Constraining the capacity of physical side channels for AI verification and security

Researchers explore methods to limit information leakage through physical side channels in AI systems, aiming to enhance verification and security against unauthorized data extraction.