{"id":"W4298394324","doi":"10.1145/3552466.3556523","title":"A Comparative Study on Physical and Perceptual Features for Deepfake Audio Detection","year":2022,"lang":"en","type":"article","venue":"","topic":"Music and Audio Processing","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Computer science; Perception; Feature (linguistics); Artificial intelligence; Feature extraction; Speech recognition; Machine learning","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":[0.002159915,0.001336735,0.0008300316,0.002177146,0.0004017297,0.001061345,0.0006401776,0.0009559049,0.002198721],"category_scores_gemma":[0.007854424,0.0002051383,0.0007100063,0.0008353597,0.0004557332,0.001905191,0.001117777,0.001020814,0.0006962141],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004334916,"about_ca_system_score_gemma":0.000439678,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003593893,"about_ca_topic_score_gemma":0.005507919,"domain_scores_codex":[0.9987966,0.0002198384,0.00008147633,0.0002886857,0.0004476615,0.0001657359],"domain_scores_gemma":[0.9955248,0.00289431,0.0001884855,0.0003465668,0.0007897943,0.000256111],"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":[0.003610025,0.000610674,0.02725496,0.00104697,0.0003202945,0.0004397897,0.0001682949,0.03350546,0.0578072,0.00144546,0.006519695,0.8672711],"study_design_scores_gemma":[0.0002713959,0.00340074,0.123037,0.0003424659,0.0005539297,0.001333185,0.0008315872,0.7593438,0.09445658,0.003377498,0.01286245,0.0001893545],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8191164,0.01222923,0.1508474,0.000818741,0.0006853436,0.0002612676,0.002108559,0.003207664,0.01072541],"genre_scores_gemma":[0.9458989,0.001190666,0.04581378,0.0001277728,0.0001079118,0.00006672552,0.003887497,0.0001506742,0.002756033],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003593893,"threshold_uncertainty_score":0.01142281,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03890443042575863,"score_gpt":0.3082676728180773,"score_spread":0.2693632423923186,"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."}}