978861],"domain_scores_gemma":[0.9958704,0.0019915982,0.00022092533,0.00041839384,0.0013665279,0.00013205296],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021274185,0.0011828121,0.002000881,0.007935201,0.0009974279,0.0029176234,0.0019952045,0.001630756,0.007648352],"category_scores_gemma":[0.007987824,0.00046307655,0.0012903546,0.015224217,0.0006779229,0.005571523,0.0012108363,0.0016692892,0.007023884],"study_design_candidate":"not_applicable","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":[0.00006768613,0.00006738979,0.0012495228,0.0010722763,0.000043310654,0.00006980303,0.000060388236,0.0037803745,0.0008701941,0.009171241,0.029544277,0.9540035],"study_design_scores_gemma":[0.000041220737,0.00034057643,0.0048810244,0.0017644265,0.00012143812,0.0019636275,0.00031133002,0.10993775,0.0054537226,0.09812224,0.77692693,0.00013581268],"about_ca_topic_score_codex":0.0036233652,"about_ca_topic_score_gemma":0.0020263956,"teacher_disagreement_score":0.007935201,"about_ca_system_score_codex":0.0010716435,"about_ca_system_score_gemma":0.0019881253,"threshold_uncertainty_score":0.025586247},"labels":[],"label_agreement":null},{"id":"W2011376672","doi":"10.1145/1007730.1007737","title":"Class imbalances versus small disjuncts","year":2004,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":681,"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; Class (philosophy); Focus (optics); Classifier (UML); Artificial intelligence; Machine learning","score_opus":0.068174256664078,"score_gpt":0.27984716191671505,"score_spread":0.21167290525263704,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2011376672","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":[0.36870503,0.0046552797,0.5936243,0.0073585566,0.0009433451,0.00071733753,0.0010579459,0.0019117009,0.021026492],"genre_scores_gemma":[0.8322751,0.0009214469,0.16078272,0.0008417431,0.0006566821,0.00028809416,0.00073556905,0.00021509724,0.0032835419],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9904151,0.0029244546,0.00077084574,0.0022462148,0.0032189472,0.00042437296],"domain_scores_gemma":[0.95879877,0.031767804,0.00329871,0.0036173624,0.0017112487,0.00080603996],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0092356,0.0008459234,0.0013939284,0.0015737632,0.0013205374,0.003126877,0.0012401928,0.0015657701,0.0034872266],"category_scores_gemma":[0.040353566,0.00035542558,0.00053087855,0.0020809239,0.0024763802,0.0052493415,0.0025416163,0.0030038343,0.00094696606],"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":[0.0030045998,0.0007708713,0.034990087,0.0009170283,0.0002916762,0.0005246293,0.0009155563,0.042836174,0.03376814,0.045408044,0.019129766,0.8174434],"study_design_scores_gemma":[0.00041665617,0.0023447475,0.033157986,0.00027935492,0.00029368227,0.0023193117,0.0015828132,0.5031462,0.07975813,0.33451673,0.042007945,0.00017641777],"about_ca_topic_score_codex":0.0003409783,"about_ca_topic_score_gemma":0.0005857431,"teacher_disagreement_score":0.0092356,"about_ca_system_score_codex":0.000862524,"about_ca_system_score_gemma":0.00058546045,"threshold_uncertainty_score":0.048843145},"labels":[],"label_agreement":null},{"id":"W2024539262","doi":"10.1145/1294301.1294307","title":"Novel approaches for small biomolecule classification and structural similarity search","year":2007,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Computational Drug Discovery Methods","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":"Simon Fraser University","funders":"","keywords":"Computer science; Nearest neighbor search; Classifier (UML); Minkowski distance; Pruning; Similarity (geometry); Artificial intelligence; Data mining; Pattern recognition (psychology); Binary classification; k-nearest neighbors algorithm; Machine learning; Euclidean distance; Support vector machine","score_opus":0.2537296575736704,"score_gpt":0.3552435413062274,"score_spread":0.10151388373255699,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2024539262","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":[0.0037744886,0.0006022376,0.9937596,0.00009531013,0.000044543976,0.00005871624,0.000055851808,0.00041778525,0.0011916353],"genre_scores_gemma":[0.056849893,0.000765164,0.9381091,0.00015566064,0.000107559026,0.00025106367,0.00044228393,0.00011358687,0.0032056503],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.998718,0.00014915547,0.00010901089,0.00028687614,0.0006678045,0.00006914711],"domain_scores_gemma":[0.9990526,0.00030240236,0.00013315433,0.00018904608,0.00028163916,0.000041102034],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000973081,0.00075914955,0.0016257532,0.0022927271,0.0005619989,0.0015202074,0.0021781537,0.0013870238,0.002425392],"category_scores_gemma":[0.002324812,0.00048434973,0.0012638561,0.002636962,0.0009136142,0.0020643058,0.0013960727,0.0013482197,0.0013729739],"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":[0.00013868067,0.00023538082,0.0012146876,0.00053881545,0.00011491759,0.00014291854,0.000120031385,0.12851049,0.03283333,0.09315337,0.004377791,0.73861957],"study_design_scores_gemma":[0.000021178015,0.000068330744,0.00044646653,0.000029013134,0.000022499311,0.00014689834,0.000027979519,0.92662305,0.008486012,0.052995913,0.011110115,0.000022594548],"about_ca_topic_score_codex":0.0015474344,"about_ca_topic_score_gemma":0.0017881214,"teacher_disagreement_score":0.002425392,"about_ca_system_score_codex":0.0010410789,"about_ca_system_score_gemma":0.0010634231,"threshold_uncertainty_score":0.0081138015},"labels":[],"label_agreement":null},{"id":"W2028572284","doi":"10.1145/846183.846208","title":"Discovering geographic knowledge in data rich environments","year":2000,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","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":"Simon Fraser University","funders":"City, University of London; University of Utah; University of Florida; Florida State University","keywords":"Data science; Identification (biology); Computer science; Geographic information system; Knowledge extraction; Data mining; Geography; Cartography","score_opus":0.05390222068445348,"score_gpt":0.2862366298714703,"score_spread":0.2323344091870168,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2028572284","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":[0.21038339,0.014947167,0.67418057,0.021293798,0.0017261382,0.00054514717,0.03311635,0.004486203,0.039321225],"genre_scores_gemma":[0.5976043,0.0103276735,0.34288597,0.0013835528,0.00041132295,0.00025706962,0.03137835,0.00021700078,0.015534799],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9989304,0.00027743986,0.00006953018,0.00026229178,0.0003460106,0.0001142323],"domain_scores_gemma":[0.9979961,0.0011692488,0.0001426311,0.00028388205,0.00029010602,0.000118122836],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011998524,0.00039760215,0.000814627,0.0032485072,0.0014871774,0.003877477,0.0010756904,0.001293856,0.0042831516],"category_scores_gemma":[0.006962259,0.0005834575,0.00066494936,0.0049819555,0.0013361851,0.005466469,0.0046564555,0.001232277,0.0014128118],"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":[0.0005830548,0.00021164995,0.020178815,0.0013621047,0.00021731223,0.0031682295,0.00504635,0.08456948,0.004602861,0.19400726,0.09891484,0.58713806],"study_design_scores_gemma":[0.0000683147,0.000101797734,0.009708973,0.00041988443,0.00012855094,0.0009594487,0.0071066306,0.12153248,0.006282609,0.26859343,0.58495563,0.00014233634],"about_ca_topic_score_codex":0.004602766,"about_ca_topic_score_gemma":0.013865982,"teacher_disagreement_score":0.004602766,"about_ca_system_score_codex":0.0010814989,"about_ca_system_score_gemma":0.0008717801,"threshold_uncertainty_score":0.014328599},"labels":[],"label_agreement":null},{"id":"W2034607275","doi":"10.1145/1046456.1046459","title":"Learning important models for web page blocks based on layout and content analysis","year":2004,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":40,"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; Web page; Block (permutation group theory); Information retrieval; Partition (number theory); Page view; Static web