{"meta":{"query_hash":"5250998b0ea9","filters":{"venue":"Meeting of the Association for Computational Linguistics"},"cohort_total":22,"direct_labels_cover":0,"predictions_cover":22,"exported":22,"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/5250998b0ea9","api":"https://metacan.xera.ac/api/v1/cohort?venue=Meeting+of+the+Association+for+Computational+Linguistics"},"results":[{"id":"W1458921027","doi":"","title":"Ontology-Based Extraction and Summarization of Protein Mutation Impact Information","year":2010,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","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":"Concordia University","funders":"","keywords":"Automatic summarization; Computer science; Ontology; Information retrieval; Information extraction; Mutation; Pace; Data mining; Geography; Biology; Genetics","score_opus":0.006578853222055952,"score_gpt":0.27714733627980903,"score_spread":0.2705684830577531,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1458921027","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.95515615,0.000024116158,0.043254,0.00020611189,0.000511244,0.00031891695,0.000118512304,0.000013653695,0.00039730192],"genre_scores_gemma":[0.96829224,5.657834e-7,0.031218342,0.000028441878,0.00011018476,0.00000934104,0.00030608257,0.0000045232296,0.000030288267],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9993622,0.000040889125,0.0002774417,0.00007567706,0.00016763968,0.00007616616],"domain_scores_gemma":[0.9982817,0.00019000437,0.00062813784,0.00007087355,0.000811713,0.000017525215],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.00048559686,0.000062452906,0.0000925732,0.000043287815,0.000082525454,0.000011695391,0.000068880436,0.00014887161,8.30727e-7],"category_scores_gemma":[0.012056825,0.000051584833,0.000063835476,0.00007160168,0.000057189307,0.0000032506875,0.00001988214,0.00007142694,2.021174e-7],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008261037,0.00041936888,0.40900794,0.00072203646,0.00044045446,2.1461334e-7,0.00067073357,0.18879275,0.3496899,0.022254132,0.0022341746,0.024942195],"study_design_scores_gemma":[0.0042519015,0.0011713401,0.494341,0.00019849864,0.00022711251,0.0000055811,0.00017818442,0.2192615,0.22413819,0.024171595,0.03152517,0.0005299446],"about_ca_topic_score_codex":0.000026191776,"about_ca_topic_score_gemma":0.000021123611,"teacher_disagreement_score":0.12555172,"about_ca_system_score_codex":0.000020393614,"about_ca_system_score_gemma":0.00011394496,"threshold_uncertainty_score":0.99626505},"labels":[],"label_agreement":null},{"id":"W172656279","doi":"","title":"RALI: Automatic Weighting of Text Window Distances","year":2010,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing 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":"Université de Montréal","funders":"","keywords":"Window (computing); Weighting; Word (group theory); Computer science; SemEval; Artificial intelligence; Task (project management); Word-sense disambiguation; Natural language processing; Limit (mathematics); Pattern recognition (psychology); Mathematics","score_opus":0.007342829639935821,"score_gpt":0.2629071066691153,"score_spread":0.2555642770291795,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W172656279","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.12020698,0.00043586944,0.86370313,0.0020144065,0.0059308964,0.001008065,0.00012717827,0.00072890066,0.005844598],"genre_scores_gemma":[0.5657179,2.0075542e-7,0.43406412,0.000030503634,0.000110689165,0.0000036542651,0.0000046190357,0.0000051039906,0.0000632221],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9987249,0.000052723026,0.0004446669,0.00015075531,0.00048468527,0.0001423024],"domain_scores_gemma":[0.99573344,0.0015199608,0.0011990013,0.00019778433,0.0013288474,0.00002095252],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.00095961645,0.000091394206,0.00017988145,0.00006291317,0.00017570116,0.000056223707,0.00074225187,0.00007423581,0.0000010518211],"category_scores_gemma":[0.014913843,0.000072691,0.00010934787,0.00028764564,0.000042759646,0.00005150101,0.0001450622,0.00015987757,6.145148e-7],"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":[0.000005783078,0.00009038569,0.026137076,0.00025762583,0.000066211476,2.8790345e-7,0.00080458564,0.0041038333,0.004592216,0.95921355,0.0008483223,0.0038800987],"study_design_scores_gemma":[0.00037595112,0.00003908912,0.005801028,0.00022382021,0.0000412475,0.0000013045227,0.000017883824,0.35712346,0.032459196,0.6016226,0.0020901652,0.00020428603],"about_ca_topic_score_codex":0.000014735503,"about_ca_topic_score_gemma":0.000008912626,"teacher_disagreement_score":0.4455109,"about_ca_system_score_codex":0.00005792036,"about_ca_system_score_gemma":0.00011107143,"threshold_uncertainty_score":0.99338394},"labels":[],"label_agreement":null},{"id":"W2096565906","doi":"","title":"Alignment-Based Discriminative String Similarity","year":2007,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":54,"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":"Discriminative model; Coreference; Artificial intelligence; Computer science; Substring; String metric; Similarity (geometry); Character (mathematics); Natural language processing; Longest common subsequence problem; String (physics); Word (group theory); Pattern recognition (psychology); Transliteration; Heuristic; String searching algorithm; Mathematics; Pattern matching; Resolution (logic); Algorithm; Set (abstract data