{"id":"W2738790619","doi":"10.7202/1040469ar","title":"Genre and Register in Comparable Corpora: An English/Spanish Contrastive Analysis","year":2017,"lang":"en","type":"article","venue":"Meta Journal des traducteurs","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Ottawa","funders":"Junta de Castilla y León","keywords":"Representativeness heuristic; Computer science; Register (sociolinguistics); Linguistics; Natural language processing; Corpus linguistics; Selection (genetic algorithm); Artificial intelligence; Mathematics; Statistics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.001170087,0.0002130181,0.0005842973,0.0003420907,0.0005520855,0.001993779,0.00156039,0.00006491385,0.0000120557],"category_scores_gemma":[0.0002478462,0.0001639936,0.0001636073,0.0003616109,0.0002820588,0.002929164,0.000149145,0.0005076575,9.617745e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005273888,"about_ca_system_score_gemma":0.00004537293,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001150186,"about_ca_topic_score_gemma":0.0002615006,"domain_scores_codex":[0.9983631,0.000237691,0.0003470842,0.0004111676,0.000295597,0.0003453068],"domain_scores_gemma":[0.9982492,0.00006451115,0.0004816519,0.0007038733,0.0002850074,0.000215718],"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":[0.000276035,0.0008662457,0.3227203,0.0001639394,0.008159732,0.003852488,0.01276318,0.0004113457,0.007405396,0.08742163,0.001185988,0.5547737],"study_design_scores_gemma":[0.002688057,0.0007424673,0.6064131,0.0002111123,0.004698548,0.001774733,0.0002542465,0.02625651,0.01077745,0.3407447,0.003643022,0.001796062],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4196742,0.04743944,0.5313742,0.000543923,0.0003446093,0.0001787028,0.000006347383,0.0001879154,0.0002507051],"genre_scores_gemma":[0.8048911,0.0008658208,0.1939229,0.0001133617,0.00008139413,0.000004595892,0.000001338712,0.00001030627,0.0001091484],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5529776,"threshold_uncertainty_score":0.9990423,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0464162821350621,"score_gpt":0.3024912447061955,"score_spread":0.2560749625711334,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}