page; Data mining; World Wide Web; Web navigation","score_opus":0.07156513468895788,"score_gpt":0.263723749372729,"score_spread":0.19215861468377116,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2034607275","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":[0.14887087,0.0007031263,0.8453839,0.00039213494,0.000062891806,0.0002692192,0.00056047045,0.0026848374,0.001072639],"genre_scores_gemma":[0.7508843,0.00048557977,0.24209806,0.00017084298,0.00015111576,0.0004258105,0.0022813198,0.0002894324,0.0032135905],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991167,0.00024342508,0.000068788024,0.0002612631,0.00020048211,0.000109278975],"domain_scores_gemma":[0.9950523,0.0032721024,0.0004518704,0.00038024716,0.0006836073,0.00015994116],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00182862,0.0011691232,0.0013060028,0.0045896354,0.00045901287,0.0014242451,0.001739314,0.0017638055,0.0014744546],"category_scores_gemma":[0.00860346,0.0008015702,0.0018850367,0.0016802186,0.0008838507,0.0031087585,0.00073388894,0.0020003573,0.00093794486],"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":[0.0010161531,0.00084581244,0.027305206,0.00031486584,0.00027841216,0.00034441962,0.0004048986,0.621131,0.012515796,0.010281082,0.0058882185,0.31967425],"study_design_scores_gemma":[0.00000810925,0.000022406104,0.00080966094,0.0000062870913,0.000014915551,0.0000302837,0.000010083996,0.9954496,0.0007566248,0.0027009079,0.00018595137,0.0000051776456],"about_ca_topic_score_codex":0.0067615258,"about_ca_topic_score_gemma":0.011957768,"teacher_disagreement_score":0.0067615258,"about_ca_system_score_codex":0.0015955915,"about_ca_system_score_gemma":0.00075548736,"threshold_uncertainty_score":0.0134443045},"labels":[],"label_agreement":null},{"id":"W2039552226","doi":"10.1145/2207243.2207248","title":"Mapping question items to skills with non-negative matrix factorization","year":2012,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Intelligent Tutoring Systems and Adaptive Learning","field":"Computer Science","cited_by":67,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Computer science; Non-negative matrix factorization; Matrix decomposition; Task (project management); Variance (accounting); Machine learning; Process (computing); Matrix (chemical analysis); Artificial intelligence; Automation; Data science","score_opus":0.02240098912766123,"score_gpt":0.27459348238239056,"score_spread":0.2521924932547293,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2039552226","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":[0.029262267,0.00017839941,0.968937,0.00023435226,0.0000563459,0.00025772362,0.00030756943,0.00030456937,0.00046181722],"genre_scores_gemma":[0.31106067,0.00021522157,0.68468523,0.00019653155,0.00009333855,0.0009674554,0.0015204214,0.0000926575,0.001168454],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9954235,0.00244532,0.00021505798,0.0012004023,0.00049196096,0.000223837],"domain_scores_gemma":[0.97267187,0.022991782,0.0012622246,0.0014710798,0.0014030879,0.00019987546],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007474936,0.0020607577,0.0010113235,0.001906229,0.0005919309,0.0014675362,0.0011451003,0.0018587232,0.0028155593],"category_scores_gemma":[0.039290555,0.00076329557,0.0024575407,0.0017159864,0.0014187517,0.0015546356,0.001183552,0.0030083105,0.0010695847],"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":[0.0006203665,0.000852518,0.017332463,0.0008651331,0.0005994539,0.0004034456,0.0012864692,0.46054557,0.010216353,0.022352714,0.006480274,0.47844517],"study_design_scores_gemma":[0.000055826524,0.00013712389,0.0053745937,0.000041569878,0.00003078984,0.00007116574,0.00008927819,0.9586881,0.001391768,0.032501314,0.0015761986,0.00004229552],"about_ca_topic_score_codex":0.0090944385,"about_ca_topic_score_gemma":0.00693582,"teacher_disagreement_score":0.0090944385,"about_ca_system_score_codex":0.0011174068,"about_ca_system_score_gemma":0.0015517923,"threshold_uncertainty_score":0.03953165},"labels":[],"label_agreement":null},{"id":"W2039760999","doi":"10.1145/380995.381054","title":"Report on MDM/KDD2000","year":2000,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Image Retrieval and Classification Techniques","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 Alberta","funders":"","keywords":"Citation; Library science; Computer science; World Wide Web","score_opus":0.03370207834165126,"score_gpt":0.2762263369441018,"score_spread":0.24252425860245055,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2039760999","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010652478,0.021423351,0.09096522,0.04937772,0.053673018,0.0035157918,0.27506557,0.03650503,0.4588218],"genre_scores_gemma":[0.01625372,0.011062135,0.06689351,0.005005975,0.007335693,0.001424988,0.41987786,0.0070381393,0.46510795],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9935681,0.0007583573,0.0003978781,0.0008117711,0.004075146,0.00038882956],"domain_scores_gemma":[0.9820747,0.0016550696,0.00031627165,0.0036969397,0.01015488,0.0021021478],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008331763,0.0017514762,0.0015267623,0.0068366155,0.0015766165,0.0059898756,0.0031335887,0.0017572364,0.15306011],"category_scores_gemma":[0.018949943,0.000706418,0.0009306782,0.005960707,0.00048019906,0.004888448,0.00358117,0.0020048285,0.18904181],"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":[0.00012975463,0.00010038639,0.00025171132,0.00010367083,0.000014027779,0.000030271784,0.000018717257,0.0002022399,0.00047849806,0.0016164464,0.9182311,0.07882307],"study_design_scores_gemma":[0.000049067115,0.000041902862,0.0006857741,0.000058026417,0.000012926162,0.000052341573,0.000033749235,0.0008921779,0.0016890172,0.001654089,0.9948178,0.000013015473],"about_ca_topic_score_codex":0.009454688,"about_ca_topic_score_gemma":0.008069629,"teacher_disagreement_score":0.15306011,"about_ca_system_score_codex":0.0021839319,"about_ca_system_score_gemma":0.003771511,"threshold_uncertainty_score":0.51203704},"labels":[],"label_agreement":null},{"id":"W2040010062","doi":"10.1145/1007730.1007736","title":"Learning from imbalanced data sets with boosting and data generation","year":2004,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":606,"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":"Boosting (machine learning); Machine learning; Computer science; Artificial intelligence; Benchmarking; Classifier (UML); Training set; Gradient boosting; Ensemble learning; Data mining; Random forest","score_opus":0.14587677944015098,"score_gpt":0.31587666002732545,"score_spread":0.16999988058717447,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2040010062","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":[0.024027519,0.00051821215,0.97179586,0.00030296136,0.00018263514,0.0002745624,0.00016213386,0.0014820453,0.0012539946],"genre_scores_gemma":[0.25277093,0.00044399744,0.74367756,0.00036484134,0.00022688786,0.0006070378,0.0008399132,0.00018241629,0.00088635646],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9953701,0.0019063473,0.00027150995,0.00060420186,0.0016172364,0.00023070235],"domain_scores_gemma":[0.99058694,0.0043273717,0.0005689799,0.002237852,0.002063448,0.00021542353],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01012599,0.0010297024,0.0020922448,0.0024564222,0.000702399,0.0017103489,0.002086115,0.0011331512,0.0008363509],"category_scores_gemma":[0.019886298,0.00067028566,0.0014439478,0.00230897,0.0008285598,0.0021786347,0.002025449,0.0025333914,0.00075675367],"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":[0.00040694393,0.00037182667,0.008087236,0.00021445574,0.00024258066,0.000151637,0.00028360903,0.38717157,0.0073182248,0.013985057,0.008073177,0.5736937],"study_design_scores_gemma":[0.00003896306,0.00008663099,0.00078960985,0.000024771623,0.000025946874,0.00007759319,0.000033482844,0.9731848,0.005327262,0.017017066,0.0033713577,0.000022514565],"about_ca_topic_score_codex":0.000943973,"about_ca_topic_score_gemma":0.0010771593,"teacher_disagreement_score":0.01012599,"about_ca_system_score_codex":0.00077169825,"about_ca_system_score_gemma":0.00095432566,"threshold_uncertainty_score":0.05355203},"labels":[],"label_agreement":null},{"id":"W2044306058","doi":"10.1145/568574.568580","title":"Constrained