type)","score_opus":0.016287383324153314,"score_gpt":0.2953723765151392,"score_spread":0.2790849931909859,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2096565906","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0013557576,0.000058704565,0.9961205,0.000482007,0.00060191634,0.00021019786,0.000025375808,0.00014092556,0.0010045908],"genre_scores_gemma":[0.5739122,9.164423e-8,0.425802,0.00012844545,0.00009099129,0.0000024772137,0.000008945046,0.0000050153935,0.0000498364],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99868125,0.00005412094,0.00034492367,0.0001759115,0.0005483596,0.00019542238],"domain_scores_gemma":[0.9960378,0.001865358,0.0006803975,0.00015898091,0.0012288945,0.000028605242],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0017678421,0.0000954509,0.00012815729,0.00007461118,0.00023672702,0.00006207456,0.000590612,0.00006804157,3.179561e-7],"category_scores_gemma":[0.01115894,0.00008052967,0.000103889935,0.0002699594,0.00002553938,0.000040117542,0.00014011268,0.00011895022,4.31109e-7],"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":[0.000017939941,0.00012010388,0.014409838,0.000098395736,0.00005125125,9.584106e-7,0.000540001,0.030797329,0.00035170792,0.95085806,0.0009003515,0.0018540379],"study_design_scores_gemma":[0.0006834052,0.00007447046,0.0068984805,0.00020263564,0.00005455118,5.160374e-7,0.000046793055,0.27287737,0.0717092,0.64464957,0.0025021355,0.00030089868],"about_ca_topic_score_codex":0.00001443478,"about_ca_topic_score_gemma":0.0000064010246,"teacher_disagreement_score":0.57255644,"about_ca_system_score_codex":0.00031260337,"about_ca_system_score_gemma":0.000100616526,"threshold_uncertainty_score":0.9971705},"labels":[],"label_agreement":null},{"id":"W2096960653","doi":"","title":"Selecting Query Term Alternations for Web Search by Exploiting Query Contexts","year":2008,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Query expansion; Computer science; Bigram; Web query classification; Web search query; Information retrieval; Query optimization; Query language; Sargable; Selection (genetic algorithm); Term (time); RDF query language; Context (archaeology); Search engine; Word (group theory); Natural language processing; Artificial intelligence","score_opus":0.02470986278797762,"score_gpt":0.2848725438985707,"score_spread":0.2601626811105931,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2096960653","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.040617038,0.00010785738,0.9545274,0.0007071448,0.0017960784,0.00067067635,0.0004664855,0.0001501319,0.0009571853],"genre_scores_gemma":[0.84654385,0.000004053579,0.15213042,0.00015425643,0.0005445436,0.000034834648,0.00014426094,0.000017592945,0.00042618558],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99828464,0.00009652114,0.000493121,0.00027630007,0.0005761903,0.0002732062],"domain_scores_gemma":[0.9942311,0.0030169555,0.00066725374,0.00019247952,0.0018451163,0.00004707024],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009835619,0.00012590896,0.00019480758,0.000076912685,0.00094086805,0.0000880641,0.0006080313,0.00007112412,5.762763e-7],"category_scores_gemma":[0.0072204345,0.00011573298,0.00014881353,0.00022658404,0.000029502,0.00010224795,0.0002062924,0.00013016071,0.0000015039209],"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":[0.000098345736,0.0009309679,0.1645883,0.000487604,0.00063127215,0.0000034948177,0.0051474725,0.3884861,0.008253797,0.2907028,0.13093075,0.009739125],"study_design_scores_gemma":[0.001000111,0.000064618296,0.004994031,0.00013946502,0.000024043886,0.0000038004282,0.00003610508,0.96765447,0.002917867,0.010485737,0.012443712,0.00023602639],"about_ca_topic_score_codex":0.000029551922,"about_ca_topic_score_gemma":0.000003477346,"teacher_disagreement_score":0.8059268,"about_ca_system_score_codex":0.00020980033,"about_ca_system_score_gemma":0.0002481011,"threshold_uncertainty_score":0.8644056},"labels":[],"label_agreement":null},{"id":"W2097120204","doi":"","title":"Towards Robust Abstractive Multi-Document Summarization: A Caseframe Analysis of Centrality and Domain","year":2013,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Automatic summarization; Centrality; Computer science; Redundancy (engineering); Sentence; Natural language processing; Information retrieval; Domain (mathematical analysis); Artificial intelligence; Abstraction; Multi-document summarization","score_opus":0.020833600090225578,"score_gpt":0.26804744871907904,"score_spread":0.24721384862885346,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2097120204","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07560828,0.000022208247,0.9226575,0.0005519021,0.0003673323,0.00032162506,0.00005988277,0.000023340514,0.00038792164],"genre_scores_gemma":[0.6911564,8.515854e-7,0.30868894,0.000038686794,0.00004097073,0.0000068665627,0.000019506226,0.0000032700193,0.000044541972],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987148,0.00009074327,0.0004642716,0.00019220875,0.00040431568,0.00013367407],"domain_scores_gemma":[0.99678993,0.00078659813,0.00079917454,0.00016905554,0.001420578,0.00003464952],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00071653666,0.00008428045,0.00021881447,0.000094250754,0.00012458858,0.00006578765,0.0002668058,0.000057653247,0.0000022805086],"category_scores_gemma":[0.0034413957,0.00007498434,0.00012355615,0.00038058404,0.000025659303,0.000055760367,0.00012652262,0.00006738471,4.1006774e-7],"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":[0.0000025129557,0.000055457058,0.04691549,0.000032248015,0.00037092695,9.230399e-8,0.00082795514,0.8622157,0.000028148694,0.08913369,0.000049674607,0.0003681038],"study_design_scores_gemma":[0.00025753796,0.000011245903,0.14021084,0.000018406625,0.0001366933,1.3024021e-7,0.000046196037,0.8291977,0.0001235714,0.029828915,0.00009658014,0.000072150135],"about_ca_topic_score_codex":0.0003277853,"about_ca_topic_score_gemma":0.000023957164,"teacher_disagreement_score":0.6155481,"about_ca_system_score_codex":0.00016002743,"about_ca_system_score_gemma":0.000083022285,"threshold_uncertainty_score":0.41199207},"labels":[],"label_agreement":null},{"id":"W2099233844","doi":"","title":"Even the Abstract have Color: Consensus in Word-Colour Associations","year":2011,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Categorization, perception, and language","field":"Psychology","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Lexicon; Computer science; Word Association; Focus (optics); Component (thermodynamics); Crowdsourcing; Word (group theory); Resource (disambiguation); Key (lock); Product (mathematics); Association (psychology); Visualization; Semantics (computer science); Natural language processing; Coherence (philosophical gambling strategy); Quality (philosophy); Artificial intelligence; Linguistics; World Wide