{"meta":{"query_hash":"209a364fe34f","filters":{"venue":"Software Engineering and Data Engineering"},"cohort_total":5,"direct_labels_cover":0,"predictions_cover":5,"exported":5,"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/209a364fe34f","api":"https://metacan.xera.ac/api/v1/cohort?venue=Software+Engineering+and+Data+Engineering"},"results":[{"id":"W167180509","doi":"","title":"Exploiting Schema Information for Efficient Streaming Queries.","year":2010,"lang":"en","type":"article","venue":"Software Engineering and Data Engineering","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"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; Schema (genetic algorithms); Streaming data; Information retrieval; Data mining","score_opus":0.010549608556253563,"score_gpt":0.22412563717463432,"score_spread":0.21357602861838076,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W167180509","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":[0.073246695,0.002583743,0.89325815,0.0015685504,0.00030002845,0.00038567165,0.005098045,0.016743941,0.0068151453],"genre_scores_gemma":[0.41714373,0.0016832754,0.5620608,0.00047644821,0.00027630434,0.00021696044,0.013148344,0.001169548,0.0038246578],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9970444,0.00077706587,0.000344048,0.0004250081,0.0011837562,0.00022565154],"domain_scores_gemma":[0.98717064,0.0061995834,0.0005490526,0.0044637686,0.0013146127,0.00030232346],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038237695,0.00088784343,0.0012781511,0.0023607204,0.0008079739,0.0029979611,0.002006037,0.001199702,0.0043738857],"category_scores_gemma":[0.017234545,0.0007386214,0.00087841484,0.0046047284,0.0006459225,0.008981918,0.0026535427,0.0017438276,0.0018188227],"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":[0.002505891,0.00057502266,0.010625912,0.00082362385,0.00033038476,0.00068190164,0.0013717492,0.030040722,0.066595525,0.07838292,0.07155204,0.73651433],"study_design_scores_gemma":[0.00028457821,0.00037792142,0.0021225489,0.00013014002,0.00025360446,0.0011282419,0.0007726968,0.72606057,0.05287082,0.17063066,0.0452675,0.00010065509],"about_ca_topic_score_codex":0.0035772214,"about_ca_topic_score_gemma":0.005031225,"teacher_disagreement_score":0.0043738857,"about_ca_system_score_codex":0.0007868884,"about_ca_system_score_gemma":0.0015787851,"threshold_uncertainty_score":0.020222306},"labels":[],"label_agreement":null},{"id":"W172667635","doi":"","title":"An Algebraic Query Method for Static Program Analysis and Measurement.","year":2008,"lang":"en","type":"article","venue":"Software Engineering and Data Engineering","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Query optimization; Query language; Algebraic number; Programming language; Theoretical computer science; Information retrieval; Mathematics","score_opus":0.04551195105983421,"score_gpt":0.3039288780590528,"score_spread":0.2584169269992186,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W172667635","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":[0.00069824065,0.00010258693,0.991905,0.00009363148,0.000037154347,0.00007090913,0.00018978836,0.0055765924,0.0013260962],"genre_scores_gemma":[0.05656659,0.0003151655,0.9350242,0.0002196057,0.0001723797,0.00046736756,0.0010881851,0.0018480569,0.004298406],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9892083,0.0022469228,0.0011882101,0.0016527189,0.005218364,0.0004854985],"domain_scores_gemma":[0.98967314,0.0040961993,0.0006044764,0.0030167017,0.0023323172,0.0002771459],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051495074,0.001377604,0.001783252,0.0036265757,0.0013360153,0.004310476,0.0039261673,0.0013466239,0.014167942],"category_scores_gemma":[0.02683864,0.0012171024,0.0019818246,0.0047782464,0.0023361682,0.00838291,0.0054696817,0.002888051,0.0053999065],"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":[0.000418758,0.00015630461,0.0016677581,0.00074983865,0.0001739374,0.0002124715,0.0007203282,0.0059572486,0.011723707,0.44208708,0.030182706,0.50594985],"study_design_scores_gemma":[0.0002712445,0.00028684345,0.0013869229,0.0002402305,0.0003227004,0.0011716578,0.00045309498,0.26958093,0.030706704,0.54853034,0.14685813,0.00019116448],"about_ca_topic_score_codex":0.0039142063,"about_ca_topic_score_gemma":0.0039354707,"teacher_disagreement_score":0.014167942,"about_ca_system_score_codex":0.0012526822,"about_ca_system_score_gemma":0.0030556994,"threshold_uncertainty_score":0.04739648},"labels":[],"label_agreement":null},{"id":"W205254203","doi":"","title":"Ontology-based Intelligent Network-Forensics Investigation.","year":2010,"lang":"en","type":"article","venue":"Software Engineering and Data Engineering","topic":"Digital and Cyber Forensics","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"","keywords":"Network forensics; Ontology; Computer science; Computer forensics; Domain (mathematical analysis); Knowledge base; Domain knowledge; Intrusion detection system; Digital forensics; Data mining; Computer security; World Wide Web; Software