This laboratory simulates a tactical cyber deception scenario informed by Evolutionary Game Theory (EGT) and Foundation Models (FMs).
The Mechanics
- Defender (You): Must allocate a limited number of Honeypots across the network.
- Attacker (AI): Uses reinforcement learning to probe the network. It updates its strategy based on previous encounters.
- Foundation Model: Acts as your strategic advisor, providing predictive analytics on attacker behavior to help you achieve decision dominance.
Game Theory Context
In strategic cyber warfare, interactions are often modeled as dynamic games. The payoff matrix determines the incentives. If the attacker breaches a real server, they gain utility (+10) and you lose (-10). If they hit a honeypot, the deception succeeds, flipping the utilities.
The FM predicts whether the attacker will strike central nodes (high value, high risk) or perimeter nodes (lower value, often less defended) based on their evolving policy.