{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":61,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":61,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"66b9019c05d7","filters":{"venue":"Computational Linguistics"}},"results":[{"id":"W2084046180","doi":"10.1162/coli_a_00049","title":"Lexicon-Based Methods for Sentiment Analysis","year":2011,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":3255,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia; University of Toronto; Simon Fraser University","funders":"Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Sentiment analysis; Lexicon; Polarity (international relations); Natural language processing; Artificial intelligence; Negation; Consistency (knowledge bases); Orientation (vector space); Word (group theory); Process (computing); Reliability (semiconductor); Linguistics; Mathematics","authors":[{"name":"Maite Taboada","is_ca":true},{"name":"Julian Brooke","is_ca":true},{"name":"Milan Tofiloski","is_ca":true},{"name":"Kimberly Voll","is_ca":true},{"name":"Manfred Stede","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08804081832030412,"gpt":0.3846883079456333,"spread":0.2966474896253292,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003045297,0.002071719,0.001542029,0.00900719,0.001419174,0.005393651,0.001823081,0.001376727,0.01292896],"category_scores_gemma":[0.01324147,0.0008683114,0.001970728,0.007577281,0.001017084,0.003788876,0.002279578,0.002361283,0.01427426],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001146294,"about_ca_system_score_gemma":0.001608836,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00181939,"about_ca_topic_score_gemma":0.002607475,"domain_scores_codex":[0.9958007,0.001542527,0.0004785838,0.0006362508,0.001407892,0.0001340198],"domain_scores_gemma":[0.9944744,0.002982128,0.0004846158,0.0007069867,0.001249119,0.0001027838],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001124247,0.0001679984,0.001888367,0.001421571,0.0004368785,0.0002571791,0.0006165473,0.008685959,0.01040488,0.1033624,0.06512037,0.8075254],"study_design_scores_gemma":[0.0001185435,0.00009532741,0.003233535,0.00053438,0.0002462578,0.0006830376,0.000772229,0.3646254,0.009748445,0.3800164,0.2397301,0.0001962019],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001288293,0.001431312,0.9831582,0.0003887857,0.0003100782,0.000406241,0.002149259,0.003022652,0.007845165],"genre_scores_gemma":[0.04533688,0.002230737,0.936375,0.0003902522,0.0005900786,0.001292801,0.007213166,0.000841814,0.005729201],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01292896,"threshold_uncertainty_score":0.04325169,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2136930489","doi":"10.1162/coli.2006.32.1.13","title":"Evaluating WordNet-based Measures of Lexical Semantic Relatedness","year":2006,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":1414,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"University of Toronto; University of Pennsylvania","keywords":"WordNet; Computer science; Semantic similarity; Natural language processing; Spelling; Artificial intelligence; Lexical database; Proxy (statistics); Similarity (geometry); Information retrieval; Measure (data warehouse); Linguistics; Machine learning; Data mining","authors":[{"name":"Alexander Budanitsky","is_ca":true},{"name":"Graeme Hirst","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0587311279945921,"gpt":0.346960390515805,"spread":0.2882292625212128,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008047068,0.001054583,0.001285083,0.01680863,0.0008150082,0.001834905,0.001153538,0.00157927,0.001562804],"category_scores_gemma":[0.03903022,0.000245351,0.000606706,0.009133915,0.001137412,0.00484967,0.001923252,0.0006586149,0.0006350853],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008571837,"about_ca_system_score_gemma":0.000754552,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001847559,"about_ca_topic_score_gemma":0.002994315,"domain_scores_codex":[0.9917888,0.003521932,0.001066689,0.0009157127,0.002435959,0.000270954],"domain_scores_gemma":[0.9540678,0.03282277,0.004065515,0.00266195,0.005328716,0.001053136],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002853965,0.001965601,0.2927179,0.001979236,0.00191266,0.0006222777,0.002096339,0.09737049,0.03108677,0.01492058,0.005496279,0.5469779],"study_design_scores_gemma":[0.0003842216,0.004384023,0.2630754,0.0002006437,0.0007751023,0.001771239,0.002096198,0.6443692,0.04514506,0.0313816,0.006037537,0.0003797344],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9230577,0.003059896,0.06271193,0.0002371843,0.0001257142,0.0002516559,0.001829505,0.0007901886,0.007936302],"genre_scores_gemma":[0.9472552,0.0006117703,0.04735991,0.00003593523,0.00008944846,0.0001533069,0.003636525,0.00009555333,0.0007623129],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01680863,"threshold_uncertainty_score":0.04255748,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2119168550","doi":"10.1162/0891201042544884","title":"The Alignment Template Approach to Statistical Machine Translation","year":2004,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":933,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Computer science; NIST; Machine translation; Evaluation of machine translation; Natural language processing; Example-based machine translation; Artificial intelligence; Phrase; Machine translation software usability; Generalization; Transfer-based machine translation; Translation (biology); Task (project management); Context (archaeology); Feature (linguistics); Rule-based machine translation; Word (group theory); Linguistics","authors":[{"name":"Franz Josef Och","is_ca":false},{"name":"Hermann Ney","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02186480999394037,"gpt":0.2953628508639949,"spread":0.2734980408700545,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002889534,0.001535862,0.001887775,0.002517394,0.0008623259,0.002795553,0.002571376,0.001907262,0.007118283],"category_scores_gemma":[0.007980834,0.0009584452,0.001889473,0.004376983,0.001277208,0.003065383,0.002092547,0.002932003,0.00947412],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009297866,"about_ca_system_score_gemma":0.002074727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00184641,"about_ca_topic_score_gemma":0.001676509,"domain_scores_codex":[0.9961126,0.001807908,0.0002858526,0.0006366392,0.001036355,0.0001206573],"domain_scores_gemma":[0.996811,0.001354415,0.0002526947,0.0009296159,0.0005798683,0.00007237609],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001032152,0.0001195738,0.0004844826,0.0006870677,0.0003779974,0.0002821159,0.0002048647,0.06246563,0.01354581,0.316671,0.02061995,0.5844383],"study_design_scores_gemma":[0.00004647604,0.0001810668,0.0004733773,0.00008944585,0.0001115153,0.0006899307,0.00005223907,0.4675969,0.01183612,0.4172697,0.1015195,0.000133643],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0003214897,0.0006344055,0.9952965,0.0001381774,0.0001143384,0.00006476677,0.0001183022,0.001227487,0.002084528],"genre_scores_gemma":[0.03202864,0.002575733,0.9558275,0.0003131784,0.0005394523,0.0006263044,0.001209073,0.0008778851,0.006002198],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007118283,"threshold_uncertainty_score":0.02381301,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2051593977","doi":"10.1162/coli_a_00002","title":"Generating Phrasal and Sentential Paraphrases: A Survey of Data-Driven Methods","year":2010,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":302,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Atomic Energy of Canada Limited; National Science Foundation","keywords":"Paraphrase; Computer science; Natural language processing; Task (project management); Artificial intelligence; Parallel corpora; Field (mathematics); Linguistics; Machine translation","authors":[{"name":"Nitin Madnani","is_ca":false},{"name":"Bonnie J. Dorr","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06410955906269462,"gpt":0.401568733933962,"spread":0.3374591748712674,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01047882,0.001759898,0.002634373,0.007297795,0.0009583142,0.0034426,0.005062104,0.001920267,0.003921885],"category_scores_gemma":[0.03765097,0.001163737,0.00169058,0.009864944,0.001401756,0.005885391,0.002230366,0.002093506,0.003457798],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009461999,"about_ca_system_score_gemma":0.001856804,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001743503,"about_ca_topic_score_gemma":0.002342924,"domain_scores_codex":[0.9888002,0.005245849,0.001435808,0.001457733,0.002913964,0.0001464791],"domain_scores_gemma":[0.9483508,0.03835462,0.001861513,0.004715962,0.006364801,0.0003522425],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001780643,0.0003150378,0.001234041,0.00376138,0.000109724,0.0001086523,0.0006247784,0.00402025,0.007104069,0.006241766,0.005496593,0.9708057],"study_design_scores_gemma":[0.0006051477,0.001691794,0.01181951,0.00417152,0.00088692,0.006003825,0.004444236,0.4604189,0.1392234,0.1208381,0.2492936,0.0006031058],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.01209397,0.0309357,0.943764,0.001370584,0.0001220495,0.0009578547,0.002346252,0.004271593,0.004137907],"genre_scores_gemma":[0.05460917,0.01910826,0.9149152,0.0004581482,0.000162799,0.0008902143,0.007204855,0.0008128833,0.001838503],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.01047882,"threshold_uncertainty_score":0.05541795,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1978620866","doi":"10.1162/coli.2008.34.2.145","title":"Semantic Role Labeling: An Introduction to the Special Issue","year":2008,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":278,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; Artificial Intelligence in Medicine (Canada); University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Task (project management); Field (mathematics); Computational linguistics; Identification (biology); Computational model; Semantic role labeling; Artificial intelligence; Natural language processing; Data science; Key (lock); Semantics (computer science)","authors":[{"name":"Lluı́s Màrquez","is_ca":true},{"name":"Xavier Carreras","is_ca":true},{"name":"Kenneth C. Litkowski","is_ca":true},{"name":"Suzanne Stevenson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01392567728635698,"gpt":0.2788210233112759,"spread":0.2648953460249189,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003817555,0.001616728,0.00183728,0.004672071,0.00277345,0.007790515,0.001989048,0.004152158,0.03282798],"category_scores_gemma":[0.01049643,0.001086494,0.001887413,0.003595748,0.00312039,0.01329346,0.00341756,0.01004796,0.02249422],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002023924,"about_ca_system_score_gemma":0.002021655,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00105027,"about_ca_topic_score_gemma":0.001758201,"domain_scores_codex":[0.9981395,0.0005108117,0.0002764369,0.0004275127,0.0005210843,0.0001247214],"domain_scores_gemma":[0.9905044,0.005801042,0.0004638284,0.0006946858,0.001738897,0.0007971383],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002713899,0.00006760091,0.0001974107,0.0003108014,0.00001582576,0.00009813481,0.0001373242,0.0002542073,0.0004874906,0.02225694,0.8864222,0.08972489],"study_design_scores_gemma":[0.000003936943,0.00002468284,0.0002786157,0.0002618037,0.000009207161,0.0003532872,0.0000995176,0.0003942882,0.00012994,0.0148436,0.983583,0.00001809606],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.001513104,0.1114024,0.1207027,0.1087616,0.5509495,0.0004921966,0.001412835,0.001771003,0.1029947],"genre_scores_gemma":[0.01003301,0.120644,0.05775074,0.03364803,0.6326497,0.0007762572,0.003678014,0.002526621,0.1382937],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.03282798,"threshold_uncertainty_score":0.1098205,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2092527610","doi":"10.1162/089120102760173625","title":"Near-Synonymy and Lexical Choice","year":2002,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":258,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Natural language processing; Denotation (semiotics); Artificial intelligence; Context (archaeology); Lexicon; Lexical choice; Set (abstract data type); Selection (genetic algorithm); Lexical semantics; Ontology; Lexical item; Linguistics","authors":[{"name":"Philip Edmonds","is_ca":false},{"name":"Graeme Hirst","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02487434097093336,"gpt":0.2844013043204678,"spread":0.2595269633495345,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002068811,0.00067599,0.001127268,0.002508612,0.001653466,0.004144925,0.003154934,0.001785794,0.007152132],"category_scores_gemma":[0.007947455,0.0007584203,0.002976573,0.00280112,0.004996182,0.01306055,0.00381253,0.002213376,0.001530941],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002397544,"about_ca_system_score_gemma":0.001668088,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003954945,"about_ca_topic_score_gemma":0.003367992,"domain_scores_codex":[0.9967014,0.001005468,0.0003051002,0.001024087,0.0007608715,0.0002030073],"domain_scores_gemma":[0.9966118,0.001731492,0.0003599847,0.0007326696,0.0004006295,0.0001634018],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003662437,0.00002398611,0.000564302,0.00007986676,0.00003870005,0.000151933,0.0005940836,0.02199537,0.001162462,0.9567356,0.000641158,0.01797595],"study_design_scores_gemma":[0.00001469278,0.00001766167,0.0001630168,0.00002277126,0.00001717957,0.00009750773,0.0001279735,0.09206878,0.0006908448,0.9015422,0.005215084,0.00002235286],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0215734,0.0002085187,0.9677023,0.0007176159,0.0000503013,0.0001036421,0.0002081746,0.000194744,0.009241168],"genre_scores_gemma":[0.4640301,0.0002978935,0.5263891,0.0002972171,0.00007738081,0.0004240198,0.0006255143,0.0001373309,0.007721356],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007152132,"threshold_uncertainty_score":0.02392632,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2164948578","doi":"10.1162/089120103321337458","title":"Word