Web; Psychology; Mathematics","score_opus":0.045509232035914296,"score_gpt":0.31724088188839533,"score_spread":0.27173164985248105,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2099233844","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.85394514,0.00013078186,0.00095404446,0.0013192989,0.0058501507,0.0017287822,0.0008925904,0.00012731185,0.13505192],"genre_scores_gemma":[0.9951062,0.000001138254,0.0014160436,0.00018079912,0.00041058255,0.000040463652,0.00012648752,0.000021275413,0.002696968],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9984413,0.00019918545,0.00060132664,0.00018448154,0.00034114977,0.00023259174],"domain_scores_gemma":[0.9956473,0.0022217822,0.00094390736,0.00017822387,0.0009821827,0.000026567035],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015020925,0.000119940545,0.00019264019,0.00008159353,0.00031857562,0.000024751858,0.0002621796,0.0001473179,0.000091175476],"category_scores_gemma":[0.007912361,0.00009551436,0.00014959098,0.00023793605,0.00005829966,0.000011620363,0.00003513597,0.00018240427,0.0000213973],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011092604,0.0005455359,0.7695939,0.000029305142,0.00032354545,0.0000015806396,0.03389458,0.028280782,0.000044315017,0.15215185,0.014717986,0.00030570035],"study_design_scores_gemma":[0.0006115649,0.00002731178,0.9663225,0.000021173972,0.00007999379,8.5115244e-7,0.0015688378,0.0025207358,0.000026690168,0.024895785,0.0037977505,0.00012685196],"about_ca_topic_score_codex":0.0007931062,"about_ca_topic_score_gemma":0.00084725255,"teacher_disagreement_score":0.19672856,"about_ca_system_score_codex":0.00031329237,"about_ca_system_score_gemma":0.0001295204,"threshold_uncertainty_score":0.9472407},"labels":[],"label_agreement":null},{"id":"W2099242427","doi":"","title":"Entity-Based Local Coherence Modelling Using Topological Fields","year":2010,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing Techniques","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":"University of Toronto","funders":"","keywords":"Coherence (philosophical gambling strategy); Computer science; Sentence; Component (thermodynamics); Topology (electrical circuits); Grid; Natural language; Natural language processing; Field (mathematics); Artificial intelligence; Language model; Theoretical computer science; Mathematics; Physics; Pure mathematics","score_opus":0.022983696173067523,"score_gpt":0.29087224003836043,"score_spread":0.2678885438652929,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2099242427","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007265297,0.000024926523,0.9909387,0.0002767312,0.00096729526,0.00013667597,0.00001077659,0.00011230252,0.00026729118],"genre_scores_gemma":[0.524128,5.5260927e-8,0.47564524,0.00009700775,0.00010080228,0.0000022547692,0.0000035732735,0.0000034308039,0.000019619247],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99893224,0.00005208386,0.00028598704,0.00017930036,0.00039223142,0.00015814493],"domain_scores_gemma":[0.99725485,0.0009648943,0.00047054983,0.00017196378,0.0011111642,0.000026606689],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007391617,0.000088634326,0.00013046079,0.00004854589,0.00024799578,0.000081395745,0.00063521095,0.0001349428,0.0000012748064],"category_scores_gemma":[0.004431317,0.000073283605,0.00010002607,0.00019809726,0.00004875143,0.00003661312,0.0001278054,0.00025926,4.6224858e-7],"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":[0.000004110468,0.00003511839,0.0021994058,0.000027262653,0.000009978698,2.8887717e-7,0.000059359245,0.74653715,0.0002566007,0.25054228,0.00007948154,0.00024897227],"study_design_scores_gemma":[0.00010349038,0.000013692489,0.00004334348,0.000027229962,0.000010202422,4.5120538e-7,0.000002235628,0.7718817,0.0038882955,0.22369663,0.0002615603,0.00007117189],"about_ca_topic_score_codex":0.000043352615,"about_ca_topic_score_gemma":0.000009020398,"teacher_disagreement_score":0.5168627,"about_ca_system_score_codex":0.000088198314,"about_ca_system_score_gemma":0.00015939475,"threshold_uncertainty_score":0.5305021},"labels":[],"label_agreement":null},{"id":"W2110750141","doi":"","title":"Automatic detection of deception in child-produced speech using syntactic complexity features","year":2013,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Deception detection and forensic psychology","field":"Psychology","cited_by":27,"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":"Utterance; Deception; Sentence; Support vector machine; Computer science; Set (abstract data type); Artificial intelligence; Random forest; Measure (data warehouse); Natural language processing; Speech recognition; Pattern recognition (psychology); Machine learning; Psychology; Data mining; Social psychology","score_opus":0.03243081732255438,"score_gpt":0.32348086160658635,"score_spread":0.291050044284032,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2110750141","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9874063,0.000029081275,0.0033654391,0.00022339735,0.004065541,0.0006868414,0.000031814525,0.000051468076,0.0041401307],"genre_scores_gemma":[0.99271023,5.4885913e-7,0.0067346967,0.00007188527,0.00025940925,0.000022522487,0.000019657446,0.000017665516,0.00016339267],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99837077,0.00024619655,0.0006561288,0.00021529273,0.0003201128,0.00019152548],"domain_scores_gemma":[0.99710584,0.0006224396,0.0011225501,0.00018582062,0.0009360791,0.00002724575],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00076431915,0.000120048666,0.00026009965,0.00020456206,0.00014568996,0.000018042654,0.00015332166,0.0001507796,0.000078096426],"category_scores_gemma":[0.0048761037,0.00011336033,0.00014423399,0.00038929953,0.00007449533,0.000028219076,0.000031196185,0.00018435903,0.0000176088],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009778397,0.0045192204,0.45084798,0.0013246821,0.002258034,0.000003843327,0.013691176,0.25774086,0.054701015,0.12344754,0.011936535,0.07855128],"study_design_scores_gemma":[0.0017912819,0.00013629234,0.8587577,0.00016116227,0.000105265884,0.0000237106,0.00048311258,0.08702328,0.0037058936,0.047079165,0.0004890181,0.00024416347],"about_ca_topic_score_codex":0.00038543882,"about_ca_topic_score_gemma":0.00006906747,"teacher_disagreement_score":0.4079097,"about_ca_system_score_codex":0.000251493,"about_ca_system_score_gemma":0.000029643921,"threshold_uncertainty_score":0.5837504},"labels":[],"label_agreement":null},{"id":"W2115791615","doi":"","title":"Deep