量化地理域偏移以解耦人类移动流生成模型的空间可迁移性
文章背景与核心概要
本研究探讨了人类移动生成模型的空间可迁移性（geospatial transferability），即评估在一个区域训练的模型应用于未见过的全新位置时的性能表现。该研究利用涵盖美国 2,265 个县的人口普查区块层级通勤流的大规模基准数据集以及四种代表性生成模型，引入了受机器学习领域自适应理论启发的地理域偏移（geographic domain shift）概念。作者提出了两种定量指标——互信息（mutual information）和空间偏移（spatial shift），用于衡量源区域与目标区域之间地理特征分布和空间结构的差异。通过线性混合效应回归，研究结果表明，空间可迁移性主要受内在地理差异的显著影响，而非仅取决于模型架构，这为鲁棒且公平的人类移动合成提供了一个新颖的框架。
Summary
This research paper investigates the geospatial transferability of human mobility generation models—evaluating how well models trained in one region perform when applied to unseen or new locations. Using a benchmark dataset of census tract-level commuting flows across 2,265 U.S. counties and four representative generation models, the study introduces the concept of geographic domain shift (inspired by machine learning domain adaptation theory). The authors propose two quantitative metrics—mutual information and spatial shift—to measure differences in geographic feature distributions and spatial structures between source and target regions. Utilizing linear mixed-effects regression, the findings reveal that geospatial transferability is significantly influenced by intrinsic geographic differences rather than model architecture alone, offering a novel framework for robust and fair human mobility synthesis.
文章元数据 (Article Metadata)
| 属性 (Attribute) | 详情 (Details) |
|---|---|
| arXiv ID | arXiv:2608.21567 [cs.AI] |
| 学科领域 (Subjects) | 人工智能 (cs.AI) |
| ACM 类别 (ACM Classes) | I.2 |
| 作者 (Authors) | Zhiyong Zhou, Song Gao, Qianheng Zhang, Feng Zhang, Zhenhong Du |
| 提交时间 (Submitted) | 2026年8月21日 |
| 期刊引用 (Journal Reference) | Annals of the American Association of Geographers, 2026 |
| 许可协议 (License) | 知识共享署名 4.0 |
Article Metadata
Attribute Details arXiv ID arXiv:2608.21567 [cs.AI] Subjects Artificial Intelligence ( cs.AI)ACM Classes I.2 Authors Zhiyong Zhou, Song Gao, Qianheng Zhang, Feng Zhang, Zhenhong Du Submitted August 21, 2026 Journal Reference Annals of the American Association of Geographers, 2026 License Creative Commons Attribution 4.0
摘要 (Abstract)
人类移动是理解城市系统中社会、经济和环境动态的重要代理指标。空间可迁移性衡量模型在全新位置或未见区域中的能力，是比较不同人类移动生成模型的一个关键维度。然而，很少有研究探讨空间可迁移性的内在特征。
为此，本研究利用包含美国 2,265 个县人口普查区块层级通勤流的大规模基准数据集，系统地研究了四种代表性人类移动生成模型的空间可迁移性。受机器学习中域自适应理论的启发，我们引入了地理域偏移来描述源区域与目标区域之间在地理特征分布和空间结构上的内在差异，这些差异可能会共同影响模型的迁移性能。
此外，我们提出了两种指标——互信息和空间偏移——来量化地理域偏移。为了检验它们与模型可迁移性的关联，我们采用线性混合效应回归来分析地理域偏移与可迁移性之间的关联。
我们的研究结果表明，跨区域的迁移性能表现出显著的空间异质性和不对称性。信息偏移和空间偏移均表现出统计学上显著且互补的解释力。这表明空间可迁移性不仅取决于模型设计，还取决于内在的地理差异。这些发现为评估和改进人类移动生成模型的空间可迁移性提供了一个新颖的方法论框架，并支持在不同区域之间进行更具鲁棒性和公平性的人类移动数据合成。它同时也为 GeoAI 模型开发中的空间可迁移性提供了深刻见解。
Human mobility serves as an essential proxy for understanding social, economic, and environmental dynamics in urban systems. Geospatial transferability, which measures a model's capability in a new location or unseen region, is a critical dimension for comparing different human mobility generation models. However, few studies have studied the intrinsic characteristics of geospatial transferability.
To this end, this study systematically investigates the geospatial transferability of four representative human mobility generation models using a large-scale benchmark dataset of census tract level commuting flows across 2,265 counties in the United States. Inspired by the domain adaptation theory in machine learning, we introduce geographic domain shift to describe the intrinsic differences in geographic feature distributions and spatial structures between source and target regions, which may jointly affect model transferability.
Moreover, we propose two metrics—mutual information and spatial shift—to quantify the geographic domain shift. To examine their associations with model transferability, we employ linear mixed-effects regression to analyze the associations between geographic domain shifts and transferability.
Our results reveal substantial spatial heterogeneity and asymmetry in transfer performance across regions. Both information shift and spatial shift exhibit statistically significant and complementary explanatory power. This indicates that geospatial transferability depends not only on model design but also on intrinsic geographic differences. These findings provide a novel methodological framework for evaluating and improving the geospatial transferability of human mobility generation models and support more robust and fair human mobility data synthesis across diverse regions. It also offers insights on spatial transferability for GeoAI model development.
访问与资源 (Access & Resources)
- 全文选项： 查看 PDF | HTML 版本 | TeX 源码
- 外部引用： Google Scholar | Semantic Scholar | NASA ADS
Access & Resources
- Full-Text Options: View PDF | HTML Version | TeX Source
- External Citations: Google Scholar | Semantic Scholar | NASA ADS
许可声明 (License Notice)
License Notice