CYBERSHIELD: A Competitive Simulation Environment for Training AI in Cybersecurity

Simulation environments are crucial in cybersecurity for training AI algorithms, providing a secure and controlled space to explore the impact derived from diverse cyber threads. They offer a flexible platform for accelerated learning and development, especially useful to train Decision Support Systems in response contexts. This paper introduces CyberShield, an advanced simulation environment tailored for cybersecurity tasks, specifically focusing on supporting the implementation and evaluation of AI algorithms, with an emphasis on RL. CyberShield encompasses a comprehensive environment with multiple computers, each hosting various services with unique vulnerabilities. Within this environment, two opposing agents, defender and attacker, participate in a strategic battle, each equipped with distinct actions aimed at outsmarting the other. CyberShield is optimized for competitive multi-agent training using RL algorithms. The specifically adapted environment facilitates the exploration of novel strategies, thereby fostering the automation of defense and attack strategies. Taking inspiration from Capture The Flag (CTF) competitions, CyberShield offers a dynamic and realistic platform for refining advanced cybersecurity techniques, fostering innovation, and strengthening the resilience of AI algorithms against sophisticated cyber threats.