"Just in Time"World Modeling Supports Human Planning and Reasoning

2026-01-20

AI Summary

This work proposes a “Just-in-Time” computational framework to address the challenge that humans face in efficiently constructing task-relevant, simplified representations during mental simulation in complex environments due to limited cognitive resources. By tightly integrating probabilistic mental simulation, visual search, and dynamic representation learning in an online manner, the model encodes only those objects pertinent to the current task, thereby constructing a minimal yet sufficient world model in real time. This approach achieves, for the first time, a close coupling between simulation-guided visual attention and on-the-fly representational updating. Evaluated on grid-world navigation and physical reasoning tasks, the method attains high prediction accuracy with remarkably low computational overhead, substantially outperforming existing models and offering a novel algorithmic account of human-like efficient reasoning.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Computer Vision: Visual Reasoning & Symbolic Representations

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Probabilistic mental simulation is thought to play a key role in human reasoning, planning, and prediction, yet the demands of simulation in complex environments exceed realistic human capacity limits. A theory with growing evidence is that people simulate using simplified representations of the environment that abstract away from irrelevant details, but it is unclear how people determine these simplifications efficiently. Here, we present a"Just-in-Time"framework for simulation-based reasoning that demonstrates how such representations can be constructed online with minimal added computation. The model uses a tight interleaving of simulation, visual search, and representation modification, with the current simulation guiding where to look and visual search flagging objects that should be encoded for subsequent simulation. Despite only ever encoding a small subset of objects, the model makes high-utility predictions. We find strong empirical support for this account over alternative models in a grid-world planning task and a physical reasoning task across a range of behavioral measures. Together, these results offer a concrete algorithmic account of how people construct reduced representations to support efficient mental simulation.

Problem

• mental simulation
• simplified representations
• human reasoning
• planning
• cognitive efficiency

Innovation

• Just-in-Time modeling
• probabilistic mental simulation
• simplified representations
• visual search
• online representation construction