Reordering and a Dynamic Programming Beam Search Algorithm for Statistical Machine Translation","year":2003,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":257,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Computer science; Word (group theory); Task (project management); Machine translation; Vocabulary; Beam search; Word error rate; Sentence; Translation (biology); Artificial intelligence; Natural language processing; Dynamic programming; Search algorithm; Speech recognition; Set (abstract data type); Algorithm; Linguistics; Programming language","authors":[{"name":"Christoph Tillmann","is_ca":false},{"name":"Hermann Ney","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01830414375166094,"gpt":0.319656416215049,"spread":0.3013522724633881,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002154728,0.001103758,0.001403458,0.001525322,0.0007651085,0.001068281,0.001575622,0.001337416,0.00535028],"category_scores_gemma":[0.006271593,0.0009679737,0.001090812,0.003113655,0.001133488,0.002148295,0.001249659,0.001946921,0.002357747],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007492857,"about_ca_system_score_gemma":0.001717256,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003122549,"about_ca_topic_score_gemma":0.004319443,"domain_scores_codex":[0.9980897,0.0009716555,0.0001142469,0.000310175,0.0004245213,0.0000896112],"domain_scores_gemma":[0.9977291,0.001539801,0.0001121295,0.0002237218,0.0003580615,0.00003724744],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002129682,0.0001214303,0.0004349302,0.0002825667,0.000163471,0.0001514684,0.0002946887,0.2715408,0.01063096,0.1135885,0.008608317,0.5939699],"study_design_scores_gemma":[0.0000636851,0.00006043009,0.0001053877,0.00001891571,0.00002505707,0.00007336041,0.00002686806,0.9234301,0.002892099,0.06645893,0.006819246,0.00002590173],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001408353,0.0001378291,0.9973579,0.00006170898,0.00002311057,0.0000342612,0.0000313814,0.0003778893,0.0005675186],"genre_scores_gemma":[0.02167759,0.0001711928,0.9760864,0.0001059655,0.00003259958,0.0002771794,0.0002047074,0.0002275124,0.001216878],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00535028,"threshold_uncertainty_score":0.0178985,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2046436395","doi":"10.1162/089120100561755","title":"The Rhetorical Parsing of Unrestricted Texts: A Surface-based Approach","year":2000,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":221,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; University of Toronto","keywords":"Rhetorical question; Computer science; Automatic summarization; Parsing; Natural language processing; Artificial intelligence; Linguistics; Context (archaeology); Natural language understanding; Natural language; Philosophy; History","authors":[{"name":"Daniel Marcu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02046125002567074,"gpt":0.2820272355339836,"spread":0.2615659855083128,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00255761,0.0008841841,0.00107342,0.004258771,0.001415092,0.005042968,0.001637241,0.001328135,0.003956576],"category_scores_gemma":[0.01841367,0.0008874147,0.001124462,0.002547296,0.003182783,0.006781413,0.001728219,0.002274546,0.001929533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008654835,"about_ca_system_score_gemma":0.001369426,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006661028,"about_ca_topic_score_gemma":0.0009330409,"domain_scores_codex":[0.9977932,0.001156896,0.0001218225,0.0003718965,0.0004825299,0.00007371001],"domain_scores_gemma":[0.9881666,0.008296403,0.0006829425,0.001501563,0.001227065,0.0001252799],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001796173,0.000179363,0.002702091,0.001293525,0.0001168052,0.0004081225,0.00441657,0.04570746,0.03728511,0.2927009,0.005571573,0.6094388],"study_design_scores_gemma":[0.00004978351,0.0001900954,0.002142844,0.0002429165,0.0001090402,0.0004639716,0.001334527,0.5034836,0.03129079,0.4276727,0.03292878,0.00009093446],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01376512,0.00039679,0.9803957,0.000404182,0.00003417741,0.0001192183,0.0002051501,0.001745015,0.002934744],"genre_scores_gemma":[0.1633496,0.0005380128,0.8322905,0.0001072196,0.00009788948,0.0001596084,0.001150515,0.0008248244,0.001481868],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005042968,"threshold_uncertainty_score":0.01352608,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2123215530","doi":"10.1162/coli.08-010-r1-07-048","title":"Unsupervised Type and Token Identification of Idiomatic Expressions","year":2009,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":201,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Atomic Energy of Canada Limited; University of Toronto","keywords":"Linguistics; Identification (biology); Interpretation (philosophy); Computer science; Natural language processing; Literal (mathematical logic); Task (project management); Context (archaeology); Security token; Artificial intelligence; Semantic property; Expression (computer science); Psychology; History","authors":[{"name":"Afsaneh Fazly","is_ca":true},{"name":"Paul Cook","is_ca":true},{"name":"Suzanne Stevenson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01564023181424613,"gpt":0.2969218139515016,"spread":0.2812815821372555,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001952591,0.000612154,0.0009856471,0.003605598,0.0006721235,0.00164894,0.001460423,0.0008356387,0.001613579],"category_scores_gemma":[0.01012901,0.0003166714,0.0007107049,0.002530779,0.0008239504,0.002931914,0.001063656,0.001126331,0.001534898],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006121513,"about_ca_system_score_gemma":0.001109624,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001350142,"about_ca_topic_score_gemma":0.00213235,"domain_scores_codex":[0.9975382,0.0006523495,0.0002863831,0.000813653,0.0004878339,0.0002216994],"domain_scores_gemma":[0.9904217,0.004739991,0.001297696,0.001365272,0.001884192,0.0002911735],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001272401,0.0002930535,0.08236528,0.0005782445,0.0001484841,0.0006804428,0.001315691,0.00578367,0.08303012,0.01692225,0.01054521,0.7970653],"study_design_scores_gemma":[0.00008678516,0.0002672633,0.07606187,0.00009126233,0.0002044139,0.002858156,0.001630861,0.7434557,0.112568,0.04964045,0.01293075,0.0002044469],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.402551,0.0005095452,0.5849693,0.000363638,0.0001920137,0.0003078166,0.002072431,0.004041636,0.00499272],"genre_scores_gemma":[0.63266,0.0002006652,0.3592974,0.0001009383,0.00009571162,0.0002496105,0.003979844,0.0004924405,0.002923427],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003605598,"threshold_uncertainty_score":0.01032639,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2166391512","doi":"10.1162/089120101317066122","title":"Automatic Verb Classification Based on Statistical Distributions of Argument Structure","year":2001,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":199,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; University of Toronto; National Science Foundation","keywords":"Computer science; Natural language processing; Artificial intelligence; Verb; Sentence; Argument (complex analysis); Classifier (UML); Predicate (mathematical logic)","authors":[{"name":"Paola Merlo","is_ca":false},{"name":"Suzanne Stevenson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01714931419756963,"gpt":0.3015708389109142,"spread":0.2844215247133445,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003481698,0.0005399815,0.0007507592,0.002959898,0.0005772026,0.001778694,0.001119739,0.0009280024,0.00154683],"category_scores_gemma":[0.0184218,0.0003331444,0.0005488006,0.001384856,0.0008118129,0.002943097,0.0009137644,0.00113779,0.001325548],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007458882,"about_ca_system_score_gemma":0.0008791592,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001230054,"about_ca_topic_score_gemma":0.001947776,"domain_scores_codex":[0.9977788,0.0008937017,0.000163688,0.0005686611,0.0004475829,0.0001475149],"domain_scores_gemma":[0.9846358,0.01104658,0.00132245,0.001235742,0.001591691,0.0001677203],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0005950482,0.0004321586,0.06581441,0.0002572685,0.0001032881,0.0002026356,0.0008371713,0.04034917,0.04970795,0.01676778,0.006742012,0.8181911],"study_design_scores_gemma":[0.00003194559,0.00005498793,0.01083031,0.00002383741,0.00001468569,0.0001469279,0.0001444319,0.9518549,0.01444088,0.02080767,0.001626473,0.00002299279],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2522357,0.0001822824,0.7391645,0.0002967717,0.000038146,0.0001579158,0.0006159189,0.003941578,0.003367217],"genre_scores_gemma":[0.7937898,0.00009030243,0.2021216,0.00006938032,0.00004023989,0.0001969548,0.002203321,0.0002713548,0.001217125],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003481698,"threshold_uncertainty_score":0.01841319,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1894075015","doi":"10.1162/coli_a_00226","title":"CODRA: A Novel Discriminative Framework for Rhetorical Analysis","year":2015,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":186,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Parsing; Computer science; Natural language processing; Rhetorical question; Discriminative model; Artificial intelligence; Coherence (philosophical gambling strategy); Probabilistic logic; Parse tree; Classifier (UML); Linguistics; Margin (machine learning); Machine learning; Mathematics","authors":[{"name":"Shafiq Joty","is_ca":false},{"name":"Giuseppe Carenini","is_ca":true},{"name":"Raymond T. Ng","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07786810095225595,"gpt":0.3751987668656995,"spread":0.2973306659134436,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004381186,0.001761199,0.001558995,0.006914157,0.001737616,0.002841225,0.003873499,0.00212135,0.005917545],"category_scores_gemma":[0.01282732,0.001136831,0.001701105,0.004333582,0.002225323,0.004854862,0.003367592,0.003643182,0.003348278],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001607868,"about_ca_system_score_gemma":0.002893361,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006615682,"about_ca_topic_score_gemma":0.01394827,"domain_scores_codex":[0.9963344,0.001418656,0.0001946803,0.001113606,0.00071882,0.0002197965],"domain_scores_gemma":[0.9922807,0.004390681,0.0007575566,0.001439237,0.0008608648,0.0002709751],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002632594,0.0002956001,0.005236629,0.0007626677,0.0002014443,0.0003513607,0.0009168882,0.04663705,0.01537577,0.1785283,0.02625609,0.7251748],"study_design_scores_gemma":[0.00004108062,0.00006886266,0.00202306,0.00007312797,0.00005991268,0.0004856064,0.0001374074,0.7774611,0.005572089,0.1820549,0.03194838,0.00007450148],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003956604,0.0005497567,0.9885156,0.0002887586,0.00004819192,0.0001111568,0.0009597138,0.003866717,0.00170348],"genre_scores_gemma":[0.1784899,0.0005139387,0.8090255,0.0004036475,0.0002971166,0.0005195447,0.005409167,0.00129451,0.004046726],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006914157,"threshold_uncertainty_score":0.02317017,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2109830295","doi":"10.1162/coli.2006.32.3.379","title":"Similarity of Semantic Relations","year":2006,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":179,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"National Research Council Canada","funders":"University of Waterloo; Johns Hopkins University","keywords":"Analogy; Word (group theory); Similarity (geometry); Vector space model; Semantic similarity; Relation (database); Vector space; Latent semantic analysis","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.01936146345015635,"gpt":0.2584151920057683,"spread":0.239053728555612,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002787363,0.0006694928,0.0007959002,0.01089845,0.001171668,0.004912937,0.001128977,0.001222455,0.01031445],"category_scores_gemma":[0.02974747,0.000333848,0.001402156,0.007791879,0.002504809,0.009063483,0.003523278,0.001188519,0.00214861],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001466735,"about_ca_system_score_gemma":0.0009266726,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001424476,"about_ca_topic_score_gemma":0.001045778,"domain_scores_codex":[0.9922525,0.002154833,0.0009495632,0.001615369,0.002712154,0.000315489],"domain_scores_gemma":[0.9907351,0.004775841,0.0009562263,0.001722048,0.001557844,0.0002530577],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004952037,0.0002559614,0.03424416,0.001463902,0.000790783,0.0005878806,0.007331467,0.006190897,0.01239439,0.5322224,0.01275056,0.3912723],"study_design_scores_gemma":[0.00005771829,0.0002078196,0.0295562,0.000298174,0.0002556208,0.001243078,0.004192665,0.02487789,0.003406513,0.8615936,0.07420375,0.0001069845],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.333966,0.01042654,0.48856,0.002946021,0.0008200454,0.001115949,0.01161031,0.00169628,0.1488588],"genre_scores_gemma":[0.8810795,0.002110034,0.1001165,0.0004212104,0.0003368448,0.0006472223,0.009251874,0.0002061431,0.005830756],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01089845,"threshold_uncertainty_score":0.03450531,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4205635927","doi":"10.1162/coli_a_00426","title":"Deep Learning for Text Style Transfer: A Survey","year":2021,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":162,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Politeness; Style (visual arts); Task (project management); Variety (cybernetics); Natural language processing; Artificial intelligence; Transfer of learning; Field (mathematics); Natural language; Deep learning; Natural (archaeology); Linguistics","authors":[{"name":"Di Jin","is_ca":false},{"name":"Zhijing Jin","is_ca":false},{"name":"Zhiting Hu","is_ca":false},{"name":"Olga Vechtomova","is_ca":true},{"name":"Rada