NetDeTox: Evasion of Hardware-Security GNNs with RL-LLM Orchestration
Published:Nov 27, 2025 20:45
•1 min read
•ArXiv
Analysis
This research explores a novel method for evading hardware-security Graph Neural Networks (GNNs) using a Reinforcement Learning (RL) - Large Language Model (LLM) orchestration. The approach could have significant implications for cybersecurity and hardware design.
Key Takeaways
- •Proposes a new adversarial attack against hardware-security GNNs.
- •Utilizes RL-LLM orchestration for attack generation.
- •Focuses on efficient evasion techniques.
Reference
“NetDeTox leverages RL-LLM orchestration for adversarial evasion.”