{"id":"W2941649104","doi":"10.1109/tifs.2019.2956591","title":"A Robust Approach for Securing Audio Classification Against Adversarial Attacks","year":2019,"lang":"en","type":"preprint","venue":"IEEE Transactions on Information Forensics and Security","topic":"Digital Media Forensic Detection","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"École de Technologie Supérieure; Université du Québec à Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Spectrogram; Artificial intelligence; Codebook; Audio signal; Speech recognition; Deep learning; Smoothing; Pattern recognition (psychology); Preprocessor; Support vector machine; Feature learning; Machine learning; Speech coding; Computer vision","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007883538,0.000908542,0.000689265,0.0009023392,0.0003279021,0.0008446597,0.001197456,0.001046669,0.001756507],"category_scores_gemma":[0.00265735,0.0003541856,0.0006838089,0.000435801,0.000848017,0.001292884,0.001575891,0.001554027,0.001175978],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004972895,"about_ca_system_score_gemma":0.0005400713,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008044806,"about_ca_topic_score_gemma":0.0008547136,"domain_scores_codex":[0.9989042,0.0001524053,0.00006385837,0.0002014146,0.0005764774,0.0001016473],"domain_scores_gemma":[0.9989833,0.0002781893,0.0001617476,0.0003217823,0.0002066111,0.00004839694],"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":[0.0003769407,0.0001157848,0.0007302337,0.0001279174,0.0001075303,0.0003001289,0.00008578778,0.3738408,0.1554426,0.01939866,0.004051853,0.4454217],"study_design_scores_gemma":[0.000006504115,0.0001052401,0.0002699185,0.00001085871,0.00001516294,0.0001322117,0.00001166452,0.9728705,0.02027841,0.004195301,0.002088207,0.00001592101],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009927457,0.0003034883,0.9870948,0.0001183507,0.00006537538,0.00003441593,0.00004095956,0.00137058,0.001044555],"genre_scores_gemma":[0.657364,0.0006917604,0.3355744,0.0002863835,0.0002203873,0.0001210853,0.0002746557,0.0001629726,0.005304463],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001756507,"threshold_uncertainty_score":0.005876124,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03006476488899338,"score_gpt":0.2290495888172815,"score_spread":0.1989848239282881,"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."}}