Mihalcea","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04084954962181789,"gpt":0.2874410351691519,"spread":0.246591485547334,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001614483,0.001269903,0.001131779,0.00166255,0.0003281792,0.001311474,0.001571845,0.0009088408,0.004167637],"category_scores_gemma":[0.004238799,0.0004408585,0.0009342756,0.002016612,0.0004133914,0.002264629,0.001287787,0.001681644,0.002217255],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006905592,"about_ca_system_score_gemma":0.00074449,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002790527,"about_ca_topic_score_gemma":0.002560421,"domain_scores_codex":[0.999316,0.0002062379,0.0000646911,0.0002025334,0.0001565204,0.00005407097],"domain_scores_gemma":[0.9985286,0.0008487054,0.0000779972,0.0002325209,0.0002458788,0.00006627622],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00009439478,0.0002133605,0.002023613,0.0009177086,0.0001480327,0.00003928563,0.00009490253,0.0279512,0.002104642,0.007001833,0.01657775,0.9428334],"study_design_scores_gemma":[0.00005766159,0.0002780933,0.004819018,0.0008409182,0.0001848237,0.0001705056,0.0001793865,0.8650008,0.006584863,0.04887984,0.07294323,0.00006068838],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.05410707,0.2896169,0.618791,0.005600206,0.001502034,0.0003020327,0.002325285,0.005242757,0.02251273],"genre_scores_gemma":[0.6021081,0.1507554,0.2076003,0.002213445,0.003003543,0.0004939602,0.008685363,0.0008581356,0.02428178],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.004167637,"threshold_uncertainty_score":0.01394218,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4412158322","doi":"10.1162/coli.a.16","title":"🧜Siren’s Song in the AI Ocean: A Survey on Hallucination in Large Language Models","year":2025,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":142,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Vector Institute; University of Waterloo","funders":"","keywords":"Siren (mythology); Computer science; Natural language processing; Linguistics; Art; Literature; Philosophy","authors":[{"name":"Yue Zhang","is_ca":false},{"name":"Yafu Li","is_ca":false},{"name":"Leyang Cui","is_ca":false},{"name":"Cai Deng","is_ca":false},{"name":"Lemao Liu","is_ca":false},{"name":"Tingchen Fu","is_ca":false},{"name":"Xinting Huang","is_ca":false},{"name":"Enbo Zhao","is_ca":false},{"name":"Yanwen Zhang","is_ca":false},{"name":"Yulong Chen","is_ca":false},{"name":"Longyue Wang","is_ca":false},{"name":"Ahn Tuan Luu","is_ca":false},{"name":"Wei Bi","is_ca":false},{"name":"Freda Shi","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02578220776824297,"gpt":0.3496665190830933,"spread":0.3238843113148503,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005140209,0.0008492813,0.001062743,0.002366817,0.0004934295,0.003175358,0.001597508,0.001148357,0.001586639],"category_scores_gemma":[0.02783582,0.0005104649,0.0007931922,0.002056405,0.001366952,0.00486848,0.002398461,0.002218206,0.0006971562],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007739268,"about_ca_system_score_gemma":0.001334841,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003482481,"about_ca_topic_score_gemma":0.002808019,"domain_scores_codex":[0.9969144,0.001713533,0.0002364757,0.0003217274,0.0007179247,0.00009593038],"domain_scores_gemma":[0.9686005,0.02639811,0.0008496669,0.001659123,0.002112498,0.0003800677],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002802013,0.0002414738,0.01377635,0.003667572,0.0004307674,0.0005135548,0.00196258,0.06499913,0.00274424,0.03821485,0.01815425,0.855015],"study_design_scores_gemma":[0.00006028948,0.0004775016,0.006248761,0.002505156,0.0002659835,0.001388733,0.002371172,0.6917969,0.006658549,0.1558374,0.1321883,0.0002012651],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.06632197,0.2767638,0.6189047,0.0129323,0.0006774334,0.0003226401,0.0007851623,0.003878175,0.01941385],"genre_scores_gemma":[0.6460825,0.1689519,0.1725573,0.00301514,0.002313291,0.0002680934,0.002259748,0.0009694088,0.003582634],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.005140209,"threshold_uncertainty_score":0.02718437,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2170471032","doi":"10.1162/089120103322753365","title":"Disambiguating Nouns, Verbs, and Adjectives Using Automatically Acquired Selectional Preferences","year":2003,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":141,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Engineering and Physical Sciences Research Council; Atomic Energy of Canada Limited","keywords":"Computer science; Adjective; Noun; Natural language processing; Verb; Artificial intelligence; Word (group theory); Heuristic; Argument (complex analysis); Part of speech; Linguistics","authors":[{"name":"Diana McCarthy","is_ca":false},{"name":"John Carroll","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02433043378964063,"gpt":0.3023177727749802,"spread":0.2779873389853396,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003878665,0.000853889,0.0009596024,0.004175608,0.00097741,0.001945875,0.000637814,0.0007250617,0.0009121499],"category_scores_gemma":[0.01751803,0.0004114117,0.0006133033,0.003279594,0.000696438,0.003013552,0.001625372,0.000707844,0.0008941206],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005978033,"about_ca_system_score_gemma":0.001046327,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003658774,"about_ca_topic_score_gemma":0.007457117,"domain_scores_codex":[0.9969323,0.00126989,0.0003725471,0.0006800415,0.0006304519,0.0001147762],"domain_scores_gemma":[0.9826937,0.01171853,0.001120157,0.001783054,0.002432541,0.0002519405],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002014697,0.0006307299,0.09386285,0.0013649,0.0005414154,0.0006306854,0.004761628,0.04399338,0.1163767,0.005410477,0.007701236,0.7227112],"study_design_scores_gemma":[0.0003190016,0.0008111889,0.09614705,0.0001775334,0.0005131208,0.001870587,0.005853472,0.5893942,0.2513878,0.02315582,0.03000402,0.0003662433],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8695164,0.0006923937,0.1176627,0.0002616121,0.00005557213,0.000270693,0.002710147,0.003195768,0.005634833],"genre_scores_gemma":[0.8423027,0.0002715034,0.1492953,0.00009192689,0.00002752887,0.000150839,0.006513522,0.0002892957,0.0010573],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.004175608,"threshold_uncertainty_score":0.02051258,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2151552691","doi":"10.1162/089120102762671963","title":"Generating Indicative-Informative Summaries with SumUM","year":2002,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":140,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Université du Québec à Montréal","funders":"Université de Montréal; McGill University","keywords":"Automatic summarization; Computer science; Identification (biology); Information retrieval; Natural language processing; Process (computing); Artificial intelligence","authors":[{"name":"Horacio Saggion","is_ca":false},{"name":"Guy Lapalme","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02829016747849478,"gpt":0.2487386652571547,"spread":0.2204484977786599,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0032882,0.001026326,0.0008721785,0.001726694,0.0006755776,0.002016862,0.001017503,0.0007514986,0.003921166],"category_scores_gemma":[0.01550213,0.000294746,0.000505013,0.001574542,0.000339388,0.002873464,0.001483337,0.0006341445,0.002071163],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003729771,"about_ca_system_score_gemma":0.0005408053,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005752751,"about_ca_topic_score_gemma":0.0009960994,"domain_scores_codex":[0.9977459,0.001182842,0.0002247231,0.0003146676,0.000468852,0.0000629391],"domain_scores_gemma":[0.9915404,0.005170198,0.0007743528,0.0008941223,0.001432848,0.0001881823],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001014773,0.0002865345,0.002052643,0.00186472,0.0001974389,0.0004094316,0.002920011,0.0216078,0.04772545,0.009200629,0.01726666,0.8954539],"study_design_scores_gemma":[0.000514128,0.001936246,0.005025861,0.0002635838,0.0006082543,0.0008584086,0.00228446,0.6160976,0.2331996,0.02120404,0.1177602,0.0002476877],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09463151,0.00179078,0.8724024,0.0005623887,0.000254802,0.0005361673,0.001425215,0.02375864,0.004638192],"genre_scores_gemma":[0.2527396,0.0006821218,0.735935,0.0001666541,0.0001914647,0.0004197634,0.004523207,0.0009616152,0.004380449],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003921166,"threshold_uncertainty_score":0.01738983,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2111856253","doi":"10.1162/coli.2007.33.1.9","title":"Word-Level Confidence Estimation for Machine Translation","year":2007,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":106,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Research Council Canada; RWTH Aachen University; European Commission","keywords":"Computer science; Word (group theory); Phrase; Machine translation; Translation (biology); Natural language processing; Artificial intelligence; Example-based machine translation; Linguistics","authors":[{"name":"Nicola Ueffing","is_ca":false},{"name":"Hermann Ney","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04620772119004527,"gpt":0.3427001006526349,"spread":0.2964923794625896,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01773942,0.001595785,0.001919982,0.006614158,0.001150668,0.003867452,0.002389209,0.002545687,0.002923929],"category_scores_gemma":[0.1817769,0.0008163168,0.001477259,0.004514826,0.00246889,0.007057548,0.002816643,0.003084298,0.001177663],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001591202,"about_ca_system_score_gemma":0.00120533,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001484701,"about_ca_topic_score_gemma":0.0007071667,"domain_scores_codex":[0.977032,0.01073644,0.001773375,0.002266762,0.007690343,0.0005010507],"domain_scores_gemma":[0.8278335,0.148691,0.007185244,0.006669271,0.00886376,0.0007571947],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0009358517,0.0001472514,0.0153935,0.001005812,0.0007096624,0.000217643,0.0005819143,0.3288257,0.006068152,0.08984554,0.004041324,0.5522277],"study_design_scores_gemma":[0.00006224454,0.0001845964,0.003489918,0.0001407945,0.0001062863,0.0003692496,0.00008343089,0.8641735,0.008650952,0.1194004,0.003195381,0.0001432281],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01313833,0.001886627,0.9822025,0.0001925404,0.00006287755,0.00005381737,0.0001450533,0.0008574763,0.001460856],"genre_scores_gemma":[0.4547023,0.0008907576,0.5411777,0.0001809682,0.0003304452,0.000307122,0.001083403,0.0005740719,0.0007533101],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01773942,"threshold_uncertainty_score":0.0938161,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2168801328","doi":"10.1162/089120103322711587","title":"Embedding Web-Based Statistical Translation Models in Cross-Language Information Retrieval","year":2003,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":106,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Université du Québec à Montréal","funders":"","keywords":"Computer science; Cross-language information retrieval; Machine translation; Natural language processing; Information retrieval; Artificial intelligence; Process (computing); Translation (biology); Machine translation software usability; Language model; Embedding; Example-based machine translation; Programming language","authors":[{"name":"Wessel Kraaij","is_ca":false},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Michel Simard","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01959166820152231,"gpt":0.3257819259314265,"spread":0.3061902577299042,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002948029,0.0006858851,0.0009962313,0.002062425,0.0006153194,0.002010587,0.0008486329,0.001592291,0.002892602],"category_scores_gemma":[0.01255923,0.0008248767,0.00107615,0.002539771,0.001093389,0.004334592,0.001423631,0.001257772,0.002479332],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001088434,"about_ca_system_score_gemma":0.001222439,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007951599,"about_ca_topic_score_gemma":0.007478223,"domain_scores_codex":[0.9977093,0.001545737,0.0001244642,0.0002656126,0.0002706117,0.00008442893],"domain_scores_gemma":[0.9948868,0.003550359,0.0003219531,0.0006213525,0.0005588894,0.0000606088],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002724509,0.0002492967,0.001439004,0.0001870036,0.0001525784,0.0002326494,0.0003334114,0.6794626,0.003470942,0.06309647,0.00294401,0.2481596],"study_design_scores_gemma":[0.00001376532,0.00002664273,0.0001219701,0.000008505562,0.0000121795,0.00002934367,0.00002473704,0.9709681,0.0005532609,0.02736202,0.000867193,0.00001224516],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02968374,0.0006860976,0.9654546,0.0005011595,0.00005434126,0.00007006529,0.0001510215,0.001362248,0.002036771],"genre_scores_gemma":[0.5984632,0.00136372,0.3919006,0.0003444765,0.0001776127,0.000361001,0.00113453,0.0005012973,0.005753377],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007951599,"threshold_uncertainty_score":0.01581061,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2017292107","doi":"10.1162/coli_a_00143","title":"Computing Lexical Contrast","year":2012,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":100,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; National Research Council Canada","funders":"National Research Council Canada; Natural Sciences and Engineering Research Council of Canada; National Science Foundation","keywords":"Contrast (vision); Computer science; Natural language processing; Artificial intelligence; Word (group theory); Meaning (existential); Focus (optics); Linguistics; Information retrieval; Psychology; Philosophy","authors":[{"name":"Saif M. Mohammad","is_ca":true},{"name":"Bonnie J. Dorr","is_ca":false},{"name":"Graeme Hirst","is_ca":true},{"name":"Peter