Meng Chen

STEP introduces a novel, black-box, retraining-free detector that profiles audio samples using dual perturbation branches to detect backdoor attacks by exploiting the characteristic instability of hidden triggers.

The paper introduces AudioHijack, a framework that successfully demonstrates context-agnostic and imperceptible auditory prompt injection attacks, showing that commercial Large Audio-Language Models can be hijacked with high success rates.

The paper introduces DigitalCoach, a dataset of human expert-novice computer use coaching sessions, and evaluates the ability of state-of-the-art models to teach humans how to use computers.

This paper investigates the OS resilience to a new class of attacks called self-state attacks, where an agent's state is compromised via legitimate OS system call invocation. It characterizes the attack space, evaluates prevention, detection, and recovery strategies, and introduces a workload-conditioned view of detectability.