frequent pattern mining","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":133,"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; Data mining; Sequential Pattern Mining; Constraint (computer-aided design); K-optimal pattern discovery; Context (archaeology); Process (computing); Association rule learning; Mathematics","score_opus":0.06108663403363127,"score_gpt":0.2512890157389533,"score_spread":0.19020238170532203,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2044306058","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":[0.011723552,0.0022365502,0.97733843,0.0007698439,0.00023788183,0.0005609869,0.0023014622,0.0006671914,0.0041640257],"genre_scores_gemma":[0.17496449,0.0032336742,0.8063217,0.0007076478,0.00041334043,0.0010942535,0.0091111,0.00021229523,0.0039415807],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9903092,0.002818198,0.0012215798,0.0019026884,0.003350458,0.00039783347],"domain_scores_gemma":[0.9751687,0.01457816,0.0019232117,0.0037078424,0.004147016,0.00047493455],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0052519194,0.0012765123,0.0024557854,0.0055087307,0.001369036,0.003103514,0.0037668862,0.0018581968,0.0051526423],"category_scores_gemma":[0.03992119,0.0008629601,0.001759415,0.008675211,0.0011306653,0.0054031676,0.0025834218,0.0018689602,0.0023341463],"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":[0.0004916203,0.00036496055,0.012069325,0.002616185,0.00081826805,0.002889421,0.0005433671,0.10498316,0.006464588,0.175185,0.034748055,0.65882605],"study_design_scores_gemma":[0.00010947052,0.00013841262,0.0021214213,0.00031697194,0.00015053432,0.0030189943,0.00030537552,0.5641562,0.003941527,0.3787076,0.046944283,0.00008923117],"about_ca_topic_score_codex":0.0013770885,"about_ca_topic_score_gemma":0.0016207218,"teacher_disagreement_score":0.0055087307,"about_ca_system_score_codex":0.0006319402,"about_ca_system_score_gemma":0.0018814228,"threshold_uncertainty_score":0.027775109},"labels":[],"label_agreement":null},{"id":"W2048976003","doi":"10.1145/1540276.1540299","title":"WebKDD 2008","year":2008,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Knowledge extraction; Las vegas; Conjunction (astronomy); Data science; Library science; World Wide Web; Data mining; History; Archaeology; Tourism","score_opus":0.05751956576975635,"score_gpt":0.2554444167508769,"score_spread":0.19792485098112056,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2048976003","genre_codex":"dataset","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":[0.0020654376,0.014559869,0.109832086,0.016676806,0.005772306,0.0008737277,0.61289394,0.1760378,0.06128797],"genre_scores_gemma":[0.006576434,0.008628371,0.0863389,0.00368892,0.0006053533,0.0012929232,0.85877854,0.008541122,0.025549455],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9903261,0.0016724387,0.0010595185,0.0013788451,0.0050969347,0.0004660473],"domain_scores_gemma":[0.98890346,0.002732135,0.0005253476,0.0034786714,0.003530724,0.00082969485],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.009461475,0.002605064,0.0043343096,0.0109142065,0.002458075,0.011936029,0.0068656923,0.0042408463,0.0514448],"category_scores_gemma":[0.031796105,0.0028013117,0.004313322,0.0073802765,0.00087144173,0.0069154277,0.0053914418,0.0066468026,0.118033305],"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":[0.00005105793,0.000039840215,0.00041969022,0.00038143957,0.00011142478,0.00003832971,0.000024616316,0.001100564,0.000083407394,0.0035796112,0.96241695,0.0317531],"study_design_scores_gemma":[0.00012409364,0.000015500444,0.00045229183,0.00018269902,0.000056131532,0.00020672088,0.000024046694,0.00771473,0.00075278344,0.019426355,0.97098774,0.000057001176],"about_ca_topic_score_codex":0.012069655,"about_ca_topic_score_gemma":0.016728954,"teacher_disagreement_score":0.94855523,"about_ca_system_score_codex":0.0026501603,"about_ca_system_score_gemma":0.008745913,"threshold_uncertainty_score":0.17210001},"labels":[],"label_agreement":null},{"id":"W2067575620","doi":"10.1145/772862.772886","title":"MDM/KDD2002","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Image Retrieval and Classification Techniques","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Conjunction (astronomy); Knowledge extraction; Data science; Series (stratigraphy); Information retrieval; Data mining; Geology","score_opus":0.07236190396419816,"score_gpt":0.24746513484452987,"score_spread":0.1751032308803317,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2067575620","genre_codex":"methods","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":[0.009912218,0.020003242,0.33551422,0.020694172,0.014802703,0.0020563821,0.30064002,0.17436263,0.12201443],"genre_scores_gemma":[0.032837387,0.010527412,0.34389612,0.0034876573,0.0017684381,0.0020523204,0.545409,0.007939018,0.05208261],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99321187,0.0009729071,0.0006590746,0.0007721998,0.004033761,0.00035008325],"domain_scores_gemma":[0.9925712,0.0010663824,0.00019448891,0.0030609004,0.0026103384,0.00049664057],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0056167357,0.0023118518,0.0035043396,0.009831926,0.0021487377,0.007460354,0.0058163847,0.003082678,0.028235193],"category_scores_gemma":[0.017537411,0.0014418481,0.0020198028,0.007913313,0.00060111907,0.0044315998,0.004614318,0.0027663037,0.04392592],"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":[0.00012884387,0.00009661775,0.00051877403,0.00043807103,0.00012441484,0.000112796544,0.00004041995,0.0024367517,0.00037519258,0.0048339777,0.84210616,0.14878795],"study_design_scores_gemma":[0.00013274617,0.00003842268,0.00094654824,0.00013706085,0.000057623525,0.0004189527,0.00007865416,0.025193077,0.002918023,0.020403983,0.94962484,0.000050146326],"about_ca_topic_score_codex":0.0071607335,"about_ca_topic_score_gemma":0.007915754,"teacher_disagreement_score":0.028235193,"about_ca_system_score_codex":0.0015516246,"about_ca_system_score_gemma":0.0031165634,"threshold_uncertainty_score":0.094456136},"labels":[],"label_agreement":null},{"id":"W2069116068","doi":"10.1145/507515.507523","title":"KDD Cup 2001 report","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Machine Learning in Bioinformatics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":110,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian Imperial Bank of Commerce (Canada)","funders":"National Science Foundation","keywords":"Computer science; Artificial intelligence","score_opus":0.036488664020318924,"score_gpt":0.2723488489952492,"score_spread":0.23586018497493028,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2069116068","genre_codex":"dataset","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":[0.0058059,0.031238545,0.07683369,0.046177574,0.02568327,0.0026344648,0.51106644,0.061402757,0.23915733],"genre_scores_gemma":[0.011523217,0.011197276,0.041183095,0.0037673682,0.002392936,0.0013540309,0.8502707,0.0028892192,0.07542215],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98324674,0.0020939661,0.0016808361,0.0018790142,0.010169036,0.0009303888],"domain_scores_gemma":[0.9593453,0.0038767261,0.001049142,0.0073372335,0.025502129,0.002889462],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014277994,0.004349593,0.003919384,0.014074648,0.0034234228,0.01518375,0.007351962,0.0051190737,0.062381167],"category_scores_gemma":[0.040800318,0.0013663169,0.0027009912,0.013729033,0.00078976364,0.009432475,0.005072314,0.0045950715,0.15419509],"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":[0.000046265788,0.000045311946,0.00031551055,0.00019038202,0.000027778711,0.000027891478,0.000009346163,0.00056633697,0.000051365536,0.0015289729,0.96496946,0.032221437],"study_design_scores_gemma":[0.00008870962,0.00003425711,0.0011622779,0.0001766385,0.00004481963,0.00013628075,0.000049479502,0.0047888164,0.0007265691,0.0046260445,0.9881252,0.000040917188],"about_ca_topic_score_codex":0.020853218,"about_ca_topic_score_gemma":0.012413182,"teacher_disagreement_score":0.062381167,"about_ca_system_score_codex":0.004508577,"about_ca_system_score_gemma":0.010788027,"threshold_uncertainty_score":0.20868576},"labels":[],"label_agreement":null},{"id":"W2082225138","doi":"10.1145/1046456.1046458","title":"Extracting