D. Turney","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02047305350721997,"gpt":0.3092489102287959,"spread":0.288775856721576,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002343735,0.0009047863,0.00146481,0.01168312,0.00165069,0.004957482,0.001134953,0.001533484,0.008647076],"category_scores_gemma":[0.02471295,0.000533111,0.001011242,0.006615039,0.001119493,0.008883035,0.003286294,0.001134077,0.002612936],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001213868,"about_ca_system_score_gemma":0.001038708,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00149443,"about_ca_topic_score_gemma":0.001737922,"domain_scores_codex":[0.9952253,0.0007471924,0.0006853524,0.001697574,0.001260474,0.0003840722],"domain_scores_gemma":[0.9888359,0.006743078,0.0007787124,0.001089389,0.002070561,0.0004824192],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00270199,0.0006660275,0.160423,0.001775497,0.0007319567,0.00169896,0.002535648,0.01345038,0.0489909,0.08746205,0.02331877,0.6562448],"study_design_scores_gemma":[0.0005158137,0.001490091,0.1436443,0.0004652763,0.000872473,0.005058128,0.006372587,0.2821586,0.04905521,0.415311,0.09464452,0.0004121668],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6477583,0.002545871,0.2844786,0.000918639,0.000525663,0.0006096585,0.01243943,0.003363407,0.04736044],"genre_scores_gemma":[0.8689468,0.0003024661,0.1166332,0.0001864796,0.0001515356,0.0003600258,0.01123523,0.0002736791,0.001910549],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01168312,"threshold_uncertainty_score":0.02892739,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3201284223","doi":"10.1162/coli_a_00433","title":"Ethics Sheet for Automatic Emotion Recognition and Sentiment Analysis","year":2022,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":93,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"National Research Council Canada","funders":"","keywords":"Harm; Emotion recognition; Sentiment analysis; Key (lock); Computer science; Psychology; Social psychology; Artificial intelligence; Computer security","authors":[{"name":"Saif M. Mohammad","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0701563089762093,"gpt":0.3225153650597269,"spread":0.2523590560835176,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.1057936,0.0007726343,0.0005910545,0.002721129,0.00462578,0.007435169,0.001394963,0.005754403,0.01221155],"category_scores_gemma":[0.1564356,0.0007394137,0.0009160558,0.001331654,0.00449763,0.003522436,0.003248792,0.006277152,0.01226119],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005014472,"about_ca_system_score_gemma":0.01603108,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002178249,"about_ca_topic_score_gemma":0.002201476,"domain_scores_codex":[0.8572909,0.09847997,0.01116565,0.002928891,0.0265287,0.003605955],"domain_scores_gemma":[0.7320235,0.1177302,0.01090653,0.02744966,0.1071079,0.004782199],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0005490775,0.0005491409,0.004616072,0.0006860491,0.0000447975,0.001014575,0.008839663,0.002934768,0.0112947,0.3238254,0.442434,0.2032118],"study_design_scores_gemma":[0.0001154232,0.0003090029,0.003474853,0.00127628,0.00003009884,0.0006698286,0.002635483,0.009259662,0.008775081,0.06632021,0.9069677,0.0001663772],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03408362,0.0009790461,0.5513681,0.1461785,0.007276906,0.02666568,0.003336435,0.003416376,0.2266953],"genre_scores_gemma":[0.1964171,0.001202588,0.5750288,0.05747811,0.003886041,0.04067319,0.002848418,0.001086776,0.121379],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9942456,"threshold_uncertainty_score":0.5594967,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2131526828","doi":"10.1162/coli.2006.32.2.223","title":"Building and Using a Lexical Knowledge Base of Near-Synonym Differences","year":2006,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":82,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada; University of Toronto; University of Ottawa; University of Pennsylvania","keywords":"Computer science; Synonym (taxonomy); Natural language processing; Knowledge base; Artificial intelligence; Mistake; Natural language; Bilingual dictionary; Word (group theory); Linguistics","authors":[{"name":"Diana Inkpen","is_ca":true},{"name":"Graeme Hirst","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02656855028045379,"gpt":0.3117453974176719,"spread":0.2851768471372181,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001640353,0.000929211,0.001079903,0.005890309,0.001438933,0.002186206,0.002578247,0.001308012,0.007434053],"category_scores_gemma":[0.01197636,0.0009000483,0.001103343,0.003757343,0.0009858871,0.007264531,0.002623338,0.002039317,0.003469549],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001021726,"about_ca_system_score_gemma":0.001979793,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003366134,"about_ca_topic_score_gemma":0.006922612,"domain_scores_codex":[0.9978129,0.0005226783,0.0002644403,0.0007227087,0.0005798735,0.00009743702],"domain_scores_gemma":[0.9941181,0.003178216,0.0003196016,0.001210754,0.001044007,0.0001292833],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003297457,0.0006179366,0.007016207,0.0009936413,0.000237107,0.002488454,0.002301073,0.01915643,0.0303245,0.04218738,0.02202768,0.8723198],"study_design_scores_gemma":[0.0003274895,0.000462562,0.01088683,0.001002015,0.0006826536,0.004671974,0.003277099,0.4281553,0.09539627,0.2410135,0.213672,0.0004523266],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04265573,0.0003190085,0.9372371,0.0006542419,0.000124393,0.0005290859,0.004486458,0.005070326,0.008923642],"genre_scores_gemma":[0.1733699,0.0003936114,0.8101042,0.0002799794,0.00005396626,0.0003772211,0.01243622,0.0004521786,0.002532745],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007434053,"threshold_uncertainty_score":0.02486932,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2137006453","doi":"10.1162/coli_a_00206","title":"Towards Topic-to-Question Generation","year":2015,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":82,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Lethbridge","funders":"","keywords":"Computer science; Latent Dirichlet allocation; Predicate (mathematical logic); Natural language processing; Correctness; Artificial intelligence; Topic model; Similarity (geometry); Subsequence; Context (archaeology); String (physics); Information retrieval; Algorithm; Mathematics; Programming language","authors":[{"name":"Yllias Chali","is_ca":true},{"name":"Sadid A. Hasan","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08126377489865644,"gpt":0.3230650423337741,"spread":0.2418012674351177,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005216217,0.00144383,0.001151568,0.002428363,0.0008650417,0.001728073,0.002308184,0.002256634,0.005471166],"category_scores_gemma":[0.01845322,0.0006162217,0.001909163,0.001843386,0.0007617014,0.003313736,0.002826752,0.002180405,0.002808698],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008901734,"about_ca_system_score_gemma":0.001518257,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001405612,"about_ca_topic_score_gemma":0.001634826,"domain_scores_codex":[0.9949032,0.002966461,0.0002391392,0.0008930683,0.0007591859,0.0002389705],"domain_scores_gemma":[0.9875201,0.009356732,0.0003594699,0.001222914,0.001302,0.0002386866],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000824075,0.0006193759,0.004491291,0.001233457,0.0002386915,0.0006740702,0.002780256,0.06062264,0.02743335,0.06751461,0.03785937,0.7957088],"study_design_scores_gemma":[0.0001622592,0.0001116272,0.0006799815,0.00005745212,0.0000823119,0.0003995697,0.0003570593,0.8532419,0.01423953,0.1109343,0.01968923,0.000044702],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01372685,0.0006225101,0.9782012,0.0005552895,0.0001540051,0.0003653546,0.0006025806,0.003904623,0.001867704],"genre_scores_gemma":[0.2073292,0.0005325529,0.7800468,0.0004118953,0.0003195275,0.0008887543,0.005572757,0.0007255433,0.004172885],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005471166,"threshold_uncertainty_score":0.02758628,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2562910494","doi":"10.1162/coli_a_00278","title":"Evaluative Language Beyond Bags of Words: Linguistic Insights and Computational Applications","year":2016,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":80,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computational linguistics; Perspective (graphical); Computer science; Linguistics; Field (mathematics); Sociocultural linguistics; Subjectivity; Multidisciplinary approach; Interpretation (philosophy); Sentiment analysis; Affect (linguistics); Artificial intelligence; Natural language processing; Sociology; Natural language; Epistemology; Social science; Comprehension approach","authors":[{"name":"Farah Benamara","is_ca":false},{"name":"Maite Taboada","is_ca":true},{"name":"Yannick Mathieu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01678670568591051,"gpt":0.3042419730166649,"spread":0.2874552673307544,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004224278,0.0008680773,0.001069141,0.004118989,0.001010878,0.007102072,0.001404593,0.0009520444,0.003769204],"category_scores_gemma":[0.02244049,0.0005902937,0.001176707,0.00466716,0.003926412,0.01080331,0.002093059,0.002412041,0.0007390252],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001705676,"about_ca_system_score_gemma":0.0008935955,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00136911,"about_ca_topic_score_gemma":0.001295575,"domain_scores_codex":[0.9972778,0.001631102,0.0002426047,0.0003261395,0.0004296183,0.00009263118],"domain_scores_gemma":[0.9872922,0.01020006,0.0009754097,0.0005496491,0.0008190267,0.0001636753],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00008324729,0.00004321932,0.002396557,0.001154826,0.0001361245,0.0001829408,0.002189057,0.009194092,0.00122103,0.830022,0.006304382,0.1470725],"study_design_scores_gemma":[0.000008590311,0.00002006726,0.001332337,0.0003129858,0.0000306322,0.000115791,0.000697417,0.04603972,0.000432622,0.9341123,0.01686583,0.00003172772],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04368417,0.03796063,0.8640242,0.01329414,0.0007619414,0.0002352429,0.000699972,0.0005863984,0.03875321],"genre_scores_gemma":[0.6299666,0.02202218,0.338241,0.001899473,0.001459525,0.0004767272,0.000820159,0.0002491832,0.004865159],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007102072,"threshold_uncertainty_score":0.02234042,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2105897558","doi":"10.1162/coli.2009.35.4.35402","title":"From Annotator Agreement to Noise Models","year":2009,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Rough Sets and Fuzzy Logic","field":"Computer Science","cited_by":77,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Kellogg's (Canada)","funders":"","keywords":"Benchmarking; Computer science; Noise (video); Agreement; Data set; Quality (philosophy); Set (abstract data type); Resource (disambiguation); Data quality; Data mining; Data science; Econometrics; Artificial intelligence; Mathematics; Linguistics; Physics","authors":[{"name":"Beata Beigman Klebanov","is_ca":true},{"name":"Eyal Beigman","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02859109390663595,"gpt":0.2738249099219155,"spread":0.2452338160152796,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.2712898,0.002643172,0.004258384,0.008287196,0.00466699,0.01524129,0.006899028,0.006341117,0.003291137],"category_scores_gemma":[0.5429263,0.002324083,0.002472288,0.007787501,0.01856312,0.02144884,0.02013682,0.008231635,0.001719489],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007126923,"about_ca_system_score_gemma":0.004966308,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005219304,"about_ca_topic_score_gemma":0.00432706,"domain_scores_codex":[0.6317576,0.2636082,0.01116808,0.04028907,0.04927701,0.00390013],"domain_scores_gemma":[0.3804966,0.5186472,0.02126362,0.05124028,0.02664683,0.001705453],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0009949803,0.0001383852,0.02893059,0.001629628,0.001818451,0.0005920044,0.01359471,0.08688071,0.002239024,0.7240312,0.01169966,0.1274507],"study_design_scores_gemma":[0.00006509005,0.00009787746,0.0038932,0.0004339998,0.0001880239,0.0002606566,0.001074972,0.1218932,0.002256125,0.8602432,0.009403256,0.0001904436],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0221137,0.001952593,0.9578827,0.005809915,0.0004335112,0.0002223523,0.0005374651,0.0004714281,0.01057642],"genre_scores_gemma":[0.7322374,0.001337781,0.2534929,0.00307462,0.0008621493,0.001374946,0.001553218,0.001210021,0.004857035],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2712898,"threshold_uncertainty_score":0.8986294,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4220732108","doi":"10.1162/coli_a_00434","title":"Domain Adaptation with Pre-trained Transformers for Query-Focused Abstractive Text Summarization","year":2022,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Automatic summarization; Computer science; Transformer; Domain adaptation; Natural language processing; Adaptation (eye); Artificial intelligence; Transfer of learning; Task (project management); Information retrieval","authors":[{"name":"Md Tahmid Rahman Laskar","is_ca":true},{"name":"Enamul Hoque","is_ca":true},{"name":"Jimmy Xiangji Huang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01946222096216806,"gpt":0.2497883595253038,"spread":0.2303261385631357,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001941258,0.0009420176,0.0007263044,0.001215794,0.0003217999,0.0008211022,0.001438743,0.0007898915,0.001754289],"category_scores_gemma":[0.007472779,0.0003071917,0.0007701556,0.00102223,0.0004184331,0.002534417,0.001233129,0.001803752,0.001910271],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006258102,"about_ca_system_score_gemma":0.0009041138,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001771695,"about_ca_topic_score_gemma":0.002873156,"domain_scores_codex":[0.9990835,0.0003756149,0.00007757233,0.0002740491,0.0001356958,0.0000535664],"domain_scores_gemma":[0.9973345,0.00128653,0.0002187659,0.0005303778,0.0005381665,0.00009171294],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005876627,0.0004753089,0.00242156,0.0004932141,0.0001533142,0.0002442715,0.0005649574,0.1802954,0.05423548,0.005698717,0.0187578,0.7360723],"study_design_scores_gemma":[0.00006171205,0.0002327918,0.0006004894,0.00001887043,0.00005544244,0.00008781966,0.000124321,0.96212,0.02438392,0.007388795,0.004902005,0.0000239099],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04161377,0.001043834,0.940353,0.0003195229,0.0001037506,0.0001970527,0.0006749717,0.01413567,0.00155843],"genre_scores_gemma":[0.6028143,0.0007137877,0.3824895,0.0004051349,0.0001898425,0.0003468992,0.008061384,0.0006782035,0.00430091],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001941258,"threshold_uncertainty_score":0.01026648,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2144719390","doi":"10.1162/coli_a_00077","title":"Information