{"id":"W4224933355","doi":"10.1109/icassp43922.2022.9746261","title":"Complex IRM-Aware Training for Voice Activity Detection Using Attention Model","year":2022,"lang":"en","type":"article","venue":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University; McGill University","funders":"","keywords":"Spectrogram; Computer science; Speech recognition; Voice activity detection; Noise (video); Artificial intelligence; Feature extraction; Pattern recognition (psychology); Artificial neural network; Speech processing; Image (mathematics)","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004473766,0.0006618545,0.0005648952,0.0004855889,0.0002000991,0.000370518,0.001101851,0.0005159845,0.00221963],"category_scores_gemma":[0.00138303,0.0002941831,0.0004357612,0.0003367409,0.0002670224,0.0007295638,0.0008319824,0.0008689397,0.0006999676],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00051143,"about_ca_system_score_gemma":0.0006262485,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004383158,"about_ca_topic_score_gemma":0.005942519,"domain_scores_codex":[0.9997357,0.00004553855,0.000014428,0.00009556132,0.00006113271,0.00004761901],"domain_scores_gemma":[0.9996283,0.0001781449,0.00002977304,0.00004230432,0.00009748639,0.00002397191],"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":[0.0003328599,0.0001608351,0.001757534,0.0001211456,0.00007490993,0.0001222916,0.00008907567,0.1035293,0.06984652,0.00205121,0.002801641,0.8191127],"study_design_scores_gemma":[0.000006788771,0.00005161732,0.0007481673,0.000006272057,0.00001645754,0.00005648605,0.000008052697,0.9874549,0.01027708,0.0006806647,0.0006861855,0.000007243852],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04753609,0.001034801,0.946401,0.0001719007,0.0001109953,0.00004501021,0.00009444785,0.002306179,0.002299615],"genre_scores_gemma":[0.8118731,0.0003790693,0.1826938,0.0003278571,0.00008307946,0.0001032018,0.0003791384,0.0001415209,0.004019327],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004383158,"threshold_uncertainty_score":0.008715272,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1525965693994567,"score_gpt":0.3427553212816414,"score_spread":0.1901587518821847,"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."}}