Learning for NLP (without Magic)","year":2012,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":132,"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 de Montréal","funders":"","keywords":"Computer science; Artificial intelligence; Feature engineering; Deep learning; Artificial neural network; Paraphrase; Machine learning; Sentiment analysis; Natural language processing; Feature (linguistics); Focus (optics); MAGIC (telescope); Language model","score_opus":0.022720816901939583,"score_gpt":0.2786095360606215,"score_spread":0.25588871915868194,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2115791615","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00284011,0.000064126776,0.9922226,0.00031845408,0.0019480959,0.00028112184,0.0000055692776,0.000066719746,0.0022532044],"genre_scores_gemma":[0.68956333,4.0953213e-7,0.30924264,0.000085898595,0.00062820804,0.000010662781,0.000010335024,0.00000916903,0.00044937144],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99881715,0.000072604926,0.00034101956,0.0001474626,0.0003618822,0.00025985634],"domain_scores_gemma":[0.996572,0.0016350761,0.0005976237,0.00014093112,0.0010150798,0.000039301005],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0013860312,0.00008897814,0.00014659131,0.000044668974,0.00027157916,0.00005278375,0.00040214596,0.00006137528,5.789303e-7],"category_scores_gemma":[0.0136610055,0.00007997827,0.00013604989,0.000120516575,0.000011155771,0.000051872637,0.000118646516,0.00009763151,0.0000030596264],"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":[0.000004978131,0.00004159337,0.08693471,0.00004713894,0.000041629795,1.5278388e-8,0.00073173386,0.65992665,0.000018944711,0.25088143,0.00029006353,0.0010811136],"study_design_scores_gemma":[0.00031440836,0.000017991026,0.005148931,0.00002381664,0.000025926396,3.424852e-7,0.00001970639,0.9410582,0.00013785546,0.037343107,0.01581066,0.00009906648],"about_ca_topic_score_codex":0.000006018312,"about_ca_topic_score_gemma":0.0000013187556,"teacher_disagreement_score":0.68672323,"about_ca_system_score_codex":0.00015641903,"about_ca_system_score_gemma":0.000050875416,"threshold_uncertainty_score":0.9946473},"labels":[],"label_agreement":null},{"id":"W2120158336","doi":"","title":"Cross Lingual Adaptation: An Experiment on Sentiment Classifications","year":2010,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":87,"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; Machine translation; Natural language processing; Task (project management); Artificial intelligence; Representation (politics); Adaptation (eye); Key (lock); Sentiment analysis; Natural language; Translation (biology); Parallel corpora; Noise (video); Machine learning","score_opus":0.03508279994780869,"score_gpt":0.3290336885755691,"score_spread":0.2939508886277604,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2120158336","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.46205848,0.000057012134,0.49462038,0.0034554058,0.018796206,0.0014403249,0.00015038087,0.00040969596,0.019012097],"genre_scores_gemma":[0.8577216,3.676394e-7,0.14104204,0.00012200358,0.0004380752,0.000017236815,0.00006025962,0.000009478051,0.0005889419],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983418,0.00007446346,0.00047754997,0.00027715624,0.00065950496,0.00016951207],"domain_scores_gemma":[0.9967967,0.00076824083,0.00074803224,0.0003326317,0.0013017061,0.000052638163],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000953677,0.00011426882,0.00014469949,0.00008717085,0.00047690404,0.00019989448,0.0005706583,0.00007237362,0.0000069918087],"category_scores_gemma":[0.0027225649,0.000100698046,0.0001619396,0.00025091955,0.00003070436,0.000057867754,0.00008577656,0.00014229308,0.000007654415],"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":[0.00001099779,0.00032028396,0.011539023,0.000009576335,0.000100250785,1.4298283e-7,0.0016737622,0.46940854,0.0008166462,0.5146855,0.00058479677,0.00085046596],"study_design_scores_gemma":[0.0004999574,0.00007538285,0.012939735,0.000024214858,0.000031635795,3.0792125e-7,0.00010215941,0.9576163,0.006624462,0.009493971,0.012421159,0.00017066805],"about_ca_topic_score_codex":0.000010177383,"about_ca_topic_score_gemma":0.000006983507,"teacher_disagreement_score":0.5051915,"about_ca_system_score_codex":0.00009750696,"about_ca_system_score_gemma":0.0001218683,"threshold_uncertainty_score":0.4106347},"labels":[],"label_agreement":null},{"id":"W2124618303","doi":"","title":"Substring-Based Transliteration","year":2007,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":61,"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":"Transliteration; Substring; Computer science; Margin (machine learning); Artificial intelligence; Word (group theory); Natural language processing; Machine translation; Speech recognition; Programming language; Machine learning; Data structure; Mathematics","score_opus":0.011120757289096535,"score_gpt":0.27550062212974663,"score_spread":0.26437986484065007,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2124618303","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0019628261,0.00008958159,0.995899,0.00039367148,0.00061130576,0.00018248848,0.00001247981,0.0001530553,0.00069554703],"genre_scores_gemma":[0.58371013,7.615142e-8,0.41600123,0.00011674494,0.00010912091,0.00000223083,0.0000112926145,0.0000045754664,0.00004457874],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99887836,0.0000380027,0.0003499077,0.00014433054,0.00043330155,0.00015609646],"domain_scores_gemma":[0.9968214,0.0012403098,0.0005063826,0.00014064273,0.0012687802,0.000022466327],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015597423,0.000078270066,0.00010295359,0.00008023988,0.00018177523,0.000069344394,0.00046113823,0.000069058,2.9669454e-7],"category_scores_gemma":[0.0055921483,0.00006698395,0.00009981616,0.00030801824,0.000015073917,0.000039277,0.000036359572,0.00009306314,5.6671564e-7],"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":[0.000031241278,0.0001176552,0.020605722,0.00013530343,0.000044191438,8.3249387e-7,0.0006276225,0.12318338,0.00090577174,0.8503111,0.0008605451,0.0031766281],"study_design_scores_gemma":[0.00086094916,0.00007698693,0.0076336605,0.00018949386,0.000041530162,9.210525e-7,0.000009947464,0.52747494,0.069823444,0.3864295,0.0071574063,0.00030124965],"about_ca_topic_score_codex":0.0000066511425,"about_ca_topic_score_gemma":0.0000060941093,"teacher_disagreement_score":0.58174735,"about_ca_system_score_codex":0.00016478717,"about_ca_system_score_gemma":0.0000920859,"threshold_uncertainty_score":0.66947275},"labels":[],"label_agreement":null},{"id":"W2146546639","doi":"","title":"Learning