DeepTheft: Stealing DNN Model Architectures Through Power Side Channel.
45TH IEEE SYMPOSIUM ON SECURITY AND PRIVACY, SP 2024（2024）
Key words
Deep Neural Network,Model Architecture,Deep Neural Network Model,Side-channel,Power Side-channel,Sampling Rate,Hyperparameters,Deep Models,Learning Framework,Low Sampling Rate,Levenshtein Distance,CPU Frequency,Attack Performance,Training Dataset,Convolutional Layers,Feature Maps,Spatial Features,Long Short-term Memory,Cross-entropy Loss,Temporal Features,Bidirectional Long Short-term Memory,Linear Layer,Online Phase,Model Inference,Time-series Signals,Direct Regression,Number Of Kernels,Input Size,Unknown Model,Deep Neural Network Architecture
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined