General Agents Contain World Models, even under Partial Observability and Stochasticity
This work addresses the fundamental problem of understanding agent capabilities in AI, though it is incremental as it builds on existing theoretical frameworks.
The paper extends a prior theorem to show that stochastic agents in partially observable environments still contain sufficient knowledge to reconstruct their environment, proving that randomization does not prevent learning world models.
Deciding whether an agent possesses a model of its surrounding world is a fundamental step toward understanding its capabilities and limitations. In [10], it was shown that, within a particular framework, every almost optimal and general agent necessarily contains sufficient knowledge of its environment to allow an approximate reconstruction of it by querying the agent as a black box. This result relied on the assumptions that the agent is deterministic and that the environment is fully observable. In this work, we remove both assumptions by extending the theorem to stochastic agents operating in partially observable environments. Fundamentally, this shows that stochastic agents cannot avoid learning their environment through the usage of randomization. We also strengthen the result by weakening the notion of generality, proving that less powerful agents already contain a model of the world in which they operate.