EU Regulatory Changes

This paper, published on arXiv, introduces a new benchmark called Adaptive Adversaries designed to test the security of large language model (LLM) agents. Unlike previous single-turn tests, this be...

This paper, published on arXiv, proposes a new technical framework for generating synthetic data that is specifically designed to preserve privacy while maintaining domain-specific utility. It intr...

A new research paper published on arXiv on July 20, 2026, titled "Self-State Attacks on Self-Hosted AI Agents: How Far Can OS Defenses Go?" examines vulnerabilities in self-hosted AI agents where a...

This publication introduces a new technical framework, RT-SHCUA, which enables real-time, self-hosted control of unmanned aerial vehicles (UAVs) through an artificial intelligence agent. The system...

This publication introduces a novel theoretical framework for defending against side-channel attacks on encrypted network traffic, using rate-distortion theory to balance privacy protection with da...

This paper, published on arXiv, presents a new method for detecting cyberattacks on encrypted OPC UA traffic, a common industrial communication protocol. It demonstrates that even when traffic is e...

This paper, published on arXiv on July 20, 2026, presents a technical analysis of vulnerabilities in Vision-Language-Action (VLA) models, which are AI systems that process visual and language input...

A new preprint from arXiv, titled "Insecure Coding Preferences in Long-Term Memory: Security Risks for LLM-based Code Generation," published on 20 July 2026, presents evidence that large language m...

This publication from arXiv presents a new technique for obfuscating floating-point computations within deep neural network (DNN) binaries using Mixed Boolean-Arithmetic (MBA) obfuscation. The pape...