Bigrams from Unigrams","year":2008,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":9,"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":"Bigram; Perplexity; Computer science; Language model; Natural language processing; Artificial intelligence; Word (group theory); Trigram; Oracle; Set (abstract data type); Speech recognition; Mathematics; Programming language","score_opus":0.022393896773708673,"score_gpt":0.24455504663165106,"score_spread":0.22216114985794239,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2146546639","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.11523348,0.00005555198,0.8779747,0.0006009102,0.0021560788,0.00022314905,0.00002382695,0.0001749083,0.0035573815],"genre_scores_gemma":[0.8257209,0.000001755787,0.17341262,0.00008479365,0.00032732467,0.0000039644956,0.000020497215,0.00000770738,0.00042040797],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987119,0.000093120514,0.00034620968,0.00019545286,0.0004927404,0.00016061398],"domain_scores_gemma":[0.99693763,0.0014291166,0.0005488979,0.00018082422,0.00087436027,0.00002917531],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.00049971155,0.0000860983,0.00014560788,0.00004260376,0.00038542724,0.00003823262,0.0005293936,0.000063261636,6.4847524e-7],"category_scores_gemma":[0.008465158,0.00007786802,0.00012196034,0.00022084804,0.000021309625,0.000029875418,0.00014194785,0.00013779238,0.000003884287],"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":[0.0000033777505,0.00003773203,0.085734256,0.000009505736,0.000050908402,4.885418e-7,0.0011339646,0.8445374,0.000039620303,0.067160204,0.0004813809,0.0008112015],"study_design_scores_gemma":[0.00033147313,0.00002674421,0.01265597,0.00003600351,0.000015076614,6.5240187e-7,0.000021673775,0.91942745,0.0002571492,0.058001366,0.009110992,0.00011543031],"about_ca_topic_score_codex":0.00010420244,"about_ca_topic_score_gemma":0.000005143499,"teacher_disagreement_score":0.7104874,"about_ca_system_score_codex":0.00012643149,"about_ca_system_score_gemma":0.00010708113,"threshold_uncertainty_score":0.999887},"labels":[],"label_agreement":null},{"id":"W2161427841","doi":"","title":"Summarizing Emails with Conversational Cohesion and Subjectivity","year":2008,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":66,"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":"Automatic summarization; Cohesion (chemistry); Computer science; PageRank; Cosine similarity; Natural language processing; Artificial intelligence; Graph; Sentence; Empirical research; Information retrieval; Subjectivity; Theoretical computer science; Pattern recognition (psychology); Mathematics","score_opus":0.013367564952236293,"score_gpt":0.24368251737275914,"score_spread":0.23031495242052286,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2161427841","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07226845,0.000032120835,0.9253579,0.00046629237,0.00021114189,0.0002940832,0.000018889226,0.0001294875,0.0012216306],"genre_scores_gemma":[0.76673037,0.000002116247,0.23304176,0.00005520412,0.00006129601,0.0000057758175,0.000008838416,0.000005578057,0.000089050984],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.998931,0.00006599862,0.0002368311,0.000188357,0.00045778154,0.000120025135],"domain_scores_gemma":[0.99629694,0.0018564148,0.00056423567,0.000132519,0.0011248821,0.000025014782],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005475976,0.00008702647,0.00015575516,0.000064769236,0.00036617054,0.000028048758,0.00023278751,0.00004476954,2.6889884e-7],"category_scores_gemma":[0.0047084196,0.00007067721,0.000061153245,0.00024367787,0.00005630504,0.00006903361,0.00010414114,0.00007745086,4.074625e-7],"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":[0.000036087247,0.00010572252,0.6116477,0.000052736876,0.0001581404,0.000001944237,0.0006961052,0.070892565,0.00025738848,0.31483656,0.00085511,0.00045989355],"study_design_scores_gemma":[0.0013151745,0.00020750951,0.2991844,0.0001450173,0.00010556375,0.000018357006,0.00003321941,0.51260924,0.0061377236,0.1761356,0.003662907,0.00044528788],"about_ca_topic_score_codex":0.000016009984,"about_ca_topic_score_gemma":0.000007850972,"teacher_disagreement_score":0.69446194,"about_ca_system_score_codex":0.00014413641,"about_ca_system_score_gemma":0.00011173052,"threshold_uncertainty_score":0.5636758},"labels":[],"label_agreement":null},{"id":"W2166202273","doi":"","title":"Bootstrapping a Stochastic Transducer for Arabic-English Transliteration Extraction","year":2007,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":24,"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":"Bootstrapping (finance); Computer science; Transducer; Scripting language; Artificial intelligence; Natural language processing; Metric (unit); Transliteration; Task (project management); Speech recognition; Similarity (geometry); Arabic; Pattern recognition (psychology); Programming language; Linguistics; Acoustics; Engineering; Mathematics","score_opus":0.014098183086053236,"score_gpt":0.2948093699806726,"score_spread":0.2807111868946193,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2166202273","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0012924362,0.000113715876,0.995942,0.0002525005,0.0013596106,0.00054945826,0.000035631234,0.0001889923,0.0002656652],"genre_scores_gemma":[0.6566775,2.2709048e-7,0.34271806,0.000071017304,0.00041862216,0.000017351269,0.000026316946,0.000009989126,0.000060898234],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99858516,0.00004039236,0.0005023945,0.00022550902,0.00042144433,0.00022509026],"domain_scores_gemma":[0.994791,0.0020106903,0.0006373667,0.00013603322,0.0023940871,0.000030874544],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0018438789,0.000118563454,0.00015745728,0.00010875762,0.00029472046,0.00010908743,0.00038340295,0.00011033245,3.237761e-7],"category_scores_gemma":[0.008877081,0.00010779967,0.00016911281,0.00027231232,0.00001818872,0.00011097609,0.000021418786,0.00013221813,2.0791454e-7],"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":[0.00015807965,0.00026011426,0.0010011632,0.00045522946,0.00017840212,5.42667e-7,0.0070718895,0.2793137,0.004677092,0.69701064,0.0010730922,0.008800037],"study_design_scores_gemma":[0.0011400672,0.00012264849,0.0013343917,0.0002758622,0.0001040861,0.000001910839,0.000068657566,0.6004904,0.014480793,0.3775532,0.004064955,0.00036304828],"about_ca_topic_score_codex":0.0000053332337,"about_ca_topic_score_gemma":0.0000071399686,"teacher_disagreement_score":0.6553851,"about_ca_system_score_codex":0.00021463909,"about_ca_system_score_gemma":0.00009423139,"threshold_uncertainty_score":0.99947155},"labels":[],"label_agreement":null},{"id":"W2250341893","doi":"","title":"Mapping