Status Distinctions and Referring Expressions: An Empirical Study of References to People in News Summaries","year":2011,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":45,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Atomic Energy of Canada Limited","keywords":"Automatic summarization; Computer science; Salience (neuroscience); Affect (linguistics); Context (archaeology); Information retrieval; Natural language processing; Artificial intelligence; Linguistics; History","authors":[{"name":"Advaith Siddharthan","is_ca":false},{"name":"Ani Nenkova","is_ca":false},{"name":"Kathleen McKeown","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09209636236490953,"gpt":0.3350728771866981,"spread":0.2429765148217886,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009096691,0.0002345626,0.0003684439,0.001619774,0.001191128,0.001577958,0.000752532,0.001220795,0.002076461],"category_scores_gemma":[0.1518565,0.0002896214,0.0002641439,0.002290794,0.001258168,0.00436219,0.001294583,0.001625359,0.000477684],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004787404,"about_ca_system_score_gemma":0.0002707919,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001786819,"about_ca_topic_score_gemma":0.002123353,"domain_scores_codex":[0.9910316,0.006907159,0.0005690258,0.0005279836,0.000839199,0.0001250504],"domain_scores_gemma":[0.7994827,0.1751615,0.01436729,0.004947285,0.005133091,0.0009082447],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001522487,0.001055163,0.543169,0.001358224,0.0002909294,0.00221982,0.2690125,0.003597548,0.009458583,0.01213987,0.006337396,0.1498384],"study_design_scores_gemma":[0.0001426301,0.001670352,0.7745522,0.0005968929,0.0003603824,0.004066068,0.116781,0.04575997,0.007867604,0.01467105,0.03330232,0.0002295362],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9927873,0.0003792575,0.003992212,0.0003475175,0.00001123191,0.00004806582,0.0001619791,0.00004958251,0.002222833],"genre_scores_gemma":[0.9963242,0.0002051149,0.002289032,0.00007654227,0.00002307464,0.00004247927,0.0003893563,0.0000297697,0.0006204222],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009096691,"threshold_uncertainty_score":0.04810852,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2129909699","doi":"10.1162/coli_a_00085","title":"Learning Entailment Relations by Global Graph Structure Optimization","year":2011,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":39,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"University of Haifa; Azrieli Foundation; Israel Science Foundation","keywords":"Logical consequence; Computer science; Textual entailment; Inference; Graph; Transitive relation; Artificial intelligence; Constraint (computer-aided design); Theoretical computer science; Natural language processing; Mathematics; Combinatorics","authors":[{"name":"Jonathan Berant","is_ca":false},{"name":"Ido Dagan","is_ca":false},{"name":"Jacob Goldberger","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01091614832774298,"gpt":0.2556109568737796,"spread":0.2446948085460366,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001717985,0.001596563,0.001590756,0.002204729,0.000857718,0.001560144,0.001988153,0.001555094,0.004457093],"category_scores_gemma":[0.006474453,0.0009055479,0.002096435,0.001814361,0.00148291,0.006042568,0.001984664,0.002907624,0.000997568],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001692576,"about_ca_system_score_gemma":0.001842785,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00375409,"about_ca_topic_score_gemma":0.01060202,"domain_scores_codex":[0.9986562,0.0003864969,0.00007623815,0.0005651977,0.0002367716,0.00007920425],"domain_scores_gemma":[0.996351,0.00262397,0.0002312548,0.0004877726,0.000250383,0.00005555394],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002186846,0.0003363875,0.002827089,0.000493319,0.0002410617,0.0002206814,0.0002685291,0.4430498,0.009199821,0.06211682,0.009620806,0.4714069],"study_design_scores_gemma":[0.00003200519,0.00004147325,0.0002516431,0.0000166196,0.00005176607,0.00004545004,0.00004944417,0.8845704,0.003293291,0.1103105,0.001327733,0.000009776571],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01760551,0.0002164884,0.9782619,0.0003144516,0.00001569492,0.00009843785,0.0002908559,0.002016207,0.001180451],"genre_scores_gemma":[0.2442569,0.0003384626,0.7493918,0.0002378316,0.00005225518,0.0001861436,0.003054315,0.0005597311,0.001922573],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004457093,"threshold_uncertainty_score":0.01491046,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2891212627","doi":"10.1162/coli_a_00333","title":"Introduction to the Special Issue on Language in Social Media: Exploiting Discourse and Other Contextual Information","year":2018,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University; University of Ottawa","funders":"","keywords":"Computer science; Social media; Context (archaeology); Perspective (graphical); Process (computing); Interpretation (philosophy); Task (project management); Linguistics; Artificial intelligence; Natural language processing; World Wide Web","authors":[{"name":"Farah Benamara","is_ca":false},{"name":"Diana Inkpen","is_ca":true},{"name":"Maite Taboada","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02197099707484514,"gpt":0.2991216231871204,"spread":0.2771506261122753,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024758,0.002102314,0.002479559,0.006466801,0.00244117,0.008639533,0.002198193,0.004735237,0.04339493],"category_scores_gemma":[0.009385281,0.0008029722,0.002009787,0.004791105,0.002303127,0.00780822,0.004597565,0.007658649,0.02022468],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001756341,"about_ca_system_score_gemma":0.002156202,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009946963,"about_ca_topic_score_gemma":0.001558742,"domain_scores_codex":[0.9978353,0.0004554582,0.000253652,0.0006019369,0.0006813648,0.0001723608],"domain_scores_gemma":[0.9881014,0.006624631,0.0006490845,0.0008626998,0.002160587,0.001601613],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003180056,0.00006577958,0.0003406664,0.0007645957,0.00003756266,0.0001482057,0.0002521125,0.0001865742,0.0005620068,0.007325102,0.922327,0.06795875],"study_design_scores_gemma":[0.000005871868,0.00003377226,0.0006152632,0.0005293423,0.00002113571,0.0003022273,0.000131619,0.0003168013,0.0001293658,0.006135643,0.9917572,0.00002172149],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.001212596,0.09839412,0.02111729,0.07413128,0.773286,0.0002012588,0.001198216,0.0007549201,0.02970428],"genre_scores_gemma":[0.00386405,0.06141495,0.004767849,0.01216064,0.8773366,0.0002200871,0.001598169,0.0007488618,0.03788888],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.04339493,"threshold_uncertainty_score":0.1451705,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2077818552","doi":"10.1162/coli.2010.36.1.36104","title":"Automatically Identifying the Source Words of Lexical Blends in English","year":2010,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; University of Toronto","funders":"University of Toronto","keywords":"Computer science; Natural language processing; Lexicon; Artificial intelligence; Task (project management); Set (abstract data type); Identification (biology); Word (group theory); Source text; Linguistics; Programming language","authors":[{"name":"Paul Cook","is_ca":true},{"name":"Suzanne Stevenson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01327449925434027,"gpt":0.2962019672436924,"spread":0.2829274679893521,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001054521,0.000705841,0.0005604802,0.0021539,0.0008432464,0.001372754,0.0005703391,0.0006234822,0.004629331],"category_scores_gemma":[0.004873028,0.0005177396,0.0004955903,0.001638864,0.0007904175,0.004932688,0.001313445,0.0006782518,0.001779746],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005172645,"about_ca_system_score_gemma":0.0004126126,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003005474,"about_ca_topic_score_gemma":0.004736926,"domain_scores_codex":[0.9990075,0.0002442419,0.0001882162,0.000358569,0.0001485434,0.00005304737],"domain_scores_gemma":[0.9957615,0.002716227,0.0005959516,0.0003164616,0.0005270848,0.00008294561],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001730852,0.0003958741,0.1301558,0.003000718,0.0002253904,0.004543259,0.01285499,0.006506252,0.1806267,0.01509992,0.01548809,0.6293721],"study_design_scores_gemma":[0.0002342961,0.0006191266,0.3282594,0.0006218338,0.0005810839,0.01104208,0.01911733,0.3299426,0.179192,0.03495542,0.095079,0.0003557636],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.9248342,0.001066625,0.06256165,0.0004195592,0.00008583207,0.00011024,0.003035278,0.001577151,0.00630935],"genre_scores_gemma":[0.9448154,0.0004835913,0.0476118,0.00007942822,0.00004621991,0.00006215484,0.004107028,0.0003303954,0.002463989],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.004629331,"threshold_uncertainty_score":0.01548672,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2064296938","doi":"10.1162/coli_a_00064","title":"A Strategy for Information Presentation in Spoken Dialog Systems","year":2011,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Speech and dialogue systems","field":"Computer Science","cited_by":35,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Ames Research Center; Atomic Energy of Canada Limited; National Aeronautics and Space Administration","keywords":"Dialog box; Computer science; Structuring; Dialog system; Process (computing); Presentation (obstetrics); User satisfaction; Human–computer interaction; Information retrieval; World Wide Web","authors":[{"name":"Vera Demberg","is_ca":false},{"name":"Andi Winterboer","is_ca":false},{"name":"Johanna D. Moore","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06398151558250546,"gpt":0.2799794875189091,"spread":0.2159979719364036,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004152914,0.002362533,0.001108317,0.001725439,0.001687028,0.004986439,0.002877066,0.003443064,0.006273338],"category_scores_gemma":[0.01336023,0.0009415048,0.001157013,0.001012913,0.002904812,0.00517786,0.003332205,0.002055932,0.003810667],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001171994,"about_ca_system_score_gemma":0.001153396,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00115227,"about_ca_topic_score_gemma":0.0008716763,"domain_scores_codex":[0.9938861,0.003331413,0.0004494616,0.000913428,0.001253006,0.0001665992],"domain_scores_gemma":[0.9948985,0.002863002,0.0003455353,0.0009449609,0.0007223396,0.0002257847],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008624864,0.0004604195,0.001485658,0.001301192,0.0002405092,0.001207192,0.009580569,0.02599184,0.130729,0.2933275,0.01811546,0.5166982],"study_design_scores_gemma":[0.0004728776,0.001691512,0.001146676,0.0004682843,0.0002755365,0.002326997,0.002502964,0.4424582,0.08469159,0.2773451,0.1862178,0.0004023235],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007388277,0.0006360658,0.979122,0.0009274661,0.0001627967,0.0004782302,0.00008491366,0.003731086,0.007469171],"genre_scores_gemma":[0.1702448,0.0005164403,0.8162823,0.0008474101,0.0001734515,0.001271576,0.0002477801,0.0005221314,0.009894004],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006273338,"threshold_uncertainty_score":0.02196294,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2884082915","doi":"10.1162/coli_a_00327","title":"Anaphora With Non-nominal Antecedents in Computational Linguistics: a Survey","year":2018,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Ruhr-Universität Bochum; Universität Hamburg","keywords":"Anaphora (linguistics); Antecedent (behavioral psychology); Computer science; Automatic summarization; Natural language processing; Linguistics; Sentence; Artificial intelligence; Machine translation; Computational linguistics; Field (mathematics); Resolution (logic); Psychology; Philosophy","authors":[{"name":"Varada Kolhatkar","is_ca":true},{"name":"Adam Roussel","is_ca":false},{"name":"Stefanie Dipper","is_ca":false},{"name":"Heike Zinsmeister","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01827161686791995,"gpt":0.3053307266107789,"spread":0.2870591097428589,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005660324,0.001001442,0.002127039,0.01762878,0.002328815,0.006641808,0.002359903,0.00332195,0.005592528],"category_scores_gemma":[0.01635506,0.001467719,0.001615489,0.02583436,0.003978933,0.01889019,0.004417152,0.003491941,0.002361002],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00214878,"about_ca_system_score_gemma":0.003422385,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002814775,"about_ca_topic_score_gemma":0.002565559,"domain_scores_codex":[0.9955891,0.001525695,0.0006506155,0.0008553069,0.001182605,0.0001966324],"domain_scores_gemma":[0.9741697,0.02257136,0.0006855502,0.001018261,0.001285871,0.0002693151],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005989927,0.0001636045,0.003417302,0.01171264,0.0001779407,0.0005077212,0.005240136,0.001206061,0.0006736683,0.1637609,0.02174421,0.7913359],"study_design_scores_gemma":[0.0000162344,0.00005213575,0.003461252,0.006864672,0.0001102577,0.001798333,0.003143752,0.003433366,0.0006677992,0.1486544,0.831718,0.00007979961],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.003967312,0.9221705,0.04271307,0.004644895,0.0005447061,0.00008641436,0.0001790894,0.0002720597,0.02542192],"genre_scores_gemma":[0.03356872,0.9287457,0.02929895,0.001884281,0.002291133,0.0001638685,0.0005530498,0.0001870044,0.003307239],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.01762878,"threshold_uncertainty_score":0.029935,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2602799059","doi":"10.1162/coli_a_00290","title":"Identifying