Latent State Design for World Models under Sufficiency Constraints
For researchers in world models and reinforcement learning, this provides a new lens to compare and design models based on task-specific sufficiency, but the contribution is primarily taxonomic and evaluative rather than empirical.
This paper proposes a functional taxonomy for world models based on what their latent state is designed for, rather than architecture, and introduces an evaluation framework that judges models by the sufficiency constraints their latent states satisfy. The conclusion is that an actionable world model matches its state construction to the task, not maximizing information preservation.
A world model matters to an agent only through the state it constructs. That state must preserve some information, discard other information, and support some future function: prediction, control, planning, memory, grounding, or counterfactual reasoning. This paper treats world-model research as latent state design under sufficiency constraints. We propose a functional taxonomy that groups methods by what their latent state is for, rather than by architecture or application domain: predictive embedding, recurrent belief state, object/causal structure, latent action interface, grounded planning interface, and memory substrate. These roles expose distinctions that architecture-based groupings hide, including the gap between predictive sufficiency and control sufficiency, and the gap between passive video prediction and counterfactual action modeling. The taxonomy supports an evaluation framework that judges a model by the sufficiency constraint its latent state was built to satisfy. We compare methods along seven axes: representation, prediction, planning, controllability, causal/counterfactual support, memory, and uncertainty. We use the resulting matrix as a diagnostic for what a latent state preserves, discards, and enables. The conclusion that follows is that an actionable world model is the one whose state construction matches the task, not the one that preserves the most information.