{"id":"W3217650841","doi":"10.1007/978-3-031-17143-7_19","title":"Real-Time Adversarial Perturbations Against Deep Reinforcement Learning Policies: Attacks and Defenses","year":2022,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Adversarial system; Computer science; Reinforcement learning; Artificial intelligence; Computer security","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":[0.001882721,0.00113144,0.0008429194,0.0004292791,0.0004044587,0.001040098,0.001022706,0.001766892,0.00190573],"category_scores_gemma":[0.00766368,0.0004243766,0.0005716995,0.0004626595,0.001623875,0.001797812,0.002518326,0.003631673,0.000520824],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000945474,"about_ca_system_score_gemma":0.0006573024,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006204593,"about_ca_topic_score_gemma":0.0004340622,"domain_scores_codex":[0.9988096,0.0003867014,0.00004125381,0.0001804014,0.0004036504,0.0001785289],"domain_scores_gemma":[0.9962374,0.002519845,0.0002920713,0.0005754762,0.0002551513,0.0001201419],"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":[0.0003781078,0.000104044,0.000479088,0.0001034648,0.00008549724,0.0001108692,0.00007274334,0.7299228,0.01127119,0.1317469,0.006261737,0.1194636],"study_design_scores_gemma":[0.000007945107,0.00006021032,0.000103244,0.00001468714,0.000006676396,0.00005990715,0.00001045477,0.9571596,0.002346989,0.03923658,0.0009854641,0.000008298448],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03360349,0.001973915,0.94732,0.001423901,0.0003528557,0.00008722227,0.00006872464,0.001095154,0.01407463],"genre_scores_gemma":[0.9274696,0.001388378,0.06089395,0.0004593671,0.0002063602,0.0001092407,0.00008875432,0.0001343199,0.009250124],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00190573,"threshold_uncertainty_score":0.009956896,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01209706532778705,"score_gpt":0.249908953409619,"score_spread":0.2378118880818319,"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."}}