and Avoiding Confusion in Dialogue with People with Alzheimer's Disease","year":2017,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Speech and dialogue systems","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Toronto Rehabilitation Institute; University of Toronto","funders":"National Institute on Deafness and Other Communication Disorders; National Institutes of Health; Alzheimer Society; AGE-WELL; Carnegie Mellon University; Natural Sciences and Engineering Research Council of Canada; University of Pittsburgh","keywords":"Computer science; Confusion; Dementia; Cognitive psychology; Vocabulary; Function (biology); Cognition; Process (computing); Parsing; Spoken language; Decision tree; Artificial intelligence; Natural language processing; Psychology; Disease; Linguistics; Medicine","authors":[{"name":"Hamidreza Chinaei","is_ca":true},{"name":"Leila Chan Currie","is_ca":true},{"name":"Andrew Danks","is_ca":true},{"name":"Hubert Lin","is_ca":true},{"name":"Tejas Mehta","is_ca":true},{"name":"Frank Rudzicz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03038340881292735,"gpt":0.2762257908780106,"spread":0.2458423820650832,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004651327,0.001097584,0.0008444529,0.001444012,0.001162401,0.002510287,0.0007909318,0.001458607,0.0009651415],"category_scores_gemma":[0.02274182,0.000391551,0.0005500496,0.0003589019,0.0008199176,0.002613697,0.002357197,0.001013349,0.0008960385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005984734,"about_ca_system_score_gemma":0.001175967,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003275659,"about_ca_topic_score_gemma":0.003939286,"domain_scores_codex":[0.9939054,0.004453421,0.0002513342,0.0007625032,0.000365353,0.0002619748],"domain_scores_gemma":[0.9884255,0.00895701,0.0008494447,0.000287081,0.001102721,0.0003782755],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.002418248,0.001041012,0.1104458,0.001089597,0.0002325927,0.0009254878,0.03799742,0.03216984,0.04281686,0.002950502,0.006262101,0.7616506],"study_design_scores_gemma":[0.0001094004,0.001349374,0.06467492,0.0003926085,0.0002816675,0.001506606,0.01951103,0.850176,0.03619826,0.01563695,0.009867811,0.0002953853],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8380473,0.001028155,0.1530204,0.001030482,0.0001106158,0.0002441776,0.0002675577,0.001992485,0.004258815],"genre_scores_gemma":[0.9259674,0.0002596741,0.07192369,0.0001825362,0.00003552452,0.00008616488,0.0004289525,0.0000689463,0.001047052],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.004651327,"threshold_uncertainty_score":0.02459884,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3193068792","doi":"10.1162/coli_a_00422","title":"Probing Classifiers: Promises, Shortcomings, and Advances","year":2021,"lang":"en","type":"preprint","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Azrieli Foundation; Israel Science Foundation","keywords":"Computer science; Artificial intelligence; Variety (cybernetics); Classifier (UML); Machine learning; Property (philosophy); Artificial neural network; Deep neural networks; Natural language processing; Epistemology","authors":[{"name":"Yonatan Belinkov","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03569677429684543,"gpt":0.2887908698049079,"spread":0.2530940955080625,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03587117,0.001675193,0.002665594,0.003246215,0.001899018,0.007862362,0.004325789,0.005503535,0.003792456],"category_scores_gemma":[0.09453927,0.001011254,0.001144988,0.003755189,0.004837242,0.02449579,0.003842504,0.01181683,0.003270743],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003302902,"about_ca_system_score_gemma":0.00318854,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004203292,"about_ca_topic_score_gemma":0.001786568,"domain_scores_codex":[0.98483,0.007672116,0.0004602659,0.002329351,0.004201488,0.000506747],"domain_scores_gemma":[0.9011201,0.0717757,0.001939408,0.009236936,0.01419413,0.00173382],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002787897,0.0001345977,0.003670814,0.0005024605,0.0001354047,0.00008146407,0.0006282135,0.0166449,0.0009037099,0.4699956,0.04591766,0.4611063],"study_design_scores_gemma":[0.00002460462,0.00007951927,0.0006542942,0.0003222412,0.00005599769,0.0001294786,0.0002350286,0.1799554,0.001206681,0.7693937,0.04788057,0.00006249973],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0174049,0.09857094,0.7594551,0.09758005,0.00273385,0.0001559836,0.0005860307,0.00252346,0.02098971],"genre_scores_gemma":[0.5606536,0.06184852,0.3392277,0.01191531,0.01477395,0.0005718343,0.001094398,0.00100626,0.008908446],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.03587117,"threshold_uncertainty_score":0.1897071,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2154889602","doi":"10.1162/coli.2010.36.1.36101","title":"A Graph-Theoretic Framework for Semantic Distance","year":2010,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":20,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; University of Toronto","funders":"","keywords":"Computer science; Semantic similarity; Natural language processing; Artificial intelligence; Coherence (philosophical gambling strategy); Information retrieval; Set (abstract data type); Semantic computing; Semantic Web; Mathematics","authors":[{"name":"Vivian Tsang","is_ca":true},{"name":"Suzanne Stevenson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01206336355466009,"gpt":0.3012501952456227,"spread":0.2891868316909626,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005251364,0.001531167,0.001360385,0.00934088,0.002518809,0.004853337,0.003445572,0.002823514,0.004236909],"category_scores_gemma":[0.0175606,0.0006947228,0.002114119,0.01020504,0.004660551,0.01310754,0.003722858,0.0035528,0.001407531],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003720278,"about_ca_system_score_gemma":0.001690487,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004344617,"about_ca_topic_score_gemma":0.002701092,"domain_scores_codex":[0.9945321,0.002396101,0.0004554129,0.001178558,0.001269531,0.0001682966],"domain_scores_gemma":[0.9925092,0.004974782,0.0005084528,0.0008555551,0.000952975,0.0001991096],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0000121557,0.00002098151,0.0002277768,0.0001250976,0.00004025683,0.00005421604,0.0002349698,0.01366429,0.0003201131,0.952262,0.001748468,0.0312895],"study_design_scores_gemma":[0.000005881496,0.00001266623,0.0001168972,0.00002575084,0.00001110348,0.00006387986,0.00005642557,0.04207969,0.0001778888,0.9482375,0.009197041,0.00001531683],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001957671,0.0007313719,0.9910617,0.0006621294,0.00007597204,0.00007563854,0.0002697293,0.0001211728,0.005044563],"genre_scores_gemma":[0.1409475,0.002608794,0.8482317,0.0005111282,0.000594462,0.0008032148,0.001358081,0.0002096842,0.00473541],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00934088,"threshold_uncertainty_score":0.02777219,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3008170323","doi":"10.1162/coli_a_00380","title":"The Limitations of Stylometry for Detecting Machine-Generated Fake News","year":2020,"lang":"en","type":"preprint","venue":"Computational Linguistics","topic":"Misinformation and Its Impacts","field":"Social Sciences","cited_by":15,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Stylometry; Misinformation; Computer science; Artificial intelligence; Authorship attribution; Natural language processing; Machine learning; Computer security","authors":[{"name":"Tal Schuster","is_ca":true},{"name":"Roei Schuster","is_ca":false},{"name":"Darsh Shah","is_ca":true},{"name":"Regina Barzilay","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1728446409505819,"gpt":0.3757561247219955,"spread":0.2029114837714136,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01009577,0.001156817,0.001047256,0.004547549,0.001701289,0.005096241,0.001337985,0.001859277,0.001908962],"category_scores_gemma":[0.1089529,0.000600989,0.000523128,0.002643849,0.00304778,0.006087,0.00265026,0.001660897,0.001890434],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001154153,"about_ca_system_score_gemma":0.001222887,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001800201,"about_ca_topic_score_gemma":0.001997179,"domain_scores_codex":[0.9783839,0.01158027,0.001359219,0.002312078,0.005887893,0.0004767102],"domain_scores_gemma":[0.8483234,0.0929692,0.0152003,0.03249928,0.009995834,0.001012003],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001449094,0.0003219229,0.1271355,0.002436969,0.0003415423,0.0007151462,0.005668254,0.04297574,0.02925049,0.05894589,0.01530848,0.7154509],"study_design_scores_gemma":[0.00008383459,0.0005733504,0.07861069,0.001105495,0.0001862561,0.00359343,0.00348781,0.5919183,0.09039558,0.1771207,0.0525193,0.0004052907],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4338149,0.00645186,0.5081814,0.005847703,0.0005134669,0.0005825383,0.002959309,0.005309538,0.03633934],"genre_scores_gemma":[0.9215909,0.0008141394,0.07467264,0.0003276078,0.0001795324,0.0001525918,0.0006009091,0.0002047085,0.001456951],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01009577,"threshold_uncertainty_score":0.05339217,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2048760284","doi":"10.1162/089120104323093267","title":"Inferable Centers, Centering Transitions, and the Notion of Coherence","year":2004,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Cohesion (chemistry); Coherence (philosophical gambling strategy); Computer science; Indeterminacy (philosophy); Transition (genetics); Linguistics; Identity (music); Natural language processing; Epistemology; Mathematics; Philosophy","authors":[{"name":"Laurel Fais","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0108289784077711,"gpt":0.2574355551626614,"spread":0.2466065767548903,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003285441,0.0004245016,0.0005436722,0.004552905,0.002915886,0.00317598,0.001165586,0.001006961,0.002152728],"category_scores_gemma":[0.01347225,0.0005575725,0.0006674048,0.004312677,0.006104614,0.0104867,0.00324822,0.001722391,0.0002134518],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001524592,"about_ca_system_score_gemma":0.001445094,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002940483,"about_ca_topic_score_gemma":0.003577682,"domain_scores_codex":[0.9964207,0.001502152,0.0003399249,0.00100609,0.0004702565,0.0002607597],"domain_scores_gemma":[0.9881436,0.007367213,0.001439336,0.001711052,0.00101622,0.0003224969],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002330735,0.00004758703,0.005720591,0.0002152302,0.00002705667,0.0003328026,0.01102716,0.005095951,0.005330502,0.9142862,0.001007782,0.05667603],"study_design_scores_gemma":[0.00003942158,0.0001022626,0.01043744,0.0001000227,0.0001266795,0.0003509023,0.004512637,0.06351352,0.01484917,0.8836205,0.02223746,0.0001099358],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2987657,0.001327004,0.6879303,0.0009388783,0.00004734003,0.0001339851,0.0008495854,0.0006238596,0.009383372],"genre_scores_gemma":[0.8705785,0.0002949682,0.1270922,0.00008487521,0.00005693977,0.0001481005,0.00076864,0.0001382573,0.0008375297],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004552905,"threshold_uncertainty_score":0.01737529,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1502796908","doi":"10.1162/coli.2002.28.4.560","title":"<i>Patterns of Text: In Honour of Michael Hoey</i> Mike Scott and Geoff Thompson (editors) (University of Liverpool) Amsterdam: John Benjamins, 2001, vii+323 pp; hardbound, ISBN 90-272-2572-9 and 1-55619-792-6, $100.00, € 110.00","year":2002,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":13,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; University of Pennsylvania","keywords":"Honour; Sociology; Media studies; Political science; Law","authors":[{"name":"Warren J. Plath","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02927748883288635,"gpt":0.2119638490646402,"spread":0.1826863602317539,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005373801,0.002220632,0.001903389,0.004931834,0.001617414,0.008956538,0.002650982,0.002234509,0.01828921],"category_scores_gemma":[0.01058545,0.001027613,0.001069405,0.007494085,0.005291115,0.02323675,0.005173719,0.003554475,0.01855168],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001625721,"about_ca_system_score_gemma":0.002132762,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006451047,"about_ca_topic_score_gemma":0.01283483,"domain_scores_codex":[0.997221,0.0009778745,0.0002805398,0.0005633148,0.0008492219,0.0001080592],"domain_scores_gemma":[0.9899616,0.005245571,0.00051262,0.0008167813,0.002633029,0.0008303625],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004220065,0.00001086972,0.0002049242,0.0006996752,0.00002185406,0.00007427692,0.0007370842,0.0001392257,0.0004204666,0.005555779,0.8074936,0.1846002],"study_design_scores_gemma":[0.000007748016,0.000009592868,0.0003227541,0.0004065579,0.00001417012,0.0002459575,0.0005931858,0.0002308119,0.0002244964,0.008233084,0.9896932,0.00001835517],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.001684145,0.5540652,0.05853755,0.2302014,0.1277897,0.0001266526,0.001423919,0.001207778,0.02496354],"genre_scores_gemma":[0.01686466,0.5506685,0.07087468,0.03694977,0.1586917,0.0003313161,0.004756992,0.003263402,0.157599],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.01828921,"threshold_uncertainty_score":0.06118351,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1521465908","doi":"10.1162/089120101300346831","title":"Book