{"id":"W4318541558","doi":"10.1145/3575693.3575738","title":"HuffDuff: Stealing Pruned DNNs from Sparse Accelerators","year":2023,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Exploit; Computer science; Edge device; Enhanced Data Rates for GSM Evolution; Side channel attack; Channel (broadcasting); State (computer science); Edge computing; Deep learning; Power (physics); Distributed computing; Computer network; Artificial intelligence; Computer security; Operating system; Cloud computing; Cryptography; Programming language","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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0003406186,0.0001773524,0.0002005851,0.0001720439,0.0002335277,0.0002573805,0.001231546,0.00008484639,0.0002168126],"category_scores_gemma":[0.000177515,0.0001573796,0.00007265303,0.0009796771,0.0000278875,0.0006572325,0.0008415598,0.0002762316,0.0009662335],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004160493,"about_ca_system_score_gemma":0.00005206618,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000306151,"about_ca_topic_score_gemma":0.000043384,"domain_scores_codex":[0.9983197,0.0001117074,0.0002481331,0.0005286857,0.0003744523,0.0004172715],"domain_scores_gemma":[0.9987654,0.0003420955,0.00008665395,0.0006404354,0.0000441253,0.0001212781],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003562541,0.000153808,0.06992537,0.00004129053,0.0002465123,0.0005989145,0.01031789,0.2722356,0.01052557,0.4812067,0.02558904,0.1291236],"study_design_scores_gemma":[0.0005601182,0.00003257134,0.01252795,0.00002351092,0.000009574338,0.000002141975,0.0003468657,0.9670814,0.003608317,0.008275173,0.007106608,0.0004257837],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1793496,0.000016907,0.8086714,0.001504072,0.001235306,0.0001322774,0.000002062192,0.0019761,0.007112344],"genre_scores_gemma":[0.9474859,0.000007072827,0.05012049,0.0004630999,0.0002932465,0.00001111345,0.00001325777,0.00002337279,0.001582458],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7681363,"threshold_uncertainty_score":0.9998116,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0318325200345579,"score_gpt":0.2771317073615917,"score_spread":0.2452991873270338,"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."}}