What Must a World Model Distinguish for Planning?

2026-09-26

AI Summary

This study addresses the limitation of world models in planning, where they excessively retain physical details required for prediction while lacking a well-defined boundary of decision sufficiency. To overcome this, we construct a hierarchical framework encompassing mechanism, response, and decision sufficiency, revealing how planning requirements depend on query characteristics. Furthermore, we propose a modular architecture integrating query localization with action-conditioned prediction to enable prediction reuse, validated through collision simulation and nonlinear dynamics experiments. The proposed architecture significantly outperforms joint generative models by effectively reducing regret on known targets and substantially improving prediction reuse efficiency when generalizing to unseen targets.

Abstract

World models simulate the consequences of action candidates, but good planning need not preserve every physical distinction required for accurate prediction. We formalize this gap through a hierarchy of mechanism, response, and decision sufficiency. Given a candidate set, the planning query determines which physical variations matter and how precisely they must be preserved: coarse decisions can discard much of the information needed for prediction, whereas fine decisions may require nearly the same resolution. In practice, planners often adaptively search to construct candidates, and information unnecessary for final selection may still be needed to discover good candidates. What a world model must preserve therefore depends on the query, the candidate set, and the planner. We study these effects in a collision system, nonlinear dynamics, and robotic planning. These varying requirements raise a design question: where should query information enter the planning system? A model that jointly generates actions and outcomes conditioned on the query achieves lower regret than an action-conditioned world model on seen objectives, but this advantage largely disappears when generalizing to unseen objectives. Motivated by this, we propose a modular design in which the query determines where to look and an action-conditioned model predicts what will happen, allowing the same predictions to be reused across objectives.

Problem

• world model
• planning
• decision sufficiency
• modular design
• generalization

Innovation

• World Models
• Decision Sufficiency
• Modular Design
• Robotic Planning
• Generalization