Reviews","year":2001,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Language, Metaphor, and Cognition","field":"Psychology","cited_by":8,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science","authors":[{"name":"Graeme Hirst","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04113790453060918,"gpt":0.34483954249689,"spread":0.3037016379662808,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0005074873,0.0007081263,0.0007794102,0.003146709,0.000801998,0.003786743,0.00142969,0.001402542,0.4409769],"category_scores_gemma":[0.002249364,0.0003312571,0.0005030032,0.003753849,0.000536686,0.002529816,0.001148021,0.001713641,0.4230461],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001159643,"about_ca_system_score_gemma":0.001512072,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002638655,"about_ca_topic_score_gemma":0.006597462,"domain_scores_codex":[0.9994771,0.00004377377,0.00002075149,0.00008516949,0.0003267279,0.00004658703],"domain_scores_gemma":[0.9989611,0.0002104842,0.00005673655,0.00009302871,0.0004957977,0.0001829685],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000008630473,0.00001427755,0.00002736579,0.0001626621,0.000002224839,0.00001726019,0.00001943318,0.00003194081,0.0001082719,0.00361189,0.8975185,0.09847755],"study_design_scores_gemma":[0.000001759987,0.000002511222,0.00006527056,0.00007123148,0.00000113546,0.00003006808,0.00001123796,0.000007660838,0.00001781483,0.0007235822,0.9990661,0.000001743761],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"review","genre_scores_codex":[0.0003987297,0.1222511,0.002000866,0.01127422,0.01851772,0.00008476105,0.002066456,0.0008164279,0.8425897],"genre_scores_gemma":[0.001147167,0.03535794,0.001112536,0.003843562,0.003741412,0.00004556914,0.001637296,0.0002691749,0.9528454],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.4409769,"threshold_uncertainty_score":0.7973785,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2986265153","doi":"10.1162/coli_a_00367","title":"On the Linguistic Representational Power of Neural Machine Translation Models","year":2020,"lang":"en","type":"preprint","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Computer science; Natural language processing; Machine translation; Artificial intelligence; Interpretability; Semantics (computer science); Rule-based machine translation; Word (group theory); Contrast (vision); Linguistics","authors":[{"name":"Yonatan Belinkov","is_ca":true},{"name":"Nadir Durrani","is_ca":false},{"name":"Fahim Dalvi","is_ca":false},{"name":"Hassan Sajjad","is_ca":false},{"name":"James Glass","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05564198337220849,"gpt":0.3233539129337652,"spread":0.2677119295615567,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004271738,0.000984534,0.0007294641,0.0009573865,0.0005879973,0.002544622,0.001313966,0.001573735,0.002177163],"category_scores_gemma":[0.02894598,0.0006340668,0.0007393414,0.001106649,0.001836443,0.005028845,0.001631675,0.002975306,0.0007315818],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001296785,"about_ca_system_score_gemma":0.0007839593,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005089027,"about_ca_topic_score_gemma":0.004803835,"domain_scores_codex":[0.9983962,0.0009278156,0.00008252268,0.0002550364,0.0002538452,0.00008459042],"domain_scores_gemma":[0.9872896,0.009721479,0.0005635863,0.001447899,0.0008556562,0.0001217276],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002506825,0.00007495966,0.002945811,0.0002172839,0.0001474907,0.0001461853,0.0004006136,0.7520987,0.005330381,0.1041379,0.002280048,0.13197],"study_design_scores_gemma":[0.000007961075,0.00002995369,0.0003561497,0.00003673336,0.00001807575,0.00002778068,0.00002882234,0.9441421,0.00109564,0.05355465,0.0006911599,0.00001095171],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1415518,0.003968054,0.8315929,0.008338322,0.0001519323,0.00005080632,0.0006413536,0.001973169,0.0117316],"genre_scores_gemma":[0.910987,0.001844711,0.08250754,0.0006262787,0.0001819679,0.00008072705,0.0007805653,0.0002231962,0.002767991],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005089027,"threshold_uncertainty_score":0.02259135,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4391017630","doi":"10.1162/coli_a_00510","title":"Context-aware Transliteration of Romanized South Asian Languages","year":2023,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Institute for Catastrophic Loss Reduction","keywords":"Transliteration; Computer science; Romanization; Natural language processing; Context (archaeology); Artificial intelligence; Sentence; Word (group theory); Linguistics; History","authors":[{"name":"Christo Kirov","is_ca":false},{"name":"Cibu Johny","is_ca":false},{"name":"Anna Katanova","is_ca":false},{"name":"Alexander Gutkin","is_ca":false},{"name":"Brian Roark","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01638905617639773,"gpt":0.2955521738544959,"spread":0.2791631176780982,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001427873,0.001698411,0.001161224,0.001040873,0.0009955268,0.001261575,0.001122901,0.0007116284,0.005664816],"category_scores_gemma":[0.005619008,0.0003899137,0.001165528,0.0008862714,0.0004469261,0.00194704,0.002112757,0.001646776,0.006483205],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008058232,"about_ca_system_score_gemma":0.001420765,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007757365,"about_ca_topic_score_gemma":0.01398284,"domain_scores_codex":[0.9985709,0.0004901661,0.0001405739,0.0005720725,0.0001135139,0.0001127341],"domain_scores_gemma":[0.9970973,0.001161666,0.0001982802,0.000694044,0.0007234067,0.000125327],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001799345,0.0003660614,0.01171306,0.002209255,0.0005949676,0.001829821,0.002187291,0.05405498,0.06950298,0.001980131,0.04487806,0.808884],"study_design_scores_gemma":[0.0003247408,0.0007622305,0.01509288,0.0003722511,0.0007037601,0.001688891,0.003320197,0.6984207,0.1945854,0.007834885,0.07663178,0.0002623498],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.7515755,0.004817926,0.1586324,0.001304183,0.00158213,0.0005128802,0.01262435,0.05276458,0.01618602],"genre_scores_gemma":[0.8527513,0.0008536799,0.108625,0.0003733602,0.0002031145,0.0002233585,0.02753987,0.001527794,0.007902643],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.007757365,"threshold_uncertainty_score":0.0189507,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4416872671","doi":"10.1162/coli.a.583","title":"LLMs and Cultural Values: The Impact of Prompt Language and Explicit Cultural Framing","year":2025,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Mila - Quebec Artificial Intelligence Institute","funders":"Canada First Research Excellence Fund","keywords":"Framing (construction); Cultural diversity; Cultural values; Cultural bias; Hofstede's cultural dimensions theory; Perspective (graphical); Set (abstract data type); Cultural issues; Cultural psychology","authors":[{"name":"Bram Bulté","is_ca":false},{"name":"Ayla Rigouts Terryn","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03104474476676788,"gpt":0.4318547238264057,"spread":0.4008099790596378,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04709833,0.001033164,0.0006769628,0.0009717423,0.000883093,0.005422696,0.0007642927,0.001084359,0.004683445],"category_scores_gemma":[0.2376297,0.0004675722,0.0009006272,0.001599498,0.002211279,0.003936927,0.003894011,0.002583631,0.0006597272],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001545129,"about_ca_system_score_gemma":0.001382441,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00246549,"about_ca_topic_score_gemma":0.002195548,"domain_scores_codex":[0.9329697,0.05893893,0.00172442,0.003202742,0.002451809,0.0007122788],"domain_scores_gemma":[0.5311048,0.4348668,0.01275219,0.01474445,0.004976541,0.001555159],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.004437341,0.0008453458,0.7266786,0.001950408,0.00084685,0.0005828952,0.06991451,0.01869827,0.007713811,0.01106268,0.00339927,0.1538701],"study_design_scores_gemma":[0.0005302024,0.003710986,0.6349207,0.003283108,0.001688459,0.0007180386,0.1315732,0.1281829,0.0170911,0.04536349,0.03227883,0.0006589169],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9796156,0.0003008916,0.01329967,0.000660287,0.00007517185,0.0001381698,0.0004281821,0.0001374968,0.005344557],"genre_scores_gemma":[0.9951663,0.00005831769,0.00401159,0.0001294768,0.00001325262,0.0001120612,0.0002487335,0.00004131433,0.0002188903],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04709833,"threshold_uncertainty_score":0.2490828,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4404534114","doi":"10.1162/coli_a_00545","title":"The Emergence of Chunking Structures with Hierarchical RNN","year":2024,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Time Series Analysis and Forecasting","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canadian Institute for Advanced Research; University of Waterloo; University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada; Alberta Innovates; University of Alberta; DeepMind; Canadian Institute for Advanced Research","keywords":"Chunking (psychology); Computer science; Natural language processing; Artificial intelligence","authors":[{"name":"Zijun Wu","is_ca":true},{"name":"Anup Anand Deshmukh","is_ca":true},{"name":"Yongkang Wu","is_ca":false},{"name":"Jimmy Lin","is_ca":true},{"name":"Lili Mou","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01130154070976787,"gpt":0.251264422587314,"spread":0.2399628818775461,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00101369,0.0004275312,0.0003489245,0.000351799,0.0002744285,0.0003896313,0.0007326315,0.0004164169,0.0006241936],"category_scores_gemma":[0.004698767,0.0003589528,0.0003099872,0.0003281898,0.0005431269,0.001066823,0.0005416692,0.0009429739,0.0002405621],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000676911,"about_ca_system_score_gemma":0.000516836,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006840246,"about_ca_topic_score_gemma":0.007987703,"domain_scores_codex":[0.9996663,0.00014041,0.00001976948,0.00009687729,0.00004229443,0.00003430518],"domain_scores_gemma":[0.9976941,0.001381743,0.0002185399,0.0003450995,0.0003026015,0.00005794863],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002675041,0.0001050327,0.005610714,0.0001188431,0.0001300684,0.0003425189,0.0005996205,0.7366472,0.05130839,0.0143321,0.002595597,0.1879425],"study_design_scores_gemma":[0.000002202949,0.0000128607,0.000335199,0.000002632328,0.000004361086,0.000006778898,0.00000710711,0.994451,0.002230978,0.002818413,0.0001252387,0.000003211351],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3039787,0.0004611262,0.690044,0.0003303658,0.00007045723,0.00006095594,0.0001643796,0.002244219,0.002645939],"genre_scores_gemma":[0.9241479,0.00007578524,0.07441155,0.00005355646,0.00001982676,0.0000357793,0.0001841955,0.00009114685,0.0009802143],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006840246,"threshold_uncertainty_score":0.01360089,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4401117478","doi":"10.1162/coli_a_00533","title":"Exploring Temporal Sensitivity in the Brain Using Multi-timescale Language Models: An EEG Decoding Study","year":2024,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Cognitive Science and Mapping","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Electroencephalography; Decoding methods; Sensitivity (control systems); Computer science; Language model; Speech recognition; Artificial intelligence; Psychology; Neuroscience; Algorithm","authors":[{"name":"Sijie Ling","is_ca":true},{"name":"A. St. J. Murphy","is_ca":true},{"name":"Alona Fyshe","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2609003131987414,"gpt":0.3765898453257158,"spread":0.1156895321269744,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005509623,0.0004205335,0.0002554237,0.0003440769,0.00009757806,0.0005623199,0.0002540449,0.0003290062,0.001894521],"category_scores_gemma":[0.005558606,0.0001660008,0.0003910582,0.0004713114,0.0002768536,0.0006570507,0.0003788822,0.0008331588,0.0004443782],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001555949,"about_ca_system_score_gemma":0.0001447712,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001744312,"about_ca_topic_score_gemma":0.001786926,"domain_scores_codex":[0.9998384,0.00006540916,0.000008111284,0.00004564201,0.00002382505,0.00001867236],"domain_scores_gemma":[0.9988267,0.000895798,0.00008243828,0.00006748711,0.00009689058,0.00003065407],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002070166,0.0005948612,0.05615485,0.0009943305,0.0006450095,0.001488928,0.002034287,0.2845975,0.2877164,0.007338392,0.008461322,0.347904],"study_design_scores_gemma":[0.00004433974,0.0001993581,0.052409,0.00005596789,0.00009923266,0.0003183225,0.0002337188,0.9239294,0.01355711,0.007416663,0.001689162,0.00004777865],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8839622,0.001114688,0.1096902,0.0006434371,0.00009106306,0.00006297127,0.001343516,0.0004180905,0.002673934],"genre_scores_gemma":[0.9913271,0.0002900843,0.00723665,0.00003660358,0.00003867219,0.00002105614,0.0004959874,0.00005075535,0.0005030179],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001894521,"threshold_uncertainty_score":0.006337821,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4389613952","doi":"10.1162/coli_a_00501","title":"Stance Detection with Explanations","year":2023,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Automatic summarization; Artificial intelligence; Parsing; Benchmark (surveying); Machine learning; Focus (optics); Grammaticality; Natural language processing; Linguistics","authors":[{"name":"Rudra Saha","is_ca":true},{"name":"Laks V. S. Lakshmanan","is_ca":true},{"name":"Raymond