Hung Dang

CHRONOS is a hardware-assisted framework that significantly reduces the latency of secure federated learning by decoupling cryptographic key setup from the active training phase, while maintaining high security against various attacks.

The paper proposes extbackslash codeName, a behavioral firewall that uses a parameterized deterministic finite automaton (pDFA) to enforce verified benign tool-call sequences and parameter bounds for structured-workflow AI agents, significantly reducing the attack surface.

The paper proposes extit{codename}, an architecture that enforces verifiable workflows across untrusted networks by combining hardware-isolated control and kernel-resident data planes, achieving low-latency encrypted routing.