relational data from HTML repositories","year":2004,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Set (abstract data type); Information retrieval; Data mining; Information extraction; Quality (philosophy)","score_opus":0.12972045116927916,"score_gpt":0.3005306323875561,"score_spread":0.17081018121827696,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2082225138","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":[0.16765764,0.0019118423,0.7773164,0.000998979,0.000070925045,0.0006377461,0.020846236,0.025099998,0.005460272],"genre_scores_gemma":[0.24147119,0.001131223,0.7131192,0.0001314196,0.000077599776,0.00025762434,0.0415202,0.0008820247,0.0014095011],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9969796,0.0004648877,0.000426762,0.00039901468,0.0015794813,0.00015028278],"domain_scores_gemma":[0.9853367,0.0059638396,0.0013581554,0.004004272,0.0030000184,0.00033695696],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0023204528,0.0009021745,0.0010544589,0.010877822,0.00070956664,0.0030615823,0.0015825682,0.00096500764,0.0011589874],"category_scores_gemma":[0.019864539,0.0007102127,0.0013212337,0.009080175,0.00045675263,0.003221844,0.0018944183,0.0009342164,0.0019221323],"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":[0.00036618853,0.00038592107,0.042028774,0.0014497873,0.0002699603,0.0017714334,0.001875047,0.025569743,0.046132784,0.008904273,0.016328966,0.8549171],"study_design_scores_gemma":[0.00013233109,0.00046191522,0.03943773,0.0004592144,0.000741255,0.0037483403,0.0025866593,0.5392147,0.28624675,0.044680744,0.08199382,0.00029665176],"about_ca_topic_score_codex":0.002656304,"about_ca_topic_score_gemma":0.0040337974,"teacher_disagreement_score":0.010877822,"about_ca_system_score_codex":0.00055777666,"about_ca_system_score_gemma":0.0013296778,"threshold_uncertainty_score":0.012271881},"labels":[],"label_agreement":null},{"id":"W2085995514","doi":"10.1145/772862.772888","title":"WEBKDD 2002","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":19,"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; Conjunction (astronomy); Theme (computing); Knowledge extraction; Data science; Web mining; Data mining; Information retrieval; World Wide Web; Web service","score_opus":0.061081370656482024,"score_gpt":0.2461904553370433,"score_spread":0.18510908468056128,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2085995514","genre_codex":"dataset","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":[0.0036374233,0.015810598,0.19205558,0.020928511,0.008032216,0.0011635148,0.55526674,0.16015731,0.042948153],"genre_scores_gemma":[0.012255161,0.00993473,0.12939861,0.0047014127,0.00074239913,0.0018134863,0.8075586,0.007987105,0.02560854],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99085134,0.0017070688,0.001269858,0.0011323718,0.0046865465,0.0003527501],"domain_scores_gemma":[0.9868537,0.0037577634,0.0005820954,0.003962881,0.00382758,0.0010159618],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.009158089,0.0021146175,0.00431655,0.008350158,0.0023231315,0.010128183,0.0059408327,0.003657967,0.027852349],"category_scores_gemma":[0.03009435,0.00271893,0.003165991,0.0072922665,0.00076637784,0.0051030223,0.003918403,0.0053297905,0.055288155],"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":[0.000120318306,0.00007616068,0.00068863836,0.0007788946,0.00019587729,0.00009822108,0.000034490615,0.002502848,0.00017492742,0.004592247,0.9365782,0.05415916],"study_design_scores_gemma":[0.00024118823,0.000023608907,0.0007189346,0.00018461795,0.00008690448,0.00036715067,0.000029757573,0.014707522,0.0015955737,0.018609619,0.9633626,0.000072511764],"about_ca_topic_score_codex":0.011893584,"about_ca_topic_score_gemma":0.016142948,"teacher_disagreement_score":0.97214764,"about_ca_system_score_codex":0.0021550683,"about_ca_system_score_gemma":0.008490472,"threshold_uncertainty_score":0.09317541},"labels":[],"label_agreement":null},{"id":"W2101549186","doi":"10.1145/2594473.2594476","title":"Ensembles for unsupervised outlier detection","year":2014,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":273,"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; Anomaly detection; Outlier; Cluster analysis; Artificial intelligence; Focus (optics); Machine learning; Unsupervised learning; Core (optical fiber); Ensemble learning; Data mining; Pattern recognition (psychology); Data science","score_opus":0.025405331016139964,"score_gpt":0.2520752218889637,"score_spread":0.22666989087282374,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2101549186","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":[0.004022422,0.00049145846,0.99333465,0.0002686388,0.00011400285,0.000033732038,0.00011666661,0.0006720388,0.0009464174],"genre_scores_gemma":[0.32852703,0.0015767064,0.6613675,0.00066295394,0.0010851256,0.00037768902,0.0018477815,0.0006825387,0.0038726449],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9940708,0.0021415208,0.0003911561,0.0014042312,0.0016112378,0.00038097028],"domain_scores_gemma":[0.98400635,0.006598711,0.0014665865,0.0041843355,0.0031570944,0.0005868931],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0055647874,0.0014226604,0.0025774536,0.0031494084,0.0014646916,0.0028752936,0.0026512532,0.001954368,0.0026599497],"category_scores_gemma":[0.024562167,0.0008705167,0.002108628,0.0027751958,0.0019107868,0.0041618627,0.004794029,0.0049526435,0.0018419359],"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":[0.00020614348,0.00015378639,0.009718353,0.0003650739,0.0007180205,0.000245107,0.000540842,0.41262805,0.0071772137,0.18400909,0.015161695,0.36907664],"study_design_scores_gemma":[0.000008656581,0.000060992472,0.0008924107,0.000051202354,0.000038921396,0.00013031559,0.00006205337,0.8451937,0.002080007,0.14411709,0.007320357,0.00004427313],"about_ca_topic_score_codex":0.0017406888,"about_ca_topic_score_gemma":0.0022718648,"teacher_disagreement_score":0.0055647874,"about_ca_system_score_codex":0.0009753072,"about_ca_system_score_gemma":0.0010754659,"threshold_uncertainty_score":0.029429793},"labels":[],"label_agreement":null},{"id":"W2110788923","doi":"10.1145/1117454.1117475","title":"WebKDD 2005","year":2005,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Theme (computing); Conjunction (astronomy); Web mining; World Wide Web; Data science; Knowledge extraction; Data mining; Web service","score_opus":0.036298116636475024,"score_gpt":0.2671368829012802,"score_spread":0.23083876626480515,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2110788923","genre_codex":"dataset","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":[0.0041338904,0.01712402,0.11363368,0.015466748,0.0069161663,0.0013728549,0.6918054,0.117106125,0.032441076],"genre_scores_gemma":[0.0116817085,0.007800427,0.103749685,0.003622262,0.0005778375,0.0018115375,0.85271657,0.0036200895,0.014419737],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98773885,0.00268544,0.0017199358,0.0018344974,0.0054642297,0.00055711303],"domain_scores_gemma":[0.9831117,0.0049267723,0.0007002512,0.005750616,0.0044511864,0.0010594794],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014731056,0.0027755958,0.0051921187,0.010575866,0.0033110455,0.012002045,0.007702077,0.0045257867,0.029673463],"category_scores_gemma":[0.043874722,0.0028492769,0.004745157,0.0075570624,0.00095371937,0.005477368,0.004991178,0.005654415,0.057641063],"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":[0.0001439183,0.00009195577,0.0014114099,0.0010049277,0.0004224712,0.00010708213,0.000053530268,0.0036926325,0.00015511537,0.0053399974,0.9305846,0.056992218],"study_design_scores_gemma":[0.00033156516,0.00004225452,0.0010467878,0.00029813655,0.00018116989,0.00039669572,0.000060775757,0.021981629,0.0016068972,0.024289107,0.94966495,0.00009994803],"about_ca_topic_score_codex":0.021729093,"about_ca_topic_score_gemma":0.029961118,"teacher_disagreement_score":0.029673463,"about_ca_system_score_codex":0.003261767,"about_ca_system_score_gemma":0.011843965,"threshold_uncertainty_score":0.0992676},"labels":[],"label_agreement":null},{"id":"W2112993176","doi":"10.1145/1294301.1294305","title":"Introduction