{"id":"W2937447982","doi":"10.1109/icassp.2019.8682202","title":"Attacks on Digital Watermarks for Deep Neural Networks","year":2019,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":79,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Digital watermarking; Computer science; Watermark; Artificial intelligence; Artificial neural network; Countermeasure; Deep learning; Embedding; Property (philosophy); Deep neural networks; Task (project management); Scheme (mathematics); Data mining; Machine learning; Computer security; Image (mathematics); Mathematics","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":[0.001746627,0.0007996751,0.0005752561,0.001033208,0.0003934186,0.001053531,0.0007350862,0.001585542,0.001558219],"category_scores_gemma":[0.01433619,0.000438997,0.0008848335,0.0005521796,0.00170565,0.003484931,0.001963772,0.001670854,0.0003956519],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001110621,"about_ca_system_score_gemma":0.0003335092,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005642336,"about_ca_topic_score_gemma":0.0004502948,"domain_scores_codex":[0.9986981,0.0003318485,0.00008888704,0.0001937637,0.0005486323,0.0001387362],"domain_scores_gemma":[0.9951653,0.002635566,0.0005801523,0.001254596,0.0002957282,0.00006866407],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008579124,0.0001859374,0.002368743,0.0003127485,0.000201396,0.0006761138,0.0002350654,0.5140906,0.09793445,0.2027226,0.003437083,0.1769773],"study_design_scores_gemma":[0.00002059999,0.00006794272,0.0003121722,0.00003618859,0.00001783937,0.0001401741,0.00001370197,0.9423977,0.02701096,0.02861152,0.001355162,0.00001591576],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.138659,0.0007516105,0.8511213,0.001134812,0.00018677,0.00008210263,0.000167044,0.001511136,0.006386308],"genre_scores_gemma":[0.9358884,0.0005103704,0.06032162,0.0002302872,0.00006133468,0.00004745927,0.0001283333,0.000121289,0.002690949],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001746627,"threshold_uncertainty_score":0.00923717,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007538449835468789,"score_gpt":0.2413116925341822,"score_spread":0.2337732426987134,"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."}}