P5_21: Towards Robust Cross-Device Side-Channel Attacks
Topic Areas: Side-Channel Attacks, Machine Learning
Principal investigator: Dr. Boyang Wang, University of Cincinnati
Co-Principal investigator: Dr. J. M. Emmert, University of Cincinnati
This research proposes a new deep-learning side-channel attack to improve the robustness and accuracy in the cross-device setting. The proposed attack will be able to recover secret keys from devices, such as microcontrollers, by analyzing power traces. In the cross-device setting, training power traces are from one device while test power traces are from a different device. It is a more realistic attack setting than the single-device setting investigated in previous research. To overcome the intra-variance of power traces across different devices and the limited number of power traces from a test device, this research proposes to design two methods by leveraging transfer learning. The team will collect power traces on two 8-bit XMEGA microcontrollers and two 32-bit STM32F3 microcontrollers and build deep neural networks to evaluate the performance of the proposed attack.