Source to Target Strings without Alignment by Analogical Learning: A Case Study with Transliteration","year":2013,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Transliteration; Computer science; Natural language processing; Phrase; Analogy; Artificial intelligence; Machine translation; Task (project management); Translation (biology); Rule-based machine translation; Machine learning; Linguistics; Engineering","score_opus":0.010253602251457633,"score_gpt":0.2590147617676204,"score_spread":0.24876115951616276,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2250341893","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1377992,0.000032686712,0.8605763,0.0004071346,0.000097812954,0.00082435575,0.000007505376,0.00015684245,0.000098179575],"genre_scores_gemma":[0.63879347,5.5239123e-8,0.36080793,0.00010480315,0.000052799325,0.000041877258,0.000009286749,0.0000081661165,0.00018160255],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984739,0.00014360293,0.00036602176,0.00026680966,0.0005601733,0.0001895214],"domain_scores_gemma":[0.99768615,0.0004906527,0.00048784082,0.00013564106,0.0011476865,0.000052014828],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00075772597,0.0001320988,0.00018100748,0.0000800094,0.000333396,0.00022630613,0.00037021225,0.000052134445,0.0000010001097],"category_scores_gemma":[0.0022520015,0.0000987197,0.000054996708,0.00035509298,0.000013555047,0.000071257826,0.00009918672,0.00015012469,0.000001229178],"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":[0.00006214707,0.0010712506,0.3736756,0.00019791708,0.00044182557,0.000037896938,0.035418876,0.5595773,0.0017056891,0.018035013,0.0042998716,0.0054765944],"study_design_scores_gemma":[0.002136267,0.0015519097,0.0032711863,0.00032544782,0.00009856865,0.000100104786,0.0024405122,0.94035465,0.0034606163,0.035301387,0.010066422,0.0008929443],"about_ca_topic_score_codex":0.00015232844,"about_ca_topic_score_gemma":0.0000063031007,"teacher_disagreement_score":0.50099427,"about_ca_system_score_codex":0.00017628665,"about_ca_system_score_gemma":0.00005026969,"threshold_uncertainty_score":0.40256727},"labels":[],"label_agreement":null},{"id":"W2250669704","doi":"","title":"Probabilistic Domain Modelling With Contextualized Distributional Semantic Vectors","year":2013,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Topic Modeling","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 Toronto","funders":"","keywords":"Automatic summarization; Computer science; Distributional semantics; Artificial intelligence; Probabilistic logic; Natural language processing; Domain (mathematical analysis); Semantics (computer science); Generative grammar; Generative model; Hidden Markov model; Semantic space; Machine learning; Semantic similarity; Mathematics","score_opus":0.01520811403659151,"score_gpt":0.23146867420578987,"score_spread":0.21626056016919837,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2250669704","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.041645568,0.000016078231,0.955977,0.00063292566,0.00048623927,0.00051950914,0.00003231192,0.00006673187,0.00062364136],"genre_scores_gemma":[0.7730668,2.024573e-7,0.22655126,0.000059481157,0.00014372575,0.000033532564,0.000027064263,0.000009068507,0.00010882944],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983521,0.00009610302,0.00045259757,0.00025053052,0.0006175733,0.00023106816],"domain_scores_gemma":[0.9957606,0.0014231243,0.00062814314,0.00022109784,0.0019228862,0.0000441257],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006579137,0.00012824293,0.0002020068,0.000045363733,0.00027671852,0.00011134578,0.00048455223,0.00006021373,0.0000017703006],"category_scores_gemma":[0.0024292136,0.000098990895,0.00010374818,0.00022331273,0.000037426627,0.0000587479,0.00010549946,0.000106905936,0.000004570278],"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":[0.000003902109,0.000028439243,0.0028328474,0.000027270167,0.000034166565,8.008526e-8,0.00015146882,0.6463704,0.000010645942,0.35036144,0.00014825948,0.000031061718],"study_design_scores_gemma":[0.00043051472,0.000024634697,0.0011696547,0.00006612679,0.000018853701,7.859666e-7,0.000017081837,0.7890018,0.000046834484,0.20835325,0.0007629225,0.00010753758],"about_ca_topic_score_codex":0.00006228605,"about_ca_topic_score_gemma":0.0000044207623,"teacher_disagreement_score":0.73142123,"about_ca_system_score_codex":0.00025154327,"about_ca_system_score_gemma":0.00017268205,"threshold_uncertainty_score":0.40367317},"labels":[],"label_agreement":null},{"id":"W3036997889","doi":"","title":"Using Attention-based Bidirectional LSTM to Identify Different Categories of Offensive Language Directed Toward Female Celebrities","year":2019,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Hate Speech and Cyberbullying Detection","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":"Dalhousie University","funders":"","keywords":"Offensive; Harassment; Perspective (graphical); Margin (machine learning); Computer science; Social media; Artificial intelligence; Natural language processing; Psychology; Machine learning; Social psychology; World Wide