{"id":"W4407221739","doi":"10.48550/arxiv.2502.03128","title":"Metis: A Foundation Speech Generation Model with Masked Generative Pre-training","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Speech and dialogue systems","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Metis; Generative grammar; Foundation (evidence); Training (meteorology); Speech recognition; Computer science; Psychology; Artificial intelligence; History; Geography; World Wide Web","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001094542,0.001136112,0.0007678684,0.000462319,0.0003869937,0.0009677445,0.002218101,0.0009812411,0.006178802],"category_scores_gemma":[0.003073975,0.0006948426,0.001095948,0.0003756444,0.0006373121,0.001320153,0.002073281,0.001971101,0.003437495],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006546343,"about_ca_system_score_gemma":0.001270721,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003702056,"about_ca_topic_score_gemma":0.006649618,"domain_scores_codex":[0.9995223,0.0001231282,0.00002655795,0.0001611269,0.0001090877,0.00005774203],"domain_scores_gemma":[0.9992173,0.0004136756,0.0000421851,0.00012883,0.0001373576,0.00006067486],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006361789,0.0001885365,0.001934364,0.0002939411,0.0001964165,0.0003662453,0.0004747861,0.5927282,0.02502227,0.0315001,0.01689367,0.3297653],"study_design_scores_gemma":[0.00001619662,0.00004814821,0.0001167403,0.00000965968,0.00001276724,0.00006001583,0.00001136458,0.9896771,0.002338428,0.005332861,0.002365382,0.00001130237],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007802793,0.0002879241,0.984388,0.0002113835,0.0001059177,0.0001088028,0.0004757127,0.003989192,0.002630211],"genre_scores_gemma":[0.4902466,0.0005261471,0.4849382,0.0007257634,0.0002164307,0.001059649,0.004259859,0.001733627,0.01629365],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006178802,"threshold_uncertainty_score":0.02067012,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09950370094731856,"score_gpt":0.298433028713288,"score_spread":0.1989293277659695,"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."}}