Large Language Model-Based Reward Design for Deep Reinforcement Learning-Driven Autonomous Cyber Defense
This work addresses the problem of reward design for autonomous cyber defense, offering a novel method that could reduce reliance on subject matter experts, though it appears incremental as it applies existing LLM and DRL techniques to a specific domain.
The paper tackled the challenge of designing rewards for autonomous cyber defense agents by proposing an LLM-based reward design approach, which generated effective defense strategies against diverse adversarial behaviors in a DRL-driven simulation environment.
Designing rewards for autonomous cyber attack and defense learning agents in a complex, dynamic environment is a challenging task for subject matter experts. We propose a large language model (LLM)-based reward design approach to generate autonomous cyber defense policies in a deep reinforcement learning (DRL)-driven experimental simulation environment. Multiple attack and defense agent personas were crafted, reflecting heterogeneity in agent actions, to generate LLM-guided reward designs where the LLM was first provided with contextual cyber simulation environment information. These reward structures were then utilized within a DRL-driven attack-defense simulation environment to learn an ensemble of cyber defense policies. Our results suggest that LLM-guided reward designs can lead to effective defense strategies against diverse adversarial behaviors.