{"id":"W6976990631","doi":"10.6082/uchicago.7969","title":"One Language, Different Romanization Rules: The 20th Anniversary of the Pinyin Conversion Project","year":2023,"lang":"en","type":"article","venue":"Knowledge@UChicago (University of Chicago)","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Pinyin; Romanization; Power (physics); Register (sociolinguistics)","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":[],"consensus_categories":[],"category_scores_codex":[0.0003185623,0.00017413,0.0002760951,0.0002708707,0.0003956117,0.00003108735,0.002452078,0.0001423782,0.0000280416],"category_scores_gemma":[0.00009229383,0.0001288201,0.0001763808,0.001611568,0.000306061,0.0003569585,0.001417417,0.0003439789,0.00003952752],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006983239,"about_ca_system_score_gemma":0.00016082,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002149812,"about_ca_topic_score_gemma":0.0001702083,"domain_scores_codex":[0.998526,0.000262116,0.000174788,0.0003702415,0.0004134197,0.0002534289],"domain_scores_gemma":[0.9983774,0.0001896109,0.0002909086,0.000864537,0.000228658,0.00004888038],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.0004462953,0.002541822,0.007428804,0.002946927,0.0008985144,0.0001414421,0.3048907,0.00008237186,0.1479619,0.2389541,0.08380329,0.2099039],"study_design_scores_gemma":[0.008956478,0.001604816,0.4102939,0.004413823,0.001014038,0.0000741156,0.04117409,0.174357,0.2711713,0.05658186,0.0266863,0.003672313],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9337515,0.001739207,0.05156588,0.005331822,0.0005170444,0.001081003,0.00005369141,0.001131732,0.004828082],"genre_scores_gemma":[0.992429,0.0001057351,0.006121255,0.0000445511,0.00003807731,7.801244e-7,0.00001936043,0.00001349686,0.0012277],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4028651,"threshold_uncertainty_score":0.5253133,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01536798602523691,"score_gpt":0.2389132332534188,"score_spread":0.2235452472281819,"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."}}