World Models in Pieces: Structural Certification for General Agents

2026-06-23

AI Summary

This work addresses the challenge that general-purpose agents often lack universally robust capabilities in complex environments, while traditional worst-case analyses fail to distinguish critical bottlenecks from incidental failures. To overcome this limitation, the paper introduces a structured certification framework that reduces goal-conditioned performance guarantees to verifiable, component-wise assurances on the agent’s internal world model through a localized transition-centric perspective. By integrating a deep compositional, goal-guided transition filtering algorithm with formal verification techniques, the authors theoretically establish the inherent non-universality of general agents and construct a verifiable world model with a tight error bound of $\mathcal{O}(1/n) + \mathcal{O}(\delta)$. The tightness of this bound under small $\delta$ is rigorously proven, thereby enabling localized identification and quantitative assurance of reliability in long-horizon planning.

Abstract

In the big-world regime, agents cannot be universally capable and their ability is inevitably specialized across a world model in pieces. Consequently, standard uniform guarantees fail to distinguish between the understanding of critical bottlenecks and irrelevant failures. We first formalize this limitation by proving that general agents are not universal, rendering standard worst-case analysis uninformative. To overcome this, we introduce structural certification, a transition-local framework that maps bounded goal-conditioned performance to entry-wise guarantees on the agent's internal world model. Our main contribution is constructive. We provide algorithms that filter specific transitions using deep compositional goals and prove that a general agent on these goals has a structural world model with a $\mathcal{O}(1/n) + \mathcal{O}(δ)$ error bound. Conversely, this bound is tight in the small-$δ$ regime, whose existence is explicitly guaranteed by our certification. These results enable the certifiable deployment of general agents by localizing the specific transitions where long-horizon planning is reliable.

Problem

• World Models
• General Agents
• Structural Certification
• Bounded Rationality
• Goal-Conditioned Performance

Innovation

• structural certification
• world models
• general agents
• compositional goals
• error bounds