to the special issue on data mining for health informatics","year":2007,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":18,"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; University of British Columbia","funders":"","keywords":"Personalized medicine; Computer science; Profiling (computer programming); Computational biology; Informatics; Metabolite profiling; Biomedicine; Data science; Oligonucleotide; Bioinformatics; Data mining; DNA; Biology; Metabolomics; Genetics","score_opus":0.05493152489629841,"score_gpt":0.3437162114384144,"score_spread":0.288784686542116,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2112993176","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00040116886,0.08150359,0.08007386,0.13646957,0.6471421,0.0006250517,0.003915236,0.0046217036,0.04524778],"genre_scores_gemma":[0.0018518334,0.054238595,0.03101238,0.05941332,0.7044128,0.0005518081,0.0060704597,0.0026598903,0.13978888],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99592894,0.0009555925,0.0005478277,0.00078996806,0.001618359,0.00015936098],"domain_scores_gemma":[0.9759517,0.013895913,0.0008299267,0.0018543749,0.0050718444,0.0023962932],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006042525,0.0023150425,0.0035410374,0.0066799615,0.0015746573,0.008875033,0.0024827407,0.004387085,0.09214429],"category_scores_gemma":[0.019510914,0.0011499263,0.0024078144,0.00620027,0.0015006893,0.009598449,0.004117718,0.008608382,0.068712555],"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":[0.000012172198,0.000021032396,0.00006489298,0.00019288699,0.000017563389,0.000032403866,0.000014965292,0.00008782917,0.00012856153,0.0013720554,0.9506359,0.04741964],"study_design_scores_gemma":[0.000012432192,0.00002462443,0.00024967533,0.00023413573,0.0000125058,0.00019265983,0.000022833876,0.0005052029,0.00007816074,0.005784146,0.99286556,0.000018091838],"about_ca_topic_score_codex":0.00072523963,"about_ca_topic_score_gemma":0.0015798189,"teacher_disagreement_score":0.09214429,"about_ca_system_score_codex":0.0011886769,"about_ca_system_score_gemma":0.0015922807,"threshold_uncertainty_score":0.30825335},"labels":[],"label_agreement":null},{"id":"W2122949122","doi":"10.1145/507515.507524","title":"MDM/KDD","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Image Retrieval and Classification Techniques","field":"Computer Science","cited_by":18,"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; Data science; Work (physics); Information retrieval; World Wide Web","score_opus":0.0716650625786421,"score_gpt":0.24935752757947455,"score_spread":0.17769246500083247,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2122949122","genre_codex":"methods","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":[0.009459023,0.016253425,0.4069296,0.020643782,0.01227831,0.001708729,0.15461874,0.11546643,0.262642],"genre_scores_gemma":[0.10351502,0.016083267,0.5283687,0.005693034,0.0037747591,0.0023291884,0.22561568,0.009914721,0.10470573],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99304575,0.00095167046,0.0006179529,0.0010504823,0.003977829,0.00035639602],"domain_scores_gemma":[0.9930131,0.0010370747,0.00023876887,0.003425978,0.0018537182,0.00043136178],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0042803152,0.0013647103,0.002630912,0.0086925905,0.0018843022,0.00835224,0.0042856005,0.0023966294,0.06591492],"category_scores_gemma":[0.016830625,0.000991069,0.0013991819,0.0061527584,0.0007946876,0.0051460327,0.0056353128,0.0019194786,0.07017294],"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":[0.0002000347,0.000113782255,0.0015701876,0.0009662665,0.00016062972,0.0001974617,0.00008883083,0.0025832148,0.0010759849,0.017594378,0.66940206,0.30604717],"study_design_scores_gemma":[0.00008239333,0.000034272212,0.0012471618,0.00017764556,0.00004491883,0.00063994527,0.0001699353,0.019060528,0.0035629054,0.034407087,0.9405256,0.000047522077],"about_ca_topic_score_codex":0.0021376132,"about_ca_topic_score_gemma":0.0029996305,"teacher_disagreement_score":0.06591492,"about_ca_system_score_codex":0.0010611649,"about_ca_system_score_gemma":0.0020353778,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W2128899929","doi":"10.1145/2207243.2207247","title":"Social network analysis and mining to support the assessment of on-line student participation","year":2012,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","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":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Automatic summarization; Computer science; Toolbox; Social network analysis; Identification (biology); Visualization; Data science; World Wide Web; Exploit; Social media; Online discussion; Information retrieval; Data mining","score_opus":0.076759501794539,"score_gpt":0.40748789889979475,"score_spread":0.33072839710525576,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2128899929","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":[0.30076495,0.00039926104,0.67942554,0.00087219896,0.00009263673,0.0007819736,0.005055937,0.0058991765,0.0067083295],"genre_scores_gemma":[0.6293612,0.00021952178,0.3643605,0.0000552939,0.00006315135,0.0005083874,0.0035472144,0.00013835484,0.0017463964],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9973477,0.0012866228,0.0002473943,0.00036822076,0.00066212553,0.00008784772],"domain_scores_gemma":[0.9850154,0.009319642,0.0022346394,0.0013108735,0.0015681472,0.00055122847],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0034519306,0.0007747635,0.0005087633,0.00840867,0.0005463275,0.0010975385,0.0008988763,0.0006392974,0.002034631],"category_scores_gemma":[0.015666863,0.00022897602,0.00049338816,0.004083227,0.00029342348,0.0017424213,0.0012985645,0.0005982706,0.00071834016],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000511169,0.001127446,0.17254004,0.0006236427,0.00051793834,0.00051927223,0.0028831447,0.054227956,0.020134415,0.012010966,0.012146702,0.72275734],"study_design_scores_gemma":[0.00004461997,0.0002611475,0.06711384,0.00009345093,0.00007684124,0.0004265814,0.0014047851,0.884607,0.011470689,0.019749423,0.014683636,0.00006788074],"about_ca_topic_score_codex":0.0020324334,"about_ca_topic_score_gemma":0.0037365633,"teacher_disagreement_score":0.00840867,"about_ca_system_score_codex":0.0005031094,"about_ca_system_score_gemma":0.0005884621,"threshold_uncertainty_score":0.01825577},"labels":[],"label_agreement":null},{"id":"W2155709205","doi":"10.1145/846183.846209","title":"WEBKDD'99","year":2000,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":26,"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":"York University; DePaul University","keywords":"Computer science; Profiling (computer programming); Knowledge extraction; Data science; World Wide Web; Information retrieval; Data mining","score_opus":0.027504247362564814,"score_gpt":0.25488755574249206,"score_spread":0.22738330837992723,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2155709205","genre_codex":"dataset","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":[0.0035845572,0.016199807,0.26065108,0.01805268,0.010185373,0.0015771608,0.3695178,0.2479245,0.07230699],"genre_scores_gemma":[0.016159061,0.011708355,0.19690529,0.0047806767,0.0009781393,0.002696942,0.7118547,0.014782295,0.04013462],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9862482,0.0024692959,0.0015146155,0.0017636846,0.0074620126,0.00054208894],"domain_scores_gemma":[0.98264354,0.0049120765,0.00052888494,0.0060737436,0.0045573427,0.0012844092],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012427366,0.002564441,0.0049830307,0.012685178,0.0036367138,0.014764848,0.0073517812,0.0042032856,0.034463704],"category_scores_gemma":[0.03654352,0.002892656,0.004261038,0.00886555,0.00119494,0.0066892887,0.0073416373,0.006464954,0.076417446],"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":[0.000105751955,0.000099648016,0.0006273994,0.0005657977,0.00019196195,0.00009563291,0.000056326055,0.002290422,0.00017419044,0.006618697,0.9157326,0.07344157],"study_design_scores_gemma":[0.00022511113,0.00002655597,0.00071793736,0.0001788008,0.000074853604,0.0002712271,0.00005099546,0.01895625,0.0016051708,0.032886837,0.94492364,0.00008252347],"about_ca_topic_score_codex":0.012633856,"about_ca_topic_score_gemma":0.016697312,"teacher_disagreement_score":0.034463704,"about_ca_system_score_codex":0.0023055596,"about_ca_system_score_gemma":0.008189315,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W2160036998","doi":"10.1145/1233321.1233324","title":"Utility-based