{"id":"W2898954967","doi":"10.1515/cllt-2018-0033","title":"An information-theoretic view on language complexity and register variation: Compressing naturalistic corpus data","year":2018,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Variation (astronomy); Register (sociolinguistics); Linguistics; Formality; Context (archaeology); Conversation; Sentence; Natural language processing; Artificial intelligence","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001216151,0.0003311959,0.0003516678,0.000176654,0.0005610811,0.0009320101,0.001450587,0.0001447073,0.00001722256],"category_scores_gemma":[0.01099728,0.0002868888,0.00002515758,0.0002289266,0.0008169664,0.0002137674,0.000759956,0.0003985438,0.00001525213],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004191221,"about_ca_system_score_gemma":0.0001031485,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007793528,"about_ca_topic_score_gemma":0.00001312971,"domain_scores_codex":[0.99793,0.0002371494,0.0004901536,0.0006482003,0.0003348406,0.0003596619],"domain_scores_gemma":[0.9962308,0.0007868832,0.0003598877,0.001714887,0.0006857817,0.0002217359],"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":[0.00005102319,0.00003556044,0.00003230303,0.0001325104,0.0000145536,0.0000342098,0.002961298,6.500607e-7,0.00005628147,0.9707473,0.0001033366,0.02583099],"study_design_scores_gemma":[0.0005525411,0.0004150988,0.0004373761,0.0005123899,0.00009403037,0.0001065915,0.00007372833,0.09118986,0.0003319408,0.8788587,0.0267293,0.0006984137],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00342656,0.008459058,0.9663231,0.0001925647,0.004094658,0.0006030222,0.0002608452,0.001296958,0.01534326],"genre_scores_gemma":[0.8274653,0.00004899852,0.1697605,0.001265582,0.001252063,0.000004819646,0.0001314503,0.00002134982,0.00004994979],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8240387,"threshold_uncertainty_score":0.9999583,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0309864596405754,"score_gpt":0.3110136154834371,"score_spread":0.2800271558428618,"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."}}