{"id":"W2929750792","doi":"10.1109/jstsp.2019.2909193","title":"Unsupervised Low Latency Speech Enhancement With RT-GCC-NMF","year":2019,"lang":"en","type":"article","venue":"IEEE Journal of Selected Topics in Signal Processing","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Sherbrooke","funders":"Natural Sciences and Engineering Research Council of Canada; CHIST-ERA; Agence Nationale de la Recherche","keywords":"Non-negative matrix factorization; Computer science; Speech enhancement; Speech recognition; Intelligibility (philosophy); Spectrogram; Short-time Fourier transform; Source separation; PESQ; Artificial intelligence; Pattern recognition (psychology); Matrix decomposition; Fourier transform; Noise reduction; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0004756546,0.0002389958,0.0004098407,0.0003485382,0.000107923,0.0003466676,0.0008736254,0.0001086234,0.00005643585],"category_scores_gemma":[0.00002540414,0.000186134,0.00005793116,0.001457502,0.00003922517,0.001437464,0.00004813747,0.0006509989,0.00001633806],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001553345,"about_ca_system_score_gemma":0.000873281,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00000427033,"about_ca_topic_score_gemma":0.000005883234,"domain_scores_codex":[0.9976987,0.00006500136,0.0006901727,0.0003330505,0.0007197035,0.0004933448],"domain_scores_gemma":[0.9982268,0.00005205659,0.0005580396,0.0001994525,0.0008318267,0.000131849],"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":[0.0002345764,0.0005860877,0.02240307,0.0005717119,0.0000787203,0.0005619608,0.002329424,0.001890706,0.3719114,0.00006633013,0.0001531096,0.5992129],"study_design_scores_gemma":[0.003033809,0.0009965255,0.00244822,0.002026904,0.00002275393,0.0005815938,0.0001055711,0.01241229,0.9754927,0.002017973,0.0003470306,0.0005145918],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8064843,0.0007552647,0.1907296,0.0004780547,0.0002517471,0.0001360283,1.957709e-7,0.00003390367,0.001130833],"genre_scores_gemma":[0.92775,0.00004103501,0.07124154,0.0002630304,0.000321521,0.000001826896,4.657211e-7,0.00001780007,0.0003628252],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6035813,"threshold_uncertainty_score":0.7590322,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009430378157480598,"score_gpt":0.2302141330364347,"score_spread":0.2207837548789541,"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."}}