anonymization for privacy preservation with less information loss","year":2006,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":78,"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":"Chinese University of Hong Kong; National Natural Science Foundation of China","keywords":"Computer science; Data anonymization; Microdata (statistics); Categorical variable; Data mining; Identifier; Information loss; Heuristic; Set (abstract data type); Data quality; Information privacy; Machine learning; Artificial intelligence","score_opus":0.04348191414472581,"score_gpt":0.25740419416415156,"score_spread":0.21392228001942576,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2160036998","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":[0.010847044,0.00039669665,0.9854872,0.00055901223,0.000046862155,0.00008389771,0.00016727271,0.00036583276,0.00204616],"genre_scores_gemma":[0.4798862,0.0009647369,0.5133491,0.00042548368,0.00027067063,0.00040887395,0.00083527295,0.0002614351,0.0035982572],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9878126,0.0068614897,0.0006029273,0.0015479035,0.0026816772,0.00049349235],"domain_scores_gemma":[0.9751213,0.008528172,0.0014237545,0.013185521,0.0014836335,0.00025766547],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0073420065,0.0009197907,0.0016349754,0.0019606096,0.0015951039,0.0032511037,0.0021935166,0.0011440172,0.0017503186],"category_scores_gemma":[0.02508089,0.00059452956,0.0014409642,0.0042970027,0.0026087873,0.0075089606,0.0041410844,0.0028295466,0.0006681969],"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":[0.0004822084,0.00017586195,0.002787054,0.0003093587,0.00024747138,0.00034327377,0.0011771134,0.15601745,0.009561687,0.54843783,0.008912064,0.27154857],"study_design_scores_gemma":[0.00007661905,0.00016068353,0.0009779863,0.00008478982,0.000109390996,0.0007547281,0.00040261185,0.42917785,0.016535146,0.53070146,0.020946039,0.00007263304],"about_ca_topic_score_codex":0.00060173124,"about_ca_topic_score_gemma":0.00067848334,"teacher_disagreement_score":0.0073420065,"about_ca_system_score_codex":0.0014476894,"about_ca_system_score_gemma":0.0019117822,"threshold_uncertainty_score":0.03882873},"labels":[],"label_agreement":null},{"id":"W2246590112","doi":"10.1145/2783702.2783708","title":"New Research Directions in Knowledge Discovery and Allied Spheres","year":2015,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Systems, Applications & Products in Data Processing (Canada)","funders":"Agence Nationale de la Recherche","keywords":"Computer science; Knowledge extraction; Data science; Schema (genetic algorithms); Viewpoints; Domain knowledge; Realm; Point (geometry); Information retrieval; World Wide Web; Knowledge management; Data mining","score_opus":0.16935484731935074,"score_gpt":0.3730839166959473,"score_spread":0.20372906937659657,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2246590112","genre_codex":"review","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0068165557,0.57272446,0.14813036,0.20584884,0.0050191046,0.000086446635,0.0002321893,0.00027163522,0.060870484],"genre_scores_gemma":[0.20227464,0.58101714,0.15102687,0.023852764,0.027740642,0.00037785096,0.0005170411,0.00019622683,0.012996862],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.986639,0.006915032,0.0007385156,0.0020965224,0.0029512197,0.00065970886],"domain_scores_gemma":[0.8842483,0.098851494,0.0019931677,0.006030641,0.005607362,0.0032690784],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.024995016,0.0012125365,0.0028783195,0.00869273,0.0036734862,0.017859032,0.0038532726,0.0077083693,0.008291723],"category_scores_gemma":[0.028530773,0.0010048059,0.0023222067,0.012552004,0.02812273,0.059678964,0.0076047215,0.01203755,0.0019524985],"study_design_candidate":"not_applicable","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":[0.000035236833,0.00006969159,0.00041087053,0.00075934693,0.00003244476,0.000082700615,0.0008407016,0.00047932033,0.00009636361,0.93866205,0.0067653526,0.05176594],"study_design_scores_gemma":[0.000015792582,0.000025267778,0.0001934253,0.00045083024,0.000014382215,0.00011665377,0.0009895836,0.0024910334,0.000095177165,0.9366944,0.058886796,0.000026638965],"about_ca_topic_score_codex":0.002929097,"about_ca_topic_score_gemma":0.00189987,"teacher_disagreement_score":0.024995016,"about_ca_system_score_codex":0.006219712,"about_ca_system_score_gemma":0.0065509756,"threshold_uncertainty_score":0.13218784},"labels":[],"label_agreement":null},{"id":"W2627710665","doi":"10.1145/772862.772887","title":"Multi-relational data mining","year":2002,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Computer science; Knowledge extraction; Data science; Relational database; Data mining; Information retrieval","score_opus":0.25641928703070915,"score_gpt":0.30813698964909747,"score_spread":0.05171770261838832,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2627710665","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":[0.0037110636,0.027726285,0.9485956,0.003167174,0.00078811566,0.00053655583,0.0033512071,0.0025520802,0.009572045],"genre_scores_gemma":[0.06739901,0.028479781,0.8844896,0.0018138466,0.0011954566,0.0004931945,0.010044333,0.00039272624,0.00569206],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9904523,0.0019556857,0.0011703017,0.0020360653,0.0041171215,0.00026846823],"domain_scores_gemma":[0.98936623,0.0048116497,0.0008931778,0.002504222,0.0021018626,0.0003228285],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006857553,0.001218159,0.001647424,0.0056410367,0.0010649444,0.0058543435,0.0039257053,0.0010988682,0.0051433356],"category_scores_gemma":[0.021246817,0.0006189688,0.0018974754,0.009295062,0.00073440105,0.009080381,0.0038846664,0.002009049,0.0058824597],"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":[0.00013034651,0.00018032236,0.007866713,0.0026101544,0.0005474271,0.0010154124,0.00033674552,0.007883908,0.0037988974,0.07670861,0.047306094,0.85161537],"study_design_scores_gemma":[0.000055846063,0.00023715988,0.0042338986,0.0010543884,0.0005908032,0.0075997887,0.0007719019,0.12936562,0.01538532,0.2542057,0.5863222,0.0001773451],"about_ca_topic_score_codex":0.0009818592,"about_ca_topic_score_gemma":0.0012520732,"teacher_disagreement_score":0.006857553,"about_ca_system_score_codex":0.0008066392,"about_ca_system_score_gemma":0.0016402175,"threshold_uncertainty_score":0.036266685},"labels":[],"label_agreement":null},{"id":"W2736013904","doi":"10.1145/1809400.1809425","title":"Efficient algorithms for the mining of constrained frequent patterns from uncertain data","year":2010,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":25,"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 Manitoba","funders":"","keywords":"Computer science; Association rule learning; Data mining; K-optimal pattern discovery; Database transaction; Uncertain data; Data stream mining; Knowledge extraction; Geodetic datum; Data science; Database","score_opus":0.10363999971016609,"score_gpt":0.321903110243897,"score_spread":0.21826311053373093,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2736013904","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":[0.00493444,0.0008094885,0.9929756,0.00018502334,0.000036040194,0.00015310373,0.00019076976,0.00034819322,0.00036726485],"genre_scores_gemma":[0.04114702,0.00073054037,0.95608354,0.000090622074,0.00009723192,0.00045420323,0.00092065585,0.000050332153,0.00042577917],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99670416,0.00079511054,0.0004878687,0.00062959606,0.0011742694,0.0002090059],"domain_scores_gemma":[0.9891756,0.007793809,0.0008052234,0.0010086097,0.0010557722,0.00016096162],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0037850204,0.0014907981,0.0025709474,0.0036164238,0.0012969843,0.0023644306,0.0036204278,0.0018746885,0.0015454149],"category_scores_gemma":[0.019599667,0.0011423733,0.0016029914,0.006404051,0.00096932007,0.0049301405,0.0028520976,0.0021755397,0.00097190525],"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":[0.0002913763,0.00030832813,0.0026593609,0.00060809695,0.00025658828,0.00040541225,0.00038739902,0.25179565,0.0026393086,0.02984852,0.0071959314,0.703604],"study_design_scores_gemma":[0.00009680348,0.00009077478,0.00047633378,0.000058030233,0.00004814102,0.0004053732,0.00011823342,0.92272294,0.0013424032,0.070610024,0.0040035006,0.000027570368],"about_ca_topic_score_codex":0.0017685117,"about_ca_topic_score_gemma":0.0025570916,"teacher_disagreement_score":0.0037850204,"about_ca_system_score_codex":0.00076373975,"about_ca_system_score_gemma":0.0017613407,"threshold_uncertainty_score":0.020017326},"labels":[],"label_agreement":null},{"id":"W2945319864","doi":"10.1145/3331651.3331653","title":"Urban Human Mobility","year":2019,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":107,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Multidisciplinary approach; Data science; Context (archaeology); Mobility model; Perspective (graphical); Field (mathematics); Global Positioning System; Urban computing; Human dynamics; Human–computer interaction; Artificial intelligence; Telecommunications; Sociology","score_opus":0.036755788776791795,"score_gpt":0.3235828484170946,"score_spread":0.2868270596403028,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2945319864","genre_codex":"other","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":[0.109940454,0.03508145,0.29897866,0.01839554,0.0015437932,0.00035382324,0.01508355,0.0014873475,0.5191354],"genre_scores_gemma":[0.88244355,0.030452669,0.0278974,0.0010857387,0.0004781792,0.0002627802,0.00694996,0.0001306654,0.050298974],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.9997316,0.00007308069,0.000012160383,0.00006772072,0.00006797424,0.000047491696],"domain_scores_gemma":[0.9997645,0.00006662401,0.000033915047,0.000037573485,0.000076621116,0.000020679996],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00025190006,0.00036881847,0.00021759463,0.0010989708,0.00060826173,0.0014043989,0.0005197416,0.00051508716,0.009492499],"category_scores_gemma":[0.0013368502,0.00016958709,0.00035053148,0.0024117047,0.00052383734,0.0020503036,0.001202847,0.00047367075,0.0017126007],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000042553213,0.00003081391,0.01885827,0.00066054356,0.0000761215,0.00031426936,0.001400452,0.039417624,0.0010483105,0.61630976,0.07855998,0.24328128],"study_design_scores_gemma":[0.000010970191,0.000056052086,0.020082692,0.0005067254,0.000063625186,0.00077510386,0.0033243876,0.0594993,0.0009869995,0.19689018,0.71773684,0.00006724621],"about_ca_topic_score_codex":0.00910354,"about_ca_topic_score_gemma":0.010726765,"teacher_disagreement_score":0.009492499,"about_ca_system_score_codex":0.0012040192,"about_ca_system_score_gemma":0.00083297666,"threshold_uncertainty_score":0.031755507},"labels":[],"label_agreement":null},{"id":"W3034271075","doi":"10.1145/3262579","title":"Session details: Special issue on learning from imbalanced datasets","year":2004,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":22,"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; Canadian Imperial Bank of Commerce (Canada)","funders":"","keywords":"Computer science; Session (web analytics); Machine learning; Data science; Artificial intelligence; World Wide Web","score_opus":0.037927200654235056,"score_gpt":0.28477606922886783,"score_spread":0.24684886857463278,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3034271075","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004856523,0.025733564,0.054878015,0.08053809,0.40366766,0.003263497,0.047207195,0.017621882,0.36223355],"genre_scores_gemma":[0.008774329,0.011766003,0.010020531,0.0075625917,0.11550184,0.000975743,0.030229414,0.0042113583,0.81095827],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9988193,0.00018663902,0.00007176059,0.0003143375,0.0004737676,0.00013410543],"domain_scores_gemma":[0.99330693,0.0017327832,0.00023234249,0.00072364707,0.00211855,0.0018858297],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0031817742,0.0020459567,0.0030998618,0.002117321,0.0019418787,0.006545901,0.0016164113,0.0033064175,0.5244928],"category_scores_gemma":[0.006777999,0.0006045218,0.0014050223,0.0031987135,0.00046252972,0.004947979,0.0025772844,0.0037645784,0.3605936],"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":[0.000049903883,0.000031292373,0.000041702897,0.00008005273,0.0000064537667,0.000015209711,0.000005183033,0.000031763095,0.00027252716,0.0002055576,0.9820238,0.017236728],"study_design_scores_gemma":[0.00009478854,0.00014269058,0.000977741,0.00014681458,0.00003161931,0.0001440572,0.00004581004,0.0007412262,0.0008753337,0.0029640326,0.9938048,0.000031034542],"about_ca_topic_score_codex":0.0007136277,"about_ca_topic_score_gemma":0.002192729,"teacher_disagreement_score":0.5244928,"about_ca_system_score_codex":0.0006748054,"about_ca_system_score_gemma":0.0014233569,"threshold_uncertainty_score":0.6782533},"labels":[],"label_agreement":null},{"id":"W3112675244","doi":"10.1145/3447556.3447563","title":"A Report on the First Workshop on Document Intelligence (DI) at NeurIPS 2019","year":2021,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Computer science; Library science; Intelligence analysis; Data science; World Wide Web; Computer security","score_opus":0.051440865561647335,"score_gpt":0.2786518561287233,"score_spread":0.22721099056707594,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3112675244","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.021402722,0.040274758,0.0886662,0.14029378,0.21958995,0.004308079,0.037439145,0.0072454754,0.44077998],"genre_scores_gemma":[0.023357784,0.013450496,0.035907023,0.008208128,0.011727274,0.0010795508,0.041724637,0.0030985493,0.86144656],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9882081,0.0023860107,0.0004642483,0.0013875149,0.0059139803,0.0016400993],"domain_scores_gemma":[0.977434,0.0023340676,0.00038442045,0.0012720033,0.011187529,0.0073880805],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014243333,0.0017670501,0.0012710565,0.0036552104,0.0030212756,0.012603111,0.0024048907,0.003095025,0.10614094],"category_scores_gemma":[0.014971792,0.00071648054,0.0016136074,0.0035290616,0.0010214433,0.0075982,0.0075699193,0.0053466023,0.086400755],"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":[0.0002511496,0.00017514484,0.0004981045,0.00028444166,0.00002017323,0.00011208115,0.000308093,0.00024839767,0.002287738,0.003005009,0.90778047,0.08502915],"study_design_scores_gemma":[0.000029930658,0.000086179534,0.0006569277,0.00013583388,0.000014667918,0.00006385956,0.00037201022,0.0003112793,0.0016468472,0.0013582222,0.9952968,0.000027530477],"about_ca_topic_score_codex":0.013954963,"about_ca_topic_score_gemma":0.019817747,"teacher_disagreement_score":0.10614094,"about_ca_system_score_codex":0.0043092994,"about_ca_system_score_gemma":0.011654351,"threshold_uncertainty_score":0.35507685},"labels":[],"label_agreement":null},{"id":"W3165211483","doi":"10.1145/3468507.3468511","title":"On the Applicability of Machine Learning Fairness Notions","year":2021,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":70,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec à Montréal","funders":"Agence Nationale de la Recherche","keywords":"Subjectivity; Set (abstract data type); Computer science; Artificial intelligence; Machine learning; Fairness