{"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027911529,0.00013359927,0.96982384,0.000024321416,0.00087469746,0.00018429072,0.00022456703,0.00081803696,0.000005149653],"genre_scores_gemma":[0.12667693,0.0000085921865,0.87273,0.00001526013,0.00018024228,0.000055753386,0.00030423322,0.000023766723,0.000005183262],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988784,0.0000026024072,0.0002838397,0.00033307806,0.00015141332,0.00035065477],"domain_scores_gemma":[0.9985929,0.00023771176,0.0000624432,0.0009182273,0.00006124663,0.00012744156],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00033201475,0.00021001157,0.00018440431,0.00014854183,0.000121505014,0.0001545435,0.00045656358,0.00007069259,0.0000013141859],"category_scores_gemma":[0.0014052646,0.0002160813,0.000026996273,0.00018897737,0.000012983609,0.0021735341,0.0005368513,0.00022439598,0.000003413614],"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":[0.000013174144,0.000046509085,0.00057890743,0.0023769196,0.0001124236,0.000015496938,0.0019028968,0.41633633,0.025884325,0.4134121,0.00056428317,0.13875663],"study_design_scores_gemma":[0.00024721227,0.000016459548,0.00024670726,0.0001075244,0.0000055119817,0.000036485904,0.000043879078,0.8808303,0.0016147449,0.0000066298935,0.11651477,0.00032981083],"about_ca_topic_score_codex":0.000015205917,"about_ca_topic_score_gemma":0.0000021938954,"teacher_disagreement_score":0.46449393,"about_ca_system_score_codex":0.000016087532,"about_ca_system_score_gemma":0.000027815022,"threshold_uncertainty_score":0.881154},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0077169277,0.00070573814,0.9794314,0.000026770278,0.00010270953,0.00027532692,0.000043222164,0.011697293,5.977981e-7],"genre_scores_gemma":[0.13564023,0.000043765674,0.864017,0.000028702305,0.000052987627,0.00009910111,0.00008074271,0.00003512077,0.0000023496707],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9982553,0.000021983664,0.0002805321,0.0007447686,0.00028512927,0.00041228096],"domain_scores_gemma":[0.9978941,0.0005210312,0.00005475144,0.0012050401,0.00009874179,0.00022629797],"candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.00094900845,0.0002852155,0.0003616084,0.00043205338,0.00014296819,0.00014207565,0.0007274965,0.000083985564,5.6104267e-7],"category_scores_gemma":[0.0015236347,0.00029414674,0.000051418818,0.0007063431,0.000020535792,0.0007299885,0.00029961503,0.00016573275,3.9270597e-7],"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":[0.00003765783,0.0006423058,0.06491062,0.0029725768,0.003284759,0.00021614159,0.0034710688,0.21376614,0.0032605194,0.0048851166,0.0066033383,0.69594973],"study_design_scores_gemma":[0.00018749996,0.00012632455,0.009031465,0.000058719455,0.00011712279,0.00007401121,0.0000019220247,0.9885877,0.0001819017,0.00016130753,0.0010845066,0.0003875426],"about_ca_topic_score_codex":0.000074813375,"about_ca_topic_score_gemma":0.0000035025857,"teacher_disagreement_score":0.7748215,"about_ca_system_score_codex":0.000033853885,"about_ca_system_score_gemma":0.000041435105,"threshold_uncertainty_score":0.99995106},"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":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.024507632,0.00023173205,0.9724181,0.00012126423,0.0016015596,0.00008153668,0.000029806373,0.0009848046,0.000023552313],"genre_scores_gemma":[0.28345424,0.000010222349,0.7157902,0.0001771965,0.00031311455,0.000011396144,0.00018367915,0.000036309313,0.000023629584],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99874103,0.000004002329,0.00022182243,0.00046107304,0.00016966226,0.000402392],"domain_scores_gemma":[0.99843186,0.00018130297,0.00003584577,0.0010828222,0.00004334167,0.00022482434],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00023526192,0.00023259611,0.00018728437,0.00009059827,0.000060460297,0.00022121359,0.0009000043,0.000106593834,0.0000026378752],"category_scores_gemma":[0.00039104192,0.00023346549,0.000031839016,0.00031481803,0.00004119064,0.00067401904,0.0005516601,0.0003824414,0.000011220938],"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":[0.00000474809,0.00003508058,0.006720083,0.00029119072,0.00011191613,0.00007721117,0.0003196103,0.65648335,0.0010491274,0.21584858,0.005694917,0.11336416],"study_design_scores_gemma":[0.00015875732,0.000029181603,0.0023010771,0.00006529884,0.000011197377,0.000046320332,0.000002224015,0.95966166,0.00070939196,0.0008104139,0.035772305,0.00043214185],"about_ca_topic_score_codex":0.000011432506,"about_ca_topic_score_gemma":0.000011778472,"teacher_disagreement_score":0.3031783,"about_ca_system_score_codex":0.000012772503,"about_ca_system_score_gemma":0.000049726943,"threshold_uncertainty_score":0.95204467},"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":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008123294,0.0016316541,0.9807474,0.000076900214,0.0010974976,0.00042291716,0.00011558732,0.007761801,0.000022966575],"genre_scores_gemma":[0.054587487,0.00022711206,0.94426686,0.00012699877,0.0003071021,0.0001234988,0.00014190756,0.00014997543,0.000069037545],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9963034,0.000021380893,0.0005788232,0.0014286704,0.00060327793,0.0010644643],"domain_scores_gemma":[0.9957008,0.0006856032,0.00008985544,0.002728594,0.000103371654,0.0006918186],"candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.00061464077,0.0007752176,0.00066856487,0.00057946466,0.00018885711,0.00031374575,0.002799768,0.0002713548,0.0000058783403],"category_scores_gemma":[0.003930936,0.0008438044,0.00011254695,0.0013253724,0.000025402627,0.0017817469,0.0019979216,0.0006687451,0.00003384109],"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":[0.000045182256,0.0003531204,0.0044323895,0.0016010506,0.00056392845,0.00051885395,0.0016959268,0.9361658,0.0012275048,0.016423365,0.024814116,0.012158771],"study_design_scores_gemma":[0.0007605818,0.00015112781,0.007205013,0.0002862911,0.000052309522,0.0014771675,0.000007859019,0.7447689,0.00043238368,0.000019164088,0.24276133,0.0020778845],"about_ca_topic_score_codex":0.00006238959,"about_ca_topic_score_gemma":5.208742e-7,"teacher_disagreement_score":0.21794721,"about_ca_system_score_codex":0.00011920507,"about_ca_system_score_gemma":0.00008199299,"threshold_uncertainty_score":0.9994013},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.08990836,0.00007364827,0.90761673,0.000017508397,0.0004751861,0.00017010256,0.000051114373,0.0016858585,0.0000014662289],"genre_scores_gemma":[0.45829278,0.000007782935,0.54146236,0.000013903562,0.0000773618,0.000021457166,0.00009903748,0.00002445688,8.339876e-7],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984883,0.000008536951,0.0002861176,0.0006498699,0.0002795633,0.00028760917],"domain_scores_gemma":[0.99764085,0.00020717627,0.00006638021,0.0018803675,0.00006833434,0.00013687207],"candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0004606425,0.0002313687,0.00018001872,0.00030256316,0.00011341815,0.0002806728,0.0012767386,0.00015446762,0.000002559206],"category_scores_gemma":[0.0007189574,0.00025955262,0.000030745698,0.00055999897,0.00001727787,0.0020382628,0.0004683536,0.0003849056,0.000007924204],"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":[0.000009685257,0.00006902214,0.0021967606,0.00024768495,0.000031287927,0.000008268423,0.00006303661,0.070604876,0.63575125,0.0048779617,0.000058408936,0.28608176],"study_design_scores_gemma":[0.00020820454,0.000010570893,0.007828539,0.000043622655,0.000009296525,0.000017866623,0.0000012428145,0.971262,0.017743237,0.00009116276,0.0024973084,0.0002869733],"about_ca_topic_score_codex":0.000038957653,"about_ca_topic_score_gemma":0.000013440737,"teacher_disagreement_score":0.9006571,"about_ca_system_score_codex":0.000049902694,"about_ca_system_score_gemma":0.000034331868,"threshold_uncertainty_score":0.9999857},"labels":[],"label_agreement":null}]}