Qi Li

Securing autonomous AI agents demands a lifecycle-oriented approach, and AgentWard provides a blueprint for defense-in-depth across initialization, input processing, memory, decision-making, and execution.

VLA models introduce a fundamentally new risk landscape compared to LLMs or robotics alone, demanding a unified safety perspective that considers irreversible physical consequences and multimodal attack surfaces.

By progressively expanding boundary predictions, PBE-UNet effectively focuses on challenging segmentation error regions in ultrasound images, leading to state-of-the-art performance.

Autonomous LLM agents are riddled with vulnerabilities, as point defenses fail to address cross-temporal and multi-stage systemic risks like memory poisoning and intent drift.