Governing Cyborg Propaganda in the Digital Public Square
Explore how cyborg propaganda blends human and AI influence, challenging democratic discourse and requiring new governance frameworks.
Security Threat Modeling for AI-Agent Protocols: MCP, A2A, Agora, ANP
Explore a detailed security threat modeling and risk assessment of emerging AI-agent protocols MCP, A2A, Agora, and ANP to ensure safe deployment.
Evaluating Agentic LLMs with Semantics-Preserving CTF Challenges
Explore a novel family-based method to evaluate agentic LLMs using semantics-preserving transformations for enhanced robustness in cybersecurity tasks.
CoMeT: Efficient Long-Context Transformer for LLMs
Discover CoMeT, the Collaborative Memory Transformer that enables efficient long-context modeling with constant memory and linear time complexity.
1S-DAug: One-Shot Data Augmentation for Few-Shot AI
Discover 1S-DAug, a one-shot data augmentation method boosting few-shot learning accuracy by up to 20% without model retraining.