T. Ng","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02863169177005188,"gpt":0.2680939184902881,"spread":0.2394622267202363,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001455506,0.001054901,0.0005549068,0.008578938,0.0006047592,0.001361616,0.0009264894,0.001586091,0.002708111],"category_scores_gemma":[0.01449951,0.0003262233,0.0009229287,0.00267564,0.0004728596,0.002426243,0.001366272,0.001457186,0.001841021],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005950945,"about_ca_system_score_gemma":0.001007865,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001321899,"about_ca_topic_score_gemma":0.001906432,"domain_scores_codex":[0.998616,0.0004314936,0.0001233209,0.0003223593,0.0004017636,0.0001050363],"domain_scores_gemma":[0.9911966,0.005205011,0.001311971,0.0007714393,0.001292784,0.0002222434],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008461485,0.0002800582,0.04792295,0.001342828,0.000310598,0.001159426,0.001210749,0.02434754,0.01736165,0.01881768,0.03806642,0.8483339],"study_design_scores_gemma":[0.00008834033,0.000214263,0.01591427,0.0004342632,0.0002232487,0.0007338021,0.0006767132,0.8560657,0.02420831,0.06648106,0.03488279,0.0000772199],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2096146,0.006397285,0.7385594,0.002657881,0.0004623472,0.0006710536,0.01378661,0.01765469,0.01019615],"genre_scores_gemma":[0.7088115,0.0008199917,0.2721676,0.0002413643,0.0003297065,0.0001996738,0.01473169,0.0002538991,0.002444467],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008578938,"threshold_uncertainty_score":0.009059548,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1538392035","doi":"10.1162/coli.2003.29.3.507","title":"Book Review","year":2003,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"","field":"","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science","authors":[{"name":"Gerald Penn","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02411177161073708,"gpt":0.3140361472330895,"spread":0.2899243756223524,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007433808,0.0006557263,0.0008778538,0.004101939,0.001127195,0.003539013,0.001301508,0.001534525,0.2944602],"category_scores_gemma":[0.00427631,0.0003174619,0.0004535963,0.004568199,0.0006709411,0.002603012,0.001451635,0.001930701,0.2405877],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00165826,"about_ca_system_score_gemma":0.003098559,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003753723,"about_ca_topic_score_gemma":0.009990189,"domain_scores_codex":[0.9994692,0.00006207394,0.00001855746,0.00006840056,0.0003321523,0.00004950298],"domain_scores_gemma":[0.9978766,0.0005246506,0.000106422,0.0001551127,0.001013744,0.00032334],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000005683262,0.00000633657,0.00002405365,0.0001102558,0.000001563028,0.00001440355,0.00001045466,0.0000214343,0.00003699098,0.001696705,0.925099,0.07297317],"study_design_scores_gemma":[0.000001531727,0.000002404953,0.00005363806,0.00008232118,0.000001738415,0.00002850038,0.0000114187,0.00001098888,0.00001907453,0.0005148419,0.999272,0.000001431841],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"review","genre_scores_codex":[0.0005735532,0.2245315,0.003475903,0.05473885,0.05352562,0.0001584641,0.001651374,0.0009650218,0.6603797],"genre_scores_gemma":[0.001525317,0.05412612,0.000819212,0.007369208,0.007256446,0.00005976369,0.000831939,0.000259777,0.9277523],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.2944602,"threshold_uncertainty_score":0.9850676,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4233779307","doi":"10.1162/coli.2000.27.1.140","title":"<i>Construing Experience through Meaning: A Language-based Approach to Cognition</i> M. A. K. Halliday and Christian M. I. M. Matthiessen London: Cassell (Open linguistics series, edited by Robin Fawcett), 1999, xiii+657 pp; paperback, ISBN 0-304-70490-3, $102.00, £65.00","year":2001,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Language, Metaphor, and Cognition","field":"Psychology","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Meaning (existential); Philosophy; Linguistics; Cognitive linguistics; Cognitive science; Cognition; Series (stratigraphy); Psychology; Epistemology; Neuroscience","authors":[{"name":"John F. Sowa","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02494040335504384,"gpt":0.2963689442215807,"spread":0.2714285408665368,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001221896,0.0007445776,0.0005217683,0.001858235,0.001254616,0.00883682,0.001398579,0.002078906,0.009239553],"category_scores_gemma":[0.001945626,0.0003788004,0.0007135073,0.003043625,0.01392588,0.01427746,0.002409275,0.002938199,0.001924256],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002147564,"about_ca_system_score_gemma":0.001606956,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003420332,"about_ca_topic_score_gemma":0.005028068,"domain_scores_codex":[0.9993592,0.0003726983,0.00003765793,0.00007325124,0.0001171452,0.00004007217],"domain_scores_gemma":[0.9992435,0.0005098146,0.00006714779,0.00005646236,0.00006978451,0.00005322146],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003479868,0.00001581115,0.0003082408,0.0008258775,0.00004493502,0.0002046494,0.00858758,0.0004979257,0.000905816,0.7925398,0.09327003,0.1027645],"study_design_scores_gemma":[0.000007809267,0.00001425536,0.0008533137,0.0005528429,0.0000164173,0.000479049,0.002955025,0.0006770815,0.0002620881,0.78605,0.2081114,0.00002074548],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.007015731,0.5201509,0.0975016,0.09103999,0.004186135,0.00006003531,0.0002448877,0.0004521504,0.2793487],"genre_scores_gemma":[0.4314135,0.3778149,0.09409912,0.0132327,0.008334087,0.000407945,0.0007825761,0.0005234064,0.07339159],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.009239553,"threshold_uncertainty_score":0.03090936,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4392367355","doi":"10.1162/coli_a_00512","title":"A Novel Alignment-based Approach for PARSEVAL Measuress","year":2023,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Parsing; Sentence; Natural language processing; Artificial intelligence; Lexical analysis; Parseval's theorem; Machine translation; Word (group theory); Linguistics","authors":[{"name":"Eunkyul Leah Jo","is_ca":true},{"name":"Angela Yoonseo Park","is_ca":true},{"name":"Jungyeul Park","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05511895696161671,"gpt":0.3166822825463177,"spread":0.261563325584701,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0124691,0.001852051,0.001684632,0.008922711,0.001489676,0.006395544,0.003358921,0.001710347,0.00350377],"category_scores_gemma":[0.06463953,0.0009641323,0.001411103,0.005906713,0.001942964,0.005739184,0.003608999,0.003224266,0.002086804],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001896358,"about_ca_system_score_gemma":0.001968803,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00167457,"about_ca_topic_score_gemma":0.001553168,"domain_scores_codex":[0.9777322,0.008279352,0.002496622,0.003024687,0.007912829,0.0005543766],"domain_scores_gemma":[0.9693965,0.01100757,0.002337451,0.006130014,0.01056578,0.0005626985],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003430912,0.000233144,0.006283846,0.0004260543,0.000276684,0.0001534232,0.0008340286,0.03747939,0.02726536,0.1748169,0.008157616,0.7437305],"study_design_scores_gemma":[0.00004753155,0.0003138505,0.003511581,0.0001393565,0.0001128999,0.0004589423,0.0001682673,0.7889706,0.03853088,0.1459067,0.0216522,0.0001872009],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003035631,0.00009625126,0.993215,0.00005673958,0.00004187143,0.00007835103,0.0001180456,0.002500813,0.0008572582],"genre_scores_gemma":[0.07486638,0.00005804887,0.9228679,0.00005621077,0.0000668062,0.0003044022,0.0003279857,0.00095135,0.0005009735],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0124691,"threshold_uncertainty_score":0.06594366,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2895279374","doi":"10.1162/coli_a_00363","title":"Scalable Micro-planned Generation of Discourse from Structured Data","year":2019,"lang":"en","type":"preprint","venue":"Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Interpretability; Natural language processing; Scalability; Artificial intelligence; Natural language generation; Sentence; Pipeline (software); Paragraph; Fluency; Text simplification; Natural language understanding; Robustness (evolution); Natural language; Programming language; Database","authors":[{"name":"Anirban Laha","is_ca":true},{"name":"Parag Jain","is_ca":false},{"name":"Abhijit Mishra","is_ca":false},{"name":"Karthik Sankaranarayanan","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08137583821686917,"gpt":0.3230469773943666,"spread":0.2416711391774974,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001824372,0.001003904,0.0007394824,0.001697289,0.0006893257,0.001653522,0.001691613,0.000788645,0.007651654],"category_scores_gemma":[0.009486726,0.0005961796,0.001222102,0.001325489,0.0007098676,0.002630942,0.002192564,0.001221408,0.004386621],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007892889,"about_ca_system_score_gemma":0.001933577,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002679475,"about_ca_topic_score_gemma":0.004420707,"domain_scores_codex":[0.9986174,0.0004328662,0.0001035243,0.0004844868,0.0003039159,0.00005773268],"domain_scores_gemma":[0.995653,0.002610545,0.0002216105,0.0007322887,0.0006560761,0.0001263688],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006422142,0.0002972847,0.003181103,0.001492601,0.0001536114,0.001017334,0.002867715,0.03825135,0.05415381,0.05084812,0.07385127,0.7732434],"study_design_scores_gemma":[0.0001494971,0.0001490954,0.0009240545,0.0001088072,0.00007599808,0.0003946891,0.0007509688,0.7789135,0.07595064,0.06052438,0.08197053,0.00008784291],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01159721,0.0002391173,0.9495867,0.0004327829,0.0001281203,0.0003131178,0.004045073,0.0310253,0.002632587],"genre_scores_gemma":[0.0986229,0.0001929592,0.8817062,0.0001621944,0.00007755972,0.0003898083,0.0137347,0.002019094,0.003094569],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007651654,"threshold_uncertainty_score":0.02559733,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4235980152","doi":"10.4018/978-1-4666-6042-7.ch012","title":"Demystifying Domain Specific Languages","year":2014,"lang":"en","type":"book-chapter","venue":"Computational Linguistics","topic":"Model-Driven Software Engineering Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"École de Technologie Supérieure","funders":"","keywords":"Vagueness; Digital subscriber line; Domain-specific language; Computer science; Domain (mathematical analysis); Software engineering; Programming language; Artificial intelligence; Mathematics","authors":[{"name":"Abdelilah Kahlaoui","is_ca":true},{"name":"Alain Abran","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02048454424811195,"gpt":0.2533815354641975,"spread":0.2328969912160855,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001596037,0.0006097081,0.000311504,0.001565101,0.0009184911,0.004535414,0.001012158,0.001126059,0.005291718],"category_scores_gemma":[0.002623998,0.0003368614,0.0005720576,0.0009779428,0.004254261,0.007770147,0.002634908,0.003948759,0.001770618],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00206214,"about_ca_system_score_gemma":0.001511169,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001570663,"about_ca_topic_score_gemma":0.002305846,"domain_scores_codex":[0.999222,0.0003437383,0.00005392156,0.0001096133,0.0002217213,0.00004892152],"domain_scores_gemma":[0.9985051,0.0007726271,0.00005767495,0.000363164,0.0002444261,0.00005700628],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000002405703,0.000004277295,0.00003461084,0.00005862764,0.000002332848,0.00003968481,0.00109219,0.0004879228,0.0005251681,0.9727656,0.005054381,0.0199328],"study_design_scores_gemma":[0.0000032065,0.000007048111,0.00006600211,0.0001513069,0.000006525223,0.0002191813,0.0004465749,0.00209295,0.0009793404,0.4222426,0.573775,0.00001030949],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.01308635,0.02118402,0.4497209,0.0137198,0.002127096,0.0001197102,0.0002904134,0.001764855,0.4979869],"genre_scores_gemma":[0.236462,0.03259464,0.4310163,0.006305357,0.001336185,0.0003667567,0.00117175,0.001973156,0.2887738],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005291718,"threshold_uncertainty_score":0.01770258,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4250326346","doi":"10.1162/coli_r_00106","title":"Briefly Noted","year":2012,"lang":"en","type":"article","venue":"Computational Linguistics","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Computer science","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.02806631387154629,"gpt":0.2845006114861258,"spread":0.2564342976145795,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.001380407,0.001467881,0.001152048,0.002747133,0.002735697,0.009323098,0.002348245,0.002828803,0.6100549],"category_scores_gemma":[0.007006588,0.0005536558,0.000996172,0.004495707,0.00124228,0.008018515,0.004504401,0.002905858,0.5779294],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00223901,"about_ca_system_score_gemma":0.002899832,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003938876,"about_ca_topic_score_gemma":0.007457406,"domain_scores_codex":[0.9977049,0.0003670275,0.0002292628,0.0006009305,0.0008299167,0.0002679038],"domain_scores_gemma":[0.9976014,0.000360916,0.00017464,0.0005522084,0.0009494567,0.0003613076],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004006861,0.00001935639,0.0002513766,0.0002455148,0.000007899786,0.0000657247,0.0002263994,0.00006459303,0.0002298888,0.04176563,0.860777,0.09630644],"study_design_scores_gemma":[0.000002108826,0.000005285101,0.00009324764,0.00005429198,0.000001651963,0.00004794729,0.00006550588,0.00002270482,0.00005733715,0.00231762,0.9973283,0.000004052216],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.0009761539,0.006965677,0.01175471,0.01178208,0.01756556,0.0002635785,0.01120612,0.00380304,0.9356831],"genre_scores_gemma":[0.00459736,0.003859221,0.005202997,0.003698939,0.001728641,0.0001751037,0.007398213,0.001583571,0.971756],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.3899451,"threshold_uncertainty_score":0.5562094,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}