{"id":"W4297841830","doi":"10.21437/interspeech.2022-11066","title":"SoundChoice: Grapheme-to-Phoneme Models with Semantic Disambiguation","year":2022,"lang":"en","type":"article","venue":"Interspeech 2022","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University; Université de Montréal; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Grapheme; Computer science; 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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0003844465,0.0001726359,0.0001769186,0.0003567721,0.0003601504,0.0001801178,0.001038163,0.00002423078,0.001244545],"category_scores_gemma":[0.00003214419,0.000165218,0.00009836292,0.001119075,0.00003012318,0.0004259238,0.000501667,0.000242871,0.0001896384],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001750902,"about_ca_system_score_gemma":0.00006554546,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001104968,"about_ca_topic_score_gemma":0.00007115402,"domain_scores_codex":[0.9980057,0.0001385326,0.0002286568,0.0005497168,0.0007539185,0.0003234676],"domain_scores_gemma":[0.9989725,0.00009459046,0.0000883977,0.0006099125,0.0000988285,0.0001357826],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004918166,0.001970065,0.001109111,0.00008226929,0.0005698011,0.0007567065,0.0261226,0.009879524,0.02286215,0.1312202,0.1258645,0.6790712],"study_design_scores_gemma":[0.004395087,0.00321892,0.001925785,0.0001894835,0.0001523716,0.00144576,0.013965,0.674513,0.03725662,0.1501738,0.1091039,0.003660271],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1108626,0.00006651214,0.8707237,0.007036894,0.0007619806,0.0005013937,0.00001942812,0.0004821345,0.009545295],"genre_scores_gemma":[0.9490138,0.000006216032,0.04511077,0.003374516,0.00005219699,0.0003047211,0.00001413121,0.00002900288,0.002094601],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8381512,"threshold_uncertainty_score":0.9996685,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01880117531733043,"score_gpt":0.232155006894166,"score_spread":0.2133538315768356,"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."}}