{"id":"W4372260241","doi":"10.1109/icassp49357.2023.10097115","title":"SARdBScene: Dataset and Resnet Baseline for Audio Scene Source Counting and Analysis","year":2023,"lang":"en","type":"article","venue":"","topic":"Music and Audio Processing","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Baseline (sea); Computer science; Task (project management); Audio analyzer; Audio signal processing; Speech recognition; Artificial intelligence; Audio signal; Speech coding; Engineering","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":[0.001255518,0.004206301,0.001665696,0.004563139,0.001568446,0.001511241,0.00415946,0.002487328,0.009672198],"category_scores_gemma":[0.004070584,0.0005054994,0.001388416,0.003490577,0.0008453006,0.002337194,0.00246516,0.002059502,0.01543438],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001095688,"about_ca_system_score_gemma":0.001786263,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02179604,"about_ca_topic_score_gemma":0.05339777,"domain_scores_codex":[0.9980044,0.0002270159,0.0001802398,0.0006695866,0.0006665755,0.0002522288],"domain_scores_gemma":[0.9984717,0.0002491136,0.00008893845,0.0005258563,0.0005282199,0.0001361452],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001155584,0.0009104013,0.004379883,0.002305829,0.0002900319,0.0006416059,0.0002329667,0.008226807,0.01728897,0.001885055,0.7865355,0.1761473],"study_design_scores_gemma":[0.0006381568,0.000885185,0.0343142,0.0007700753,0.0004599167,0.002995444,0.001296881,0.07160382,0.04377123,0.007929021,0.8349159,0.0004202524],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.05541141,0.004165902,0.0562936,0.0008229894,0.002029509,0.001440276,0.8026071,0.05165574,0.02557356],"genre_scores_gemma":[0.02201817,0.0004435167,0.03335512,0.0002449705,0.0001393562,0.0006960489,0.9380698,0.0008808151,0.004152216],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.02179604,"threshold_uncertainty_score":0.04333836,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02702748367579776,"score_gpt":0.2825128516751092,"score_spread":0.2554853679993114,"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."}}