Cross-Device Profiled Side-Channel Attacks using Meta-Transfer Learning
Hosted in Virtual Platform
Hardware Security: Attack and Defense
DescriptionDeep learning (DL) based profiling side-channel analysis (SCA) poses a great threat to embedded devices. An adversary can break the target encryption engine through physical leakages collected from a profiling device. In this paper, we propose a more efficient attack mechanism that adopts meta-transfer learning to transfer DL networks among target devices by judiciously extracting information from a profiling device even using different side-channel sources (e.g., power or EM). In comparison to previous attack methodologies, we significantly reduce training costs and the number of traces required for side-channel attacks on both unprotected or masked AES implementations.