Web; Mathematics","score_opus":0.02115456455403126,"score_gpt":0.2868773790427908,"score_spread":0.26572281448875956,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3036997889","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.90971065,0.000028518663,0.086315796,0.00016521139,0.0026505745,0.0004220596,0.00010115289,0.0000979958,0.00050803483],"genre_scores_gemma":[0.9697,2.824295e-7,0.029507713,0.000057852736,0.0001609786,0.000008374888,0.000038585025,0.000011841786,0.00051439955],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99852484,0.00010653571,0.00040151205,0.00021342933,0.00058186514,0.00017182714],"domain_scores_gemma":[0.9966794,0.00080036593,0.00060698134,0.00016320425,0.0017150901,0.00003495542],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00042580772,0.000116836476,0.0002267739,0.00013365848,0.00016001615,0.000063050225,0.0002922363,0.00006488995,0.0000033612926],"category_scores_gemma":[0.0028630563,0.00010753156,0.00017303978,0.0003300072,0.000020440722,0.000029809378,0.000090876514,0.00007995607,0.0000051873344],"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":[0.000096981545,0.00029406385,0.22764732,0.00054487196,0.0003424702,9.4592514e-7,0.0027544547,0.7073027,0.035525665,0.024046553,0.0009538616,0.0004901367],"study_design_scores_gemma":[0.0014613152,0.00023567709,0.19833441,0.00047818696,0.00013945416,0.0000022175693,0.0005803712,0.6941553,0.094195224,0.0088461125,0.0010542754,0.00051749544],"about_ca_topic_score_codex":0.00017814244,"about_ca_topic_score_gemma":0.000013974095,"teacher_disagreement_score":0.059989315,"about_ca_system_score_codex":0.00021691383,"about_ca_system_score_gemma":0.00012757543,"threshold_uncertainty_score":0.43850097},"labels":[],"label_agreement":null},{"id":"W3037566932","doi":"","title":"Augmenting Named Entity Recognition with Commonsense Knowledge","year":2019,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Topic Modeling","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":"Université du Québec à Montréal; Université du Québec","funders":"","keywords":"Commonsense knowledge; Computer science; Natural language processing; Artificial intelligence; Named-entity recognition; Ambiguity; Knowledge base; Natural language understanding; Natural language; Word embedding; Question answering; Embedding; Task (project management)","score_opus":0.019435820131858655,"score_gpt":0.25008630713282126,"score_spread":0.2306504870009626,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3037566932","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.40548486,0.00004651145,0.57144904,0.0006029322,0.0033726383,0.0008809659,0.000051092426,0.00017947407,0.01793248],"genre_scores_gemma":[0.869102,2.9109734e-7,0.13017565,0.000066016204,0.00014929961,0.000006337739,0.000019744817,0.000008584409,0.000472073],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988395,0.00010199775,0.00032210455,0.00020098944,0.00037166494,0.00016374813],"domain_scores_gemma":[0.9965599,0.0012495762,0.00062899024,0.00020509325,0.0013321687,0.000024233137],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009367459,0.00008918781,0.0001490604,0.000054543132,0.00018365904,0.00006880239,0.00034822445,0.00005039432,0.0000013519295],"category_scores_gemma":[0.0029181738,0.000075716685,0.00008362982,0.00020695633,0.0000130623175,0.000043737324,0.00013035412,0.00010151642,0.000013124819],"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":[0.000052117935,0.0003186531,0.29763263,0.00036577237,0.00028397594,6.1649433e-7,0.0023249076,0.50976354,0.00026800792,0.18337485,0.0011107724,0.0045041563],"study_design_scores_gemma":[0.00080216036,0.00004953228,0.0084740035,0.00018189488,0.000036916674,0.0000010850163,0.000034812416,0.95467204,0.00053636637,0.03263614,0.0024128826,0.00016218325],"about_ca_topic_score_codex":0.000020986388,"about_ca_topic_score_gemma":0.000013923816,"teacher_disagreement_score":0.46361715,"about_ca_system_score_codex":0.0001693853,"about_ca_system_score_gemma":0.00012083318,"threshold_uncertainty_score":0.34935373},"labels":[],"label_agreement":null},{"id":"W3088355435","doi":"","title":"PACTE: A colloaborative platform for textual annotation","year":2017,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Polytechnique Montréal; École de Technologie Supérieure","funders":"","keywords":"Annotation; Computer science; Natural language processing; Artificial intelligence; Information retrieval; World Wide Web","score_opus":0.021316635145724953,"score_gpt":0.31386733643635356,"score_spread":0.2925507012906286,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3088355435","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0018531426,0.000069834,0.99234056,0.001433612,0.0011902809,0.00081910385,0.00024342448,0.00015656617,0.0018934518],"genre_scores_gemma":[0.571179,3.3650858e-7,0.42828846,0.00006822185,0.00019469058,0.000025797299,0.000025438458,0.0000067955007,0.00021127365],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9989377,0.000027483267,0.00031225436,0.00018618458,0.00037296925,0.00016336143],"domain_scores_gemma":[0.9940622,0.00135009,0.0015041117,0.00026694778,0.0027930527,0.000023606313],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0008937849,0.000100777834,0.00015938665,0.000049634735,0.0009134368,0.00028375126,0.00090973295,0.000084366686,1.9447315e-7],"category_scores_gemma":[0.033238985,0.0000841582,0.00011217836,0.00010110387,0.000038398768,0.00013437288,0.00015872768,0.00008584297,5.8281586e-7],"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":[0.000032430773,0.0000539184,0.003790161,0.00008113184,0.00006872568,1.787504e-7,0.00072388485,0.0047906614,0.0001878338,0.98339605,0.004127721,0.0027473106],"study_design_scores_gemma":[0.00085417915,0.00010527503,0.0036094738,0.00013292785,0.00004392448,5.31133e-7,0.000040236457,0.28937694,0.0073010954,0.6907694,0.0075493148,0.00021670928],"about_ca_topic_score_codex":0.000015533054,"about_ca_topic_score_gemma":0.000010373675,"teacher_disagreement_score":0.56932586,"about_ca_system_score_codex":0.00019421577,"about_ca_system_score_gemma":0.00018675307,"threshold_uncertainty_score":0.9749045},"labels":[],"label_agreement":null},{"id":"W3177369674","doi":"","title":"AdElectra - Effective Adversarial Electra","year":2021,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Classical Philosophy and Thought","field":"Arts and Humanities","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":"Université de Montréal","funders":"","keywords":"Adversarial system; Computer