World Models for Cognitive Agents: Transforming Edge Intelligence in Future Networks

The paper provides a comprehensive overview of world models in AI, emphasizing their role in enabling agents to build internal environment representations for predictive reasoning, planning, and decision-making, with particular benefits in data-limited or safety-critical contexts. It also introduces Wireless Dreamer, a novel reinforcement learning framework based on world models designed to optimize wireless edge networks, demonstrated through a weather-aware UAV trajectory planning case study.