Transferable Human Mobility Network Reconstruction with neuroGravity

2026-04-26

AI Summary

This study addresses the scarcity of mobility survey data in underdeveloped regions by proposing neuroGravity, a model that integrates the physical priors of gravity models with deep learning to reconstruct human mobility flows using only urban facility and population distribution data. The framework supports cross-city transferability and reveals, for the first time, that spatial income segregation is a key determinant of transfer performance. Leveraging this insight, the authors develop a quantitative metric that predicts transfer effectiveness. The learned area representations serve as scalable proxy variables for socioeconomic conditions and livability. Built upon a physics-informed deep learning and transfer learning architecture, the approach generates high-quality surrogate mobility data for over 1,200 cities worldwide, substantially mitigating data gaps, with transfer performance strongly aligned with the similarity in income segregation levels between source and target cities.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: The sharing economy
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

Accurate modeling of human mobility is critical for tackling urban planning and public health challenges. In undeveloped regions, the absence of comprehensive travel surveys necessitates reconstructing mobility networks from publicly available data. Here we develop neuroGravity, a physics-informed deep learning model that reliably reconstructs mobility flows from limited observations and transfers to unobserved cities. Using only urban facility and population distributions, we find that neuroGravity's regional representations strongly correlate with socioeconomic and livability status, offering scalable proxies for costly surveys. Furthermore, we uncover that spatial income segregation plays a key role in model transferability: mobility networks are most reliably reconstructed when target cities share similar segregation levels with the source. We design an index to quantify this segregation and accurately predict transferability. Finally, we generate mobility flow proxies for over 1,200 cities worldwide, highlighting neuroGravity's potential to mitigate critical data shortages in resource-limited, underdeveloped areas.

Problem

• human mobility
• mobility network reconstruction
• data scarcity
• undeveloped regions
• transferability

Innovation

• neuroGravity
• human mobility reconstruction
• physics-informed deep learning
• transferability
• spatial income segregation