science; Artificial intelligence","score_opus":0.014771906817406777,"score_gpt":0.23645921032382644,"score_spread":0.22168730350641966,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3177369674","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04701041,0.0004169041,0.004654868,0.008312715,0.019939214,0.0016886316,0.001337947,0.00027317568,0.91636616],"genre_scores_gemma":[0.99373925,6.066202e-7,0.0009108978,0.00021400556,0.0030679079,0.000010210203,0.00007563183,0.000011634055,0.0019698716],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99914324,0.0000869646,0.00023727631,0.00012050795,0.00027935937,0.00013263772],"domain_scores_gemma":[0.9965626,0.0016736125,0.00030411364,0.000067425215,0.0013731776,0.000019053337],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0002836641,0.00007630795,0.00014146534,0.000021497806,0.00034985368,0.00005099488,0.000096861135,0.000043031807,0.000018581126],"category_scores_gemma":[0.008463089,0.000065781824,0.00017818136,0.000062076804,0.00004295041,0.000014962327,0.000028508557,0.00011426672,0.0000047417825],"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":[0.000024897005,0.000073796604,0.00037413713,0.00002714288,0.000113230955,2.8291768e-7,0.000997158,0.0065784804,0.000046589917,0.98894805,0.002696969,0.00011925759],"study_design_scores_gemma":[0.0007690993,0.00007110136,0.00090037344,0.000057504945,0.00015154417,3.779428e-7,0.000077443074,0.008257603,0.0024462466,0.7150604,0.27205768,0.0001506272],"about_ca_topic_score_codex":0.0000097169,"about_ca_topic_score_gemma":0.000032817043,"teacher_disagreement_score":0.9467288,"about_ca_system_score_codex":0.00012651867,"about_ca_system_score_gemma":0.000097816424,"threshold_uncertainty_score":0.9998891},"labels":[],"label_agreement":null},{"id":"W3197573382","doi":"","title":"Proceedings of the ACL-IJCNLP 2021 Student Research Workshop.","year":2021,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Dental Research and COVID-19","field":"Dentistry","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":"McGill University","funders":"","keywords":"Computer science; Mathematics education; Data science; Psychology","score_opus":0.07961984896349812,"score_gpt":0.418582325272663,"score_spread":0.3389624763091649,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3197573382","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9509982,0.0003150201,0.00039852483,0.0023355302,0.0035915764,0.0010415912,0.00037925915,0.00003016648,0.040910173],"genre_scores_gemma":[0.98468816,0.000003914277,0.0013345539,0.000083463885,0.00042649388,0.000020085909,0.000023852856,0.000016570355,0.013402893],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99675125,0.00015478938,0.00046977022,0.00021079431,0.002130492,0.00028290862],"domain_scores_gemma":[0.98821074,0.0043014437,0.00056379696,0.00016794716,0.006698805,0.000057271478],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0025011466,0.0000886737,0.00018214504,0.00005949704,0.0004258574,0.000113656715,0.0005966373,0.0000915498,0.000025403551],"category_scores_gemma":[0.09325187,0.00006768681,0.00023141169,0.0007247775,0.00008763413,0.000020877273,0.000539073,0.0003168211,0.000012101303],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00015812322,0.0008403764,0.5191537,0.0005710551,0.00055458717,0.000004955144,0.001413265,0.027270047,0.0023808412,0.3045885,0.1426744,0.00039017884],"study_design_scores_gemma":[0.0036974985,0.00012829286,0.49832004,0.0010782361,0.00024122062,0.000009013801,0.004793567,0.013286651,0.021624614,0.3754712,0.08088764,0.00046203256],"about_ca_topic_score_codex":0.000025583771,"about_ca_topic_score_gemma":0.0000510734,"teacher_disagreement_score":0.09075072,"about_ca_system_score_codex":0.00043461405,"about_ca_system_score_gemma":0.00045900227,"threshold_uncertainty_score":0.91438603},"labels":[],"label_agreement":null},{"id":"W4531093","doi":"","title":"Extraction of Disease-Treatment Semantic Relations from Biomedical Sentences","year":2010,"lang":"en","type":"article","venue":"Meeting of the Association for Computational Linguistics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","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 Ottawa","funders":"","keywords":"Relationship extraction; Relation (database); Semantic relation; Measure (data warehouse); Computer science; Natural language processing; Focus (optics); Artificial intelligence; Semantics (computer science); Information retrieval; Information extraction; Data mining; Medicine; Programming language","score_opus":0.012831567719147929,"score_gpt":0.2901257257106199,"score_spread":0.27729415799147195,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4531093","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.98756146,0.000084359206,0.008876423,0.00072076335,0.0017898527,0.00016353033,0.00042099893,0.000014287091,0.0003683304],"genre_scores_gemma":[0.9793823,0.000004108637,0.019522127,0.000018999022,0.00046093264,0.000006162283,0.00033160904,0.0000057090097,0.00026810102],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9993033,0.000037954484,0.00024975475,0.00011785975,0.0002118608,0.000079308986],"domain_scores_gemma":[0.99858713,0.0005057341,0.00038495695,0.00009916763,0.0003896368,0.00003334764],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.00022212903,0.00006168886,0.000098632794,0.00002261564,0.000096628006,0.0000064208484,0.000100264704,0.00010012205,0.000003110756],"category_scores_gemma":[0.009224476,0.000047052043,0.000106999774,0.00005967411,0.000086035456,7.028161e-7,0.000028616078,0.000054518856,9.258395e-7],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00032468102,0.0011788643,0.7964888,0.00011201619,0.00088113954,7.5073825e-7,0.0004954164,0.023076769,0.15443574,0.008617773,0.009253093,0.0051349555],"study_design_scores_gemma":[0.0025130953,0.00067723385,0.63303536,0.0002006814,0.0006411771,0.0000014118469,0.00029803082,0.058604345,0.031720217,0.051068723,0.22078995,0.00044973995],"about_ca_topic_score_codex":0.00005089888,"about_ca_topic_score_gemma":0.000026928288,"teacher_disagreement_score":0.21153685,"about_ca_system_score_codex":0.000021687449,"about_ca_system_score_gemma":0.00011754165,"threshold_uncertainty_score":0.99912125},"labels":[],"label_agreement":null}]}