Title: DRAN: A Distribution and Relation Adaptive Network for Spatio-temporal Forecasting

Original: arXiv:2504.01531v4 Announce Type: replace Abstract: Accurate predictions of spatio-temporal systems are crucial for tasks such as system management, control, and crisis prevention. However, the inherent time variance of many spatio-temporal systems poses challenges to achieving accurate predictions whenever stationarity is not granted. In order to address non-stationarity, we propose a Distribution and Relation Adaptive Network (DRAN) capable of dynamically adapting to relation and distribution changes over time. While temporal normalization and de-normalization are frequently used techniques to adapt to distribution shifts, this operation is not suitable for the spatio-temporal context as temporal normalization scales the time series of nodes and possibly disrupts the spatial relations among nodes. In order to address this problem, a Spatial Factor Learner (SFL) module is developed that enables the normalization and de-normalization process. To adapt to dynamic changes in spatial relationships among sensors, we propose a Dynamic-Static Fusion Learner (DSFL) module that effectively integrates features learned from both dynamic and static relations through an adaptive fusion ratio mechanism. Furthermore, we introduce a Stochastic Learner to capture the noisy components of spatio-temporal representations. Our approach outperforms state-of-the-art methods on weather prediction and traffic flow forecasting tasks.Experimental results show that our SFL efficiently preserves spatial relationships across various temporal normalization operations. Visualizations of the learned dynamic and static relations demonstrate that DSFL can capture both local and distant relationships between nodes.

Rewritten: Title: DRAN: A Distribution and Relation Adaptive Network for Spatio-temporal Forecasting

Original: arXiv:2504.01531v4 Announce Type: replace Abstract: Precise forecasting of spatio-temporal systems is essential for applications like crisis prevention, control, and system management. Nevertheless, the intrinsic time-variance inherent in numerous spatio-temporal systems makes accurate prediction difficult, particularly when stationarity cannot be assumed. To tackle the issue of non-stationarity, we introduce the Distribution and Relation Adaptive Network (DRAN), which dynamically adjusts to evolving distributions and relationships over time. Although temporal normalization and de-normalization are common methods for handling distribution shifts, they are ill-suited for spatio-temporal contexts because scaling node time series can inadvertently alter the spatial connections between nodes. To resolve this, we developed a Spatial Factor Learner (SFL) module that facilitates the normalization and de-normalization process while maintaining integrity. Additionally, to accommodate shifting spatial dynamics among sensors, we present a Dynamic-Static Fusion Learner (DSFL) module. This component seamlessly merges features derived from both static and dynamic relations using an adaptive fusion ratio. We also incorporate a Stochastic Learner to extract noisy elements within spatio-temporal representations. Our method surpasses existing state-of-the-art techniques in traffic flow forecasting and weather prediction. Our findings indicate that the SFL successfully maintains spatial relationships during various temporal normalization procedures. Moreover, visualizations of the learned dynamic and static relations reveal that DSFL is capable of identifying both nearby and distant node interactions.

Source: arXiv Generated at: 2026-06-03 00:00:00 UTC