TEAM: Topological Evolution-aware Framework for Traffic Forecasting--Extended Version

TL;DR

The paper tackles long-term traffic forecasting on evolving road networks, where topology changes invalidate static models and trigger costly retraining. It introduces CAST, a Convolution Attention for Spatio-Temporal model, and TEAM, a continual learning framework that uses Wasserstein-based buffers to identify stable and changing nodes, enabling efficient incremental training on only evolved parts. Empirical results on PEMS03-Evolve and PEMS04-Evolve show CAST achieves strong accuracy, while TEAM delivers substantial runtime gains with competitive accuracy, outperforming many baselines in evolving-topology scenarios. The approach reduces computational burden without sacrificing forecast quality, offering a practical solution for scalable traffic forecasting in rapidly urbanizing environments.

Abstract

Due to the global trend towards urbanization, people increasingly move to and live in cities that then continue to grow. Traffic forecasting plays an important role in the intelligent transportation systems of cities as well as in spatio-temporal data mining. State-of-the-art forecasting is achieved by deep-learning approaches due to their ability to contend with complex spatio-temporal dynamics. However, existing methods assume the input is fixed-topology road networks and static traffic time series. These assumptions fail to align with urbanization, where time series are collected continuously and road networks evolve over time. In such settings, deep-learning models require frequent re-initialization and re-training, imposing high computational costs. To enable much more efficient training without jeopardizing model accuracy, we propose the Topological Evolution-aware Framework (TEAM) for traffic forecasting that incorporates convolution and attention. This combination of mechanisms enables better adaptation to newly collected time series, while being able to maintain learned knowledge from old time series. TEAM features a continual learning module based on the Wasserstein metric that acts as a buffer that can identify the most stable and the most changing network nodes. Then, only data related to stable nodes is employed for re-training when consolidating a model. Further, only data of new nodes and their adjacent nodes as well as data pertaining to changing nodes are used to re-train the model. Empirical studies with two real-world traffic datasets offer evidence that TEAM is capable of much lower re-training costs than existing methods are, without jeopardizing forecasting accuracy.

Figures (15)

• Figure 1: Example of a city and its corresponding evolving RN.
• Figure 2: Evolution from $G^{\pi-1}$ to $G^{\pi}$. The blue oval regions highlight the evolved parts. Red nodes and edges denote added nodes and edges, respectively. Blue nodes and edges denote removed nodes and edges, respectively.
• Figure 3: TEAM framework overview.
• Figure 4: Main model (CAST).
• Figure 5: Temporal convolution.