{"id":"W4417072973","doi":"10.1016/j.csl.2025.101923","title":"Enhanced audio-visual speech enhancement with posterior sampling methods in recurrent variational autoencoders","year":2025,"lang":"en","type":"article","venue":"Computer Speech & Language","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Speech enhancement; Inference; Intelligibility (philosophy); Sampling (signal processing); Autoencoder; Pattern recognition (psychology); Noise reduction; Posterior probability","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.000799565,0.0003719021,0.0004617402,0.0004810054,0.0001661228,0.00045981,0.001077984,0.0001112971,0.0000758449],"category_scores_gemma":[0.00004882092,0.0003336559,0.00009472396,0.001263348,0.00005604035,0.0005540515,0.0006009833,0.0003795483,0.00003041851],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002683921,"about_ca_system_score_gemma":0.0003707639,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003643167,"about_ca_topic_score_gemma":0.0000645201,"domain_scores_codex":[0.9971666,0.0001975864,0.000563702,0.0009488765,0.0004635885,0.0006596668],"domain_scores_gemma":[0.9987049,0.0002011702,0.0002130707,0.0006101768,0.0001471049,0.0001235591],"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":[0.00004009322,0.000194691,0.0001363896,0.00007184671,0.00005141617,0.0001125849,0.002022257,0.000339766,0.03157797,0.0004378413,0.0001255813,0.9648896],"study_design_scores_gemma":[0.00249995,0.0005374305,0.007237718,0.0009358193,0.00003435303,0.00011778,0.000239569,0.1259224,0.8586945,0.001464715,0.001301152,0.00101456],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06439032,0.0005260485,0.9309851,0.0007837521,0.001024128,0.0003621141,0.000001949406,0.0002445806,0.001682039],"genre_scores_gemma":[0.04759343,0.000009897601,0.9504756,0.001279722,0.0001965227,0.00003729799,0.00001546919,0.00001835935,0.0003737063],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.963875,"threshold_uncertainty_score":0.9999115,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01555674382515423,"score_gpt":0.3504463130887455,"score_spread":0.3348895692635913,"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."}}