English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep Neural Network (DNN) models are often deployed in resource-sharingclouds as Machine Learning as a Service (MLaaS) to provide inferenceservices.To steal model architectures that are of valuable intellectualproperties, a class of attacks has been proposed via different side-channelleakage, posing a serious security challenge to MLaaS.  Also targeting MLaaS, we propose a new end-to-end attack, DeepTheft, toaccurately recover complex DNN model architectures on general processors viathe RAPL-based power side channel. However, an attacker can acquire only a lowsampling rate (1 KHz) of the time-series energy traces from the RAPL interface,rendering existing techniques ineffective in stealing large and deep DNNmodels. To this end, we design a novel and generic learning-based frameworkconsisting of a set of meta-models, based on which DeepTheft is demonstrated tohave high accuracy in recovering a large number (thousands) of modelsarchitectures from different model families including the deepest ResNet152.Particularly, DeepTheft has achieved a Levenshtein Distance Accuracy of 99.75%in recovering network structures, and a weighted average F1 score of 99.60% inrecovering diverse layer-wise hyperparameters. Besides, our proposed learningframework is general to other time-series side-channel signals. To validate itsgeneralization, another existing side channel is exploited, i.e., CPUfrequency. Different from RAPL, CPU frequency is accessible to unprivilegedusers in bare-metal OSes. By using our generic learning framework trainedagainst CPU frequency traces, DeepTheft has shown similarly high attackperformance in stealing model architectures.

DeepTheft: Stealing DNN Model Architectures through Power Side Channel