{"id":"W4299806750","doi":"10.1109/icme52920.2022.9859621","title":"Mel-Spectrogram Image-Based End-to-End Audio Deepfake Detection Under Channel-Mismatched Conditions","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Multimedia and Expo (ICME)","topic":"Digital Media Forensic Detection","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"Computer Research Institute of Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Spectrogram; Codec; Robustness (evolution); Speech recognition; Spoofing attack; Speech coding; Jitter; Artificial intelligence; Computer network; Telecommunications","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","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0003512594,0.0003185154,0.0002628689,0.000593159,0.0004144254,0.0004525089,0.0008353377,0.00008014173,0.0009681969],"category_scores_gemma":[0.000116479,0.0003477023,0.0001343405,0.0005128428,0.0001503545,0.0006242229,0.0002796377,0.0005046245,0.0002154996],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000379543,"about_ca_system_score_gemma":0.0001595236,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001085458,"about_ca_topic_score_gemma":0.0001712744,"domain_scores_codex":[0.9970674,0.0001389021,0.0004029442,0.0008234914,0.001122107,0.0004451682],"domain_scores_gemma":[0.9985666,0.0002399652,0.0001951869,0.0004349027,0.0002554829,0.0003078115],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008895725,0.001943498,0.0001814072,0.00005593919,0.0005737035,0.0003983148,0.007051808,0.008986235,0.4383278,0.06764644,0.007835049,0.4661102],"study_design_scores_gemma":[0.002579371,0.001548784,0.001823567,0.00005560038,0.00003654623,0.0001632667,0.001669053,0.8502179,0.1150673,0.01967945,0.006188294,0.000970853],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3805448,0.00004209,0.5746018,0.01271857,0.01745353,0.001475587,0.0005067455,0.001062477,0.01159449],"genre_scores_gemma":[0.9938688,0.00001199583,0.003382162,0.001162025,0.0002576289,0.0005943764,0.0001691177,0.00002919544,0.0005247148],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8412316,"threshold_uncertainty_score":0.999945,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03054497750112798,"score_gpt":0.2803384128655858,"score_spread":0.2497934353644578,"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."}}