{"id":"W3048188704","doi":"10.1162/ling_a_00401","title":"Statistical Evidence for Learnable Lexical Subclasses in Japanese","year":2020,"lang":"en","type":"article","venue":"Linguistic Inquiry","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Phonotactics; Lexicon; Learnability; Linguistics; Computer science; Natural language processing; Phonology; Psychology; Artificial intelligence; Philosophy","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":[0.001053056,0.0002441073,0.0004881956,0.001909032,0.0009338559,0.00130783,0.0005539844,0.0004740902,0.002002864],"category_scores_gemma":[0.009094245,0.000458129,0.0003246271,0.001354375,0.001881777,0.001223277,0.001078531,0.0006357827,0.0003214782],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004405979,"about_ca_system_score_gemma":0.0004406848,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00856078,"about_ca_topic_score_gemma":0.01169476,"domain_scores_codex":[0.9993064,0.0001559828,0.00008821252,0.0003031807,0.00008886886,0.00005735278],"domain_scores_gemma":[0.9886321,0.006331464,0.001416842,0.001835356,0.001240281,0.0005438726],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00105534,0.0002994521,0.7584822,0.0005475041,0.0004003491,0.0009836829,0.01061679,0.01136294,0.09113906,0.01409213,0.001518617,0.109502],"study_design_scores_gemma":[0.00003696006,0.0002008795,0.9221509,0.00002907421,0.0001534887,0.0004321318,0.003269759,0.04731173,0.007261821,0.01758029,0.001488215,0.00008478479],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9959111,0.00007803074,0.002814628,0.00005076186,0.000003333662,0.000004887602,0.00006904482,0.00003449985,0.001033741],"genre_scores_gemma":[0.9988931,0.0000434234,0.0006305369,0.00001520215,0.000005297271,0.00000733758,0.0002253289,0.00001672543,0.0001632955],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.00856078,"threshold_uncertainty_score":0.01702189,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1347884067309446,"score_gpt":0.3918053928929698,"score_spread":0.2570169861620252,"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."}}