{"id":"W3135820585","doi":"10.1109/dasc-picom-cbdcom-cyberscitech52372.2021.00100","title":"Robust Black-box Watermarking for Deep Neural Network using Inverse Document Frequency","year":2021,"lang":"en","type":"article","venue":"2021 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech)","topic":"Advanced Steganography and Watermarking Techniques","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Imperial Bank of Commerce (Canada)","funders":"","keywords":"Digital watermarking; Computer science; Watermark; Robustness (evolution); Artificial intelligence; Deep neural networks; Noise (video); Artificial neural network; Speech recognition; Embedding; Image (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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication","open_science","research_integrity"],"consensus_categories":["metaepi_narrow","sts","research_integrity"],"category_scores_codex":[0.00408846,0.002495994,0.002714688,0.002219088,0.004036861,0.003598955,0.005831092,0.001356933,0.00006217489],"category_scores_gemma":[0.000895543,0.002376669,0.000407447,0.00275117,0.005729209,0.001309826,0.006147895,0.003518345,0.00002837436],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004828102,"about_ca_system_score_gemma":0.001037565,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003832941,"about_ca_topic_score_gemma":0.0002368426,"domain_scores_codex":[0.9851415,0.0007241322,0.002901604,0.006456437,0.001498357,0.00327799],"domain_scores_gemma":[0.9886227,0.002173744,0.002244032,0.003826076,0.002076845,0.001056615],"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":[0.0006107636,0.0008715123,0.005567677,0.0005295891,0.00101503,0.001235926,0.004938965,0.01364399,0.003576187,0.3109129,0.002995617,0.6541018],"study_design_scores_gemma":[0.003239716,0.003564153,0.0005321053,0.003731987,0.0003266273,0.001322,0.002256126,0.9024635,0.03164264,0.0192831,0.02718981,0.004448213],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6945546,0.001779614,0.2878561,0.003062415,0.007661948,0.00251603,0.0002059498,0.001200968,0.001162294],"genre_scores_gemma":[0.9696411,0.001044977,0.02222578,0.005542155,0.001038813,0.00003656751,0.0001584682,0.000159043,0.0001531282],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8888195,"threshold_uncertainty_score":0.9999395,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05792764735468442,"score_gpt":0.2850530221042313,"score_spread":0.2271253747495469,"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."}}