{"id":"W1990591837","doi":"10.1353/lan.2014.0076","title":"How to investigate linguistic diversity: Lessons from the Pacific Northwest","year":2014,"lang":"en","type":"article","venue":"Language","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":57,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Falsifiability; Syntax; Linguistics; Diversity (politics); Lexical diversity; Typology; Rule-based machine translation; Semantics (computer science); Historical linguistics; Computer science; Scale (ratio); Test (biology); Relative clause; Linguistic universal; Natural language processing; Artificial intelligence; Theoretical linguistics; Sociology; Geography; Epistemology; Philosophy; Archaeology; Ecology; Vocabulary","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01416453,0.0004566307,0.0007244739,0.003166746,0.008373997,0.004416541,0.002318386,0.001789323,0.002281676],"category_scores_gemma":[0.02590174,0.0003792613,0.0003477214,0.005522763,0.0164616,0.01398424,0.007539113,0.003854646,0.000202135],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004889095,"about_ca_system_score_gemma":0.005890251,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.08786052,"about_ca_topic_score_gemma":0.1397455,"domain_scores_codex":[0.9923459,0.005846518,0.0001936973,0.0004877161,0.0007571554,0.0003690012],"domain_scores_gemma":[0.9692573,0.02437776,0.001091176,0.002216557,0.002303136,0.0007540535],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.0002014377,0.0002584039,0.07271968,0.0007086844,0.00009057517,0.004413503,0.5426834,0.002990635,0.002208378,0.1335688,0.005116425,0.2350401],"study_design_scores_gemma":[0.00005350145,0.00009923893,0.04362947,0.001438173,0.00004990608,0.001866303,0.6102156,0.004630827,0.002788189,0.2484296,0.08669904,0.0001001193],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.8449806,0.005749624,0.04706172,0.03659184,0.00007795156,0.0002596267,0.0002588539,0.00008300543,0.06493668],"genre_scores_gemma":[0.9531337,0.002932303,0.03939209,0.001266205,0.00003651834,0.0002324192,0.0002077052,0.00007335306,0.002725669],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.08786052,"threshold_uncertainty_score":0.1746982,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01536529150023137,"score_gpt":0.2480672163176107,"score_spread":0.2327019248173793,"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."}}