{"id":"W4391421598","doi":"10.1016/j.csl.2024.101685","title":"Objective and subjective evaluation of speech enhancement methods in the UDASE task of the 7th CHiME challenge","year":2024,"lang":"en","type":"article","venue":"Computer Speech & Language","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"ca_institutions":"Google (Canada)","funders":"Agence Nationale de la Recherche; University of Sheffield","keywords":"Computer science; Speech recognition; Speech enhancement; Leverage (statistics); Task (project management); Noise (video); Distortion (music); Artificial intelligence; Domain adaptation; Noise reduction","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":[0.002925975,0.0001744558,0.0002457529,0.0001607449,0.00006506452,0.0001201786,0.0008336949,0.00005908322,0.00001557025],"category_scores_gemma":[0.00007531764,0.0001047922,0.00009061973,0.000797609,0.0000717138,0.0002899618,0.0003944202,0.0002541829,0.000003483703],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000767781,"about_ca_system_score_gemma":0.0001684763,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006617595,"about_ca_topic_score_gemma":0.00003866739,"domain_scores_codex":[0.9977129,0.0006705476,0.000317315,0.0004386579,0.0006351894,0.0002254529],"domain_scores_gemma":[0.9987978,0.000294551,0.0001441882,0.0005836808,0.0001488081,0.00003097924],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000004736234,0.00007511106,0.0000705994,0.00008462054,0.00004088987,0.00002448526,0.02950769,0.00003337633,0.01615166,0.0006567542,0.00003786667,0.9533122],"study_design_scores_gemma":[0.0006639838,0.0002576156,0.008803505,0.0004248424,0.00007490985,0.0000978547,0.00144918,0.143248,0.8376061,0.006938418,0.0001762389,0.0002593353],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4492462,0.009737152,0.5358797,0.0009846133,0.000667694,0.0007839339,0.000005238911,0.00005598168,0.002639566],"genre_scores_gemma":[0.8776155,0.0000340796,0.1220206,0.0001327193,0.000141777,0.00002060785,0.000001455243,0.000009298561,0.00002395668],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9530529,"threshold_uncertainty_score":0.4273303,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02378122438405992,"score_gpt":0.3505703286582682,"score_spread":0.3267891042742083,"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."}}