World Model

A world model is an internal representation that an AI system builds and uses to simulate how its environment works — capturing the rules, dynamics, and causal relationships that govern how states change in response to actions. Rather than relying solely on direct interaction with the real world, an agent equipped with a world model can mentally "roll out" hypothetical sequences of events, predicting what would happen if it took a given action before committing to it. This makes world models a cornerstone of model-based reinforcement learning, where the learned model of the environment is used to plan, generate synthetic training data, or both.

In practice, world models are typically learned from experience — the agent observes transitions in its environment and trains a neural network or other function approximator to predict future observations, rewards, or latent states. Architectures like recurrent neural networks, variational autoencoders, and transformer-based sequence models have all been used to build world models that compress high-dimensional sensory input into compact, predictive representations. Systems like Dreamer and Ha & Schmidhuber's "World Models" paper demonstrated that agents could learn entirely inside their own imagination, training a policy against the world model rather than the real environment.

World models matter because real-world interaction is often expensive, slow, or dangerous. A robot learning to manipulate objects, an autonomous vehicle navigating traffic, or a game-playing agent exploring a complex environment all benefit from being able to simulate consequences internally before acting. This also connects world models to broader ideas in AI safety and alignment: an agent that accurately models its environment is better positioned to reason about long-term consequences and avoid unintended side effects.

Beyond reinforcement learning, world models have become a unifying concept in discussions about general intelligence. Researchers argue that robust generalization — the ability to adapt to novel situations — requires something like a world model rather than purely reactive, stimulus-response behavior. This framing connects machine learning to cognitive science, where internal models of the environment are considered fundamental to how biological agents plan, imagine, and reason about counterfactuals.