engineering","score_opus":0.014491419770314975,"score_gpt":0.20635898809934036,"score_spread":0.19186756832902538,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W205254203","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":[0.005571046,0.0007770493,0.9672035,0.0036846183,0.00019241977,0.0003150267,0.0005365013,0.0008460651,0.020873912],"genre_scores_gemma":[0.13881715,0.0013352244,0.8510717,0.0008178839,0.00009412421,0.0002796177,0.0019590436,0.00012997603,0.005495309],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9963129,0.0014516276,0.0004312792,0.00041348938,0.0011836456,0.00020707335],"domain_scores_gemma":[0.9967552,0.0011915133,0.0003816937,0.0008065266,0.00064988906,0.00021523502],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00463571,0.000565363,0.00056120503,0.0034898676,0.0017518612,0.0047387993,0.0019239014,0.00139655,0.0025153991],"category_scores_gemma":[0.0066485093,0.00042209929,0.0017988656,0.0021058435,0.0028727548,0.009244426,0.0041161873,0.0019191715,0.00090929767],"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":[0.000048649927,0.00021985544,0.0021765856,0.00054002047,0.00017444954,0.00063252664,0.0021278253,0.015280496,0.0047408612,0.81314033,0.011578383,0.14934],"study_design_scores_gemma":[0.000028712984,0.000031560405,0.001180974,0.00054734986,0.00014838274,0.0008138686,0.0017964435,0.16669914,0.0059354263,0.6448445,0.17790093,0.00007273525],"about_ca_topic_score_codex":0.0062453626,"about_ca_topic_score_gemma":0.0064617232,"teacher_disagreement_score":0.0062453626,"about_ca_system_score_codex":0.0020821523,"about_ca_system_score_gemma":0.004119181,"threshold_uncertainty_score":0.024516225},"labels":[],"label_agreement":null},{"id":"W2179793844","doi":"","title":"gRUP- A Globalized Approach to Software Engineering.","year":2008,"lang":"en","type":"article","venue":"Software Engineering and Data Engineering","topic":"Software Engineering Techniques and Practices","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":"University of Ottawa","funders":"","keywords":"Globalization; Rational Unified Process; Internationalization; Computer science; Process (computing); Software; Schedule; Software engineering; Software development process; Software quality; Software development; Process management; Systems engineering; Engineering; Business; Economics; International trade","score_opus":0.030154209944975283,"score_gpt":0.24602785104300554,"score_spread":0.21587364109803026,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2179793844","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":[0.006332807,0.004373207,0.7608339,0.0064358716,0.0007214137,0.0004958317,0.0003107565,0.0018067322,0.21868946],"genre_scores_gemma":[0.12841006,0.0061197127,0.8187005,0.0026227196,0.00033981304,0.0012739842,0.0006453797,0.00068185804,0.041206043],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9939752,0.0034284538,0.0004207457,0.0006457903,0.0013034858,0.0002262267],"domain_scores_gemma":[0.99605274,0.0014245572,0.00025804856,0.0013564483,0.0004897649,0.0004184589],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0056361095,0.0006516715,0.0005220392,0.0026212714,0.0015695406,0.0068038385,0.0014466253,0.001193052,0.006432093],"category_scores_gemma":[0.0060631773,0.0004187861,0.00086904154,0.0025799775,0.00530145,0.005546168,0.007135566,0.0030950205,0.0022979216],"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":[0.000011651841,0.000024029518,0.0003423705,0.00023236663,0.00002125777,0.0001044966,0.0016121414,0.0016259543,0.00049845024,0.916132,0.006276655,0.07311865],"study_design_scores_gemma":[0.000013413713,0.000041324314,0.00041612773,0.00036902155,0.000023282842,0.0003542122,0.0011195361,0.004509698,0.0007173858,0.60267645,0.389739,0.00002050022],"about_ca_topic_score_codex":0.0006545067,"about_ca_topic_score_gemma":0.001076614,"teacher_disagreement_score":0.0068038385,"about_ca_system_score_codex":0.001592196,"about_ca_system_score_gemma":0.0046964013,"threshold_uncertainty_score":0.029806972},"labels":[],"label_agreement":null},{"id":"W45025849","doi":"","title":"Unsupervised Identity Application Fraud Detection using Rule-based Decision Tree.","year":2010,"lang":"en","type":"article","venue":"Software Engineering and Data Engineering","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"","keywords":"Decision tree; Computer science; Identity (music); Artificial intelligence; Machine learning; Identity theft; Data mining; Pattern recognition (psychology); Computer security","score_opus":0.015521125092752195,"score_gpt":0.2569089143695922,"score_spread":0.24138778927684,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W45025849","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":[0.2066913,0.0016930837,0.7796442,0.001065232,0.00030576682,0.0005104366,0.0018615737,0.0036247324,0.004603617],"genre_scores_gemma":[0.74638444,0.00033045496,0.24868749,0.00018169088,0.00012712345,0.00016428765,0.0026489522,0.00007025005,0.0014052757],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9968111,0.0010431804,0.0003224091,0.0004563474,0.0011175007,0.00024954748],"domain_scores_gemma":[0.990969,0.004788059,0.0010231527,0.0009384027,0.0020333487,0.0002480608],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036944307,0.0007143489,0.0012403358,0.0037577194,0.0006509127,0.0016954003,0.0017765184,0.0011942037,0.0008908343],"category_scores_gemma":[0.01228828,0.00026469777,0.0008278421,0.0027670392,0.00039105245,0.0021527624,0.00084696873,0.0013207344,0.0011114029],"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":[0.0010520022,0.0015412108,0.047387738,0.00036932892,0.0004560327,0.0005222269,0.00021829123,0.10129751,0.009581819,0.005557751,0.017599074,0.81441706],"study_design_scores_gemma":[0.000027045615,0.0000870138,0.004029343,0.000030404517,0.000066540735,0.00015331869,0.00005880387,0.98249084,0.0053594154,0.0060085547,0.0016740725,0.000014531928],"about_ca_topic_score_codex":0.0017036314,"about_ca_topic_score_gemma":0.0022850165,"teacher_disagreement_score":0.0037577194,"about_ca_system_score_codex":0.0005453525,"about_ca_system_score_gemma":0.0010652808,"threshold_uncertainty_score":0.019538224},"labels":[],"label_agreement":null}]}