world models

In artificial intelligence and cognitive science, **world models** refer to internal representations or learned simulations that describe how the world works. These models allow an agent (human or machine) to predict outcomes of actions, reason about cause and effect, plan future behaviors, and generalize knowledge to new situations. In AI, world models can be explicit (e.g., physics-based simulations) or learned implicitly from data (e.g., neural networks trained to forecast observations and rewards).