measure; Epistemology","score_opus":0.08874008202605167,"score_gpt":0.35175980610599283,"score_spread":0.2630197240799412,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3165211483","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":[0.066635884,0.0066391574,0.82092506,0.021750143,0.00077771395,0.00023996603,0.00027466836,0.00022358977,0.08253382],"genre_scores_gemma":[0.8743927,0.0023909898,0.11517175,0.002527869,0.0013200045,0.00033996542,0.00010486348,0.00012491192,0.0036268583],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9651904,0.020751167,0.0015110893,0.0040623415,0.006583966,0.0019010293],"domain_scores_gemma":[0.8317467,0.14502926,0.0065395674,0.008170796,0.0065346267,0.0019791014],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.040462986,0.0011123922,0.001358533,0.0031911344,0.003987904,0.0066492083,0.0023768728,0.003349282,0.0044395654],"category_scores_gemma":[0.118934795,0.0005449083,0.0015964813,0.002520363,0.013910423,0.01280024,0.005647481,0.0073442687,0.0005563529],"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":[0.000076467106,0.00004104744,0.0015004667,0.000071254864,0.000020107482,0.000051893574,0.0003968925,0.01139375,0.00016710603,0.96764684,0.0009345767,0.017699547],"study_design_scores_gemma":[0.000016163562,0.000047060163,0.0006826743,0.00011775946,0.000011961088,0.00006520068,0.00015951219,0.036393754,0.0002778718,0.9586048,0.003597314,0.000025881398],"about_ca_topic_score_codex":0.0028634013,"about_ca_topic_score_gemma":0.0012928784,"teacher_disagreement_score":0.040462986,"about_ca_system_score_codex":0.004570568,"about_ca_system_score_gemma":0.0032156832,"threshold_uncertainty_score":0.21399128},"labels":[],"label_agreement":null},{"id":"W4311079932","doi":"10.1145/3575637.3575645","title":"Finding Multidimensional Simpson's Paradox","year":2022,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Data Mining Algorithms and Applications","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; Burnaby Hospital","funders":"","keywords":"Computer science; Curse of dimensionality; Multidimensional data; Data science; Phenomenon; Data mining; Machine learning; Epistemology","score_opus":0.04358511030504669,"score_gpt":0.2740753601844798,"score_spread":0.23049024987943312,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4311079932","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":[0.48575655,0.0029588745,0.48872963,0.0068387124,0.0005146546,0.00016499004,0.0010064741,0.00039842143,0.013631698],"genre_scores_gemma":[0.89724433,0.0008312762,0.098454446,0.0009243557,0.00029593517,0.00015430403,0.0006311869,0.000074774056,0.0013893748],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99411917,0.0018854856,0.0005717825,0.0014461897,0.0015969885,0.00038037525],"domain_scores_gemma":[0.9500843,0.036470424,0.0050537284,0.0041869036,0.0027456132,0.0014590578],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0077705747,0.0006511525,0.002227119,0.004647752,0.00303803,0.0046917857,0.001790905,0.0019412054,0.0029698492],"category_scores_gemma":[0.07232878,0.0006749765,0.0014231299,0.0043415697,0.004138089,0.008352136,0.004317204,0.0036803198,0.0002700079],"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":[0.0003250371,0.00015283062,0.057932362,0.0007496676,0.00042330657,0.0019731652,0.0029789885,0.019592771,0.0027345286,0.79937786,0.0127262445,0.10103331],"study_design_scores_gemma":[0.00004030241,0.00003330301,0.0039059925,0.000076447286,0.00003841929,0.0008158768,0.00064408965,0.068817675,0.00074244046,0.92235386,0.0024592592,0.00007230094],"about_ca_topic_score_codex":0.001140653,"about_ca_topic_score_gemma":0.0009145209,"teacher_disagreement_score":0.0077705747,"about_ca_system_score_codex":0.0017554732,"about_ca_system_score_gemma":0.001618492,"threshold_uncertainty_score":0.041095197},"labels":[],"label_agreement":null},{"id":"W4311079975","doi":"10.1145/3575637.3575654","title":"Experiences with Contrastive Predictive Coding in Industrial Time-Series Classification","year":2022,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Time Series Analysis and Forecasting","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","funders":"","keywords":"Computer science; Artificial intelligence; Classifier (UML); Machine learning; Coding (social sciences); Data mining; Multivariate statistics; Pattern recognition (psychology); Artificial neural network; Time series; Statistics","score_opus":0.04476407521292988,"score_gpt":0.22947922000461182,"score_spread":0.18471514479168194,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4311079975","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":[0.21575996,0.0025474357,0.76977795,0.0029359977,0.00020719919,0.0001497438,0.00016713735,0.0021599303,0.006294702],"genre_scores_gemma":[0.77486515,0.0007044714,0.2220449,0.00028264456,0.00012991233,0.000066265704,0.0003191827,0.00010066706,0.0014867862],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99764806,0.001268742,0.00009405665,0.0003042615,0.00058317813,0.00010164587],"domain_scores_gemma":[0.9872737,0.00934541,0.0002406172,0.0013881694,0.0015549593,0.00019717103],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006951128,0.00070777995,0.00060270703,0.0007749548,0.00046864178,0.001017187,0.0015543071,0.001053642,0.00068121916],"category_scores_gemma":[0.016957402,0.00028370236,0.0005170659,0.0012169889,0.0015527199,0.0019486748,0.001009874,0.003094657,0.00022535831],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005549998,0.00052169437,0.006881026,0.00015192472,0.00010075423,0.00019204972,0.0012500685,0.3923897,0.010502162,0.016880129,0.0054730014,0.5651025],"study_design_scores_gemma":[0.000016919785,0.0001142256,0.00085921783,0.000013415654,0.000008419917,0.00005394838,0.00006682951,0.98078674,0.0085315,0.007568834,0.0019593518,0.000020620195],"about_ca_topic_score_codex":0.011453551,"about_ca_topic_score_gemma":0.0066031395,"teacher_disagreement_score":0.011453551,"about_ca_system_score_codex":0.0014059466,"about_ca_system_score_gemma":0.0006932846,"threshold_uncertainty_score":0.036761522},"labels":[],"label_agreement":null},{"id":"W4406939872","doi":"10.1145/3715073.3715081","title":"Graph Fairness via Authentic Counterfactuals: Tackling Structural and Causal Challenges","year":2025,"lang":"en","type":"article","venue":"ACM SIGKDD Explorations Newsletter","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","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":"University of Manitoba","funders":"","keywords":"Counterfactual conditional; Computer science; Psychology; Counterfactual thinking; Social psychology","score_opus":0.058918373595661724,"score_gpt":0.36371029940612337,"score_spread":0.30479192581046166,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4406939872","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":[0.01517421,0.00027382132,0.98125297,0.0014149181,0.00007178869,0.000256797,0.00010379981,0.00018325241,0.0012684386],"genre_scores_gemma":[0.56624854,0.0004239454,0.4295262,0.0009566079,0.0002879954,0.00097036053,0.00027071973,0.00016228357,0.0011534267],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.931234,0.056695916,0.0012411899,0.0058044665,0.0042371843,0.00078720954],"domain_scores_gemma":[0.6355671,0.3125336,0.012204135,0.032188304,0.005926295,0.0015805223],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.08245675,0.0016230388,0.0021420117,0.0030011397,0.0022356533,0.004892112,0.0044571646,0.0033548912,0.0032953734],"category_scores_gemma":[0.26178995,0.0012992938,0.0018307858,0.0025248744,0.008828229,0.0085280305,0.007319254,0.005969099,0.00035355287],"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":[0.0004140274,0.00017312344,0.007240899,0.0003937298,0.00029501264,0.0004008133,0.0014524189,0.18158856,0.00077261514,0.7136476,0.0019025075,0.09171874],"study_design_scores_gemma":[0.00006443323,0.00005552564,0.00031718248,0.000059390615,0.000045079476,0.000060458868,0.00008224523,0.28825715,0.0006748924,0.70870423,0.0016589855,0.000020451205],"about_ca_topic_score_codex":0.0023598797,"about_ca_topic_score_gemma":0.0023605477,"teacher_disagreement_score":0.08245675,"about_ca_system_score_codex":0.0031667557,"about_ca_system_score_gemma":0.005092664,"threshold_uncertainty_score":0.4360782},"labels":[],"label_agreement":null}]}