{"id":"W3033311119","doi":"10.1109/eurosp51992.2021.00024","title":"Sponge Examples: Energy-Latency Attacks on Neural Networks","year":2021,"lang":"en","type":"preprint","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Toronto","funders":"Engineering and Physical Sciences Research Council; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Computer science; Energy consumption; Latency (audio); Exploit; Application-specific integrated circuit; Artificial neural network; Efficient energy use; Software portability; Energy (signal processing); Embedded system; Artificial intelligence; Computer security; Operating system; Engineering; Telecommunications","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":[0.001118162,0.0007939931,0.0005565492,0.0003782005,0.0004703375,0.0007482529,0.001027747,0.00109897,0.002141821],"category_scores_gemma":[0.008417485,0.0003426199,0.000750614,0.000359012,0.001524455,0.00331747,0.001941346,0.001931299,0.0003961014],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000868938,"about_ca_system_score_gemma":0.0003979011,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000651015,"about_ca_topic_score_gemma":0.0006149427,"domain_scores_codex":[0.9985785,0.0005455312,0.00008639568,0.0001562882,0.0004263647,0.0002068578],"domain_scores_gemma":[0.9945865,0.003129927,0.0004490005,0.001462532,0.0002800757,0.00009199278],"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":[0.001850893,0.0002073593,0.005103647,0.0003685994,0.0002488336,0.001224454,0.0003888767,0.6821768,0.1036307,0.09295192,0.01546813,0.0963798],"study_design_scores_gemma":[0.0000362688,0.0002078074,0.0005523381,0.00002502304,0.0000235054,0.0003386585,0.0000497701,0.9338548,0.03540995,0.02622483,0.003249987,0.00002704906],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4390687,0.001099424,0.5358031,0.003157687,0.0003192137,0.0001859902,0.0004150376,0.00700354,0.01294735],"genre_scores_gemma":[0.9789337,0.0001639854,0.01898968,0.0003361463,0.00002537569,0.00005286362,0.0000858771,0.0001417713,0.001270581],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002141821,"threshold_uncertainty_score":0.007165134,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02620811630047215,"score_gpt":0.268355225118932,"score_spread":0.2421471088184599,"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."}}