A Survey of Large Language Model-Powered Spatial Intelligence Across Scales: Advances in Embodied Agents, Smart Cities, and Earth Science

2025-04-14

AI Summary

This study investigates disparities and interdependencies in spatial intelligence exhibited by large language models (LLMs) across three distinct scales: embodied, urban, and Earth-system. Addressing the lack of unified cross-scale modeling in prior work, we propose the first interdisciplinary, multi-scale framework for spatial intelligence evolution—integrating human spatial cognition theory with LLM capability assessment. Methodologically, we unify cognitive science principles, multimodal representation learning, spatial knowledge graphs, and vision–language–geospatial co-modeling to systematically characterize the progression from spatial memory and comprehension to reasoning and decision-making. Our results identify scale-specific bottlenecks, uncover patterns of capability transfer across scales, and establish a theoretical foundation—alongside a reusable research paradigm—for advancing LLM-driven spatial decision support. (132 words)

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Natural Language Processing: (Large) Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Over the past year, the development of large language models (LLMs) has brought spatial intelligence into focus, with much attention on vision-based embodied intelligence. However, spatial intelligence spans a broader range of disciplines and scales, from navigation and urban planning to remote sensing and earth science. What are the differences and connections between spatial intelligence across these fields? In this paper, we first review human spatial cognition and its implications for spatial intelligence in LLMs. We then examine spatial memory, knowledge representations, and abstract reasoning in LLMs, highlighting their roles and connections. Finally, we analyze spatial intelligence across scales -- from embodied to urban and global levels -- following a framework that progresses from spatial memory and understanding to spatial reasoning and intelligence. Through this survey, we aim to provide insights into interdisciplinary spatial intelligence research and inspire future studies.

Problem

• Exploring spatial intelligence differences across disciplines and scales
• Reviewing human spatial cognition implications for LLMs
• Analyzing spatial memory, knowledge, and reasoning in LLMs

Innovation

• LLMs enhance spatial memory and knowledge representation
• Spatial reasoning across embodied to global scales
• Interdisciplinary framework for spatial intelligence research