{"id":"W4392910034","doi":"10.1109/icassp48485.2024.10447992","title":"VIC-KD: Variance-Invariance-Covariance Knowledge Distillation to Make Keyword Spotting More Robust Against Adversarial Attacks","year":2024,"lang":"en","type":"article","venue":"","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institut National de la Recherche Scientifique","funders":"","keywords":"Adversarial system; Covariance; Distillation; Variance (accounting); Computer science; Spotting; Artificial intelligence; Mathematics; Statistics; Chemistry; Chromatography","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0007722102,0.0005168401,0.0005289533,0.0005068093,0.0003238715,0.0007962614,0.00159337,0.0002222099,0.00008262919],"category_scores_gemma":[0.0002584083,0.0004919075,0.0003218046,0.003168842,0.00007670218,0.001446316,0.0008222019,0.0004439646,0.0005589159],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004520765,"about_ca_system_score_gemma":0.000259505,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004132891,"about_ca_topic_score_gemma":0.0001068627,"domain_scores_codex":[0.9961448,0.000120981,0.0007668046,0.001605002,0.0005774269,0.0007849478],"domain_scores_gemma":[0.9975398,0.00032257,0.0001541255,0.001405269,0.0002842296,0.0002940213],"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":[0.00005315951,0.0001806002,0.0002777568,0.0002182598,0.0003229857,0.0002956416,0.003558591,0.06043991,0.005005901,0.455086,0.01390616,0.460655],"study_design_scores_gemma":[0.0004094706,0.0000841298,0.0008421298,0.0004949337,0.00007456205,0.00002445268,0.00006952921,0.8243631,0.002860661,0.008090473,0.1615741,0.001112548],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.000510551,0.0004172109,0.9645331,0.002744438,0.001419435,0.0005319629,0.00001103751,0.002847459,0.02698483],"genre_scores_gemma":[0.3556755,0.00005443959,0.6381239,0.0007196023,0.0008628881,0.00009756957,0.00002428312,0.00006370834,0.004378081],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.7639231,"threshold_uncertainty_score":0.9997532,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02042862100012554,"score_gpt":0.305732944577786,"score_spread":0.2853043235776605,"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."}}