Understanding the link between urban planning and commuting flows is crucial
for guiding urban development and policymaking. This research, bridging
computer science and urban studies, addresses the challenge of integrating
these fields with their distinct focuses. Traditional urban studies methods,
like the gravity and radiation models, often underperform in complex scenarios
due to their limited handling of multiple variables and reliance on overly
simplistic and unrealistic assumptions, such as spatial isotropy. While deep
learning models offer improved accuracy, their black-box nature poses a
trade-off between performance and explainability -- both vital for analyzing
complex societal phenomena like commuting flows. To address this, we introduce
TransFlower, an explainable, transformer-based model employing flow-to-flow
attention to predict urban commuting patterns. It features a geospatial encoder
with an anisotropy-aware relative location encoder for nuanced flow
representation. Following this, the transformer-based flow predictor enhances
this by leveraging attention mechanisms to efficiently capture flow
interactions. Our model outperforms existing methods by up to 30.8% Common Part
of Commuters, offering insights into mobility dynamics crucial for urban
planning and policy decisions.

通过引入 TransFlower，一种可解释的、基于 Transformer 的模型，使用流量 - 流量注意力来预测城市通勤模式，我们的研究在运动力学方面取得了 30.8% 的改进，为城市规划和政策决策提供了重要的移动动态洞察。

逐流注意力的可解释 Transformer 模型用于通勤流预测

TransFlower: An Explainable Transformer-Based Model with Flow-to-Flow  Attention for Commuting Flow Prediction

Among the realistic ingredients to be considered in the computational
modeling of infectious diseases, human mobility represents a crucial challenge
both on the theoretical side and in view of the limited availability of
empirical data. In order to study the interplay between small-scale commuting
flows and long-range airline traffic in shaping the spatio-temporal pattern of
a global epidemic we i) analyze mobility data from 29 countries around the
world and find a gravity model able to provide a global description of
commuting patterns up to 300 kms; ii) integrate in a worldwide structured
metapopulation epidemic model a time-scale separation technique for evaluating
the force of infection due to multiscale mobility processes in the disease
dynamics. Commuting flows are found, on average, to be one order of magnitude
larger than airline flows. However, their introduction into the worldwide model
shows that the large scale pattern of the simulated epidemic exhibits only
small variations with respect to the baseline case where only airline traffic
is considered. The presence of short range mobility increases however the
synchronization of subpopulations in close proximity and affects the epidemic
behavior at the periphery of the airline transportation infrastructure. The
present approach outlines the possibility for the definition of layered
computational approaches where different modeling assumptions and granularities
can be used consistently in a unifying multi-scale framework.

本文研究人类移动性在感染病毒建模中的重要性，并通过分析 29 个国家的移动数据和构建重力模型，以及将短程通勤流纳入全球疫情结构化流行病模型中，证明通勤流的引入会影响到疫情在近距离群体中的传播，但对全球模式的影响相对较小。