Model improves urban traffic forecasting by separating scales and trends
$M^2Net$: Unifying Scale-Specific Structures and Prototype Decomposition for Spatial-Temporal Forecasting
Artificial Intelligence
Summary
Forecasting how things change across time and space, like traffic patterns, is tricky because patterns shift depending on the time scale and because common trends get mixed up with unique local differences. The authors propose a model called M2Net that looks at different time scales separately and also separates out shared global patterns from individual location differences. This helps the model better understand and predict complex data by focusing on important signals and ignoring noise. Their tests on real urban data show their model works better than other methods.
What this means in practice
- •For urban planners: Predict changing traffic and resource demand patterns by distinguishing local disruptions from overall trends across multiple time scales.
- •For smart city system developers: Build smarter control systems that adapt to varying spatial and temporal dependencies in urban sensor data for improved decision-making.
Authors
Chengyang Zhou, Zijian Zhang, Hao Miao, Chunxu Zhang, Ziming Liu, Chao Tang, Juncheng Hu, Irwin King
Abstract
Spatial-temporal forecasting is pivotal for intelligent urban management yet remains challenging due to the intricate coupling of spatial dynamics and temporal non-stationarity. Existing approaches, while effective, often overlook two fundamental characteristics of real-world data: (1) Granularity-dependence, where dominant spatial correlations shift significantly across different time scales, and (2) Signal entanglement, where global common patterns are inextricably mixed with node-specific disturbances. In this paper, we propose a novel framework named $M^2Net$ (Multi-Scale Masked attention and Prototype network) to unify scale-specific structural learning and semantic decomposition. First, to capture the evolution of spatial dependencies, we introduce a Multi-Scale Masked attention mechanism. By downsampling temporal features into multiple resolutions, we dynamically generate scale-adaptive spatial masks, allowing the model to distinguish between micro-scale physical neighbors and macro-scale functional correlations. Second, to disentangle unique variations from shared trends, we propose a Prototype Residual Learning module. It retrieves global common patterns from a momentum-updated prototype bank and subtracts them from node representations, explicitly forcing the network to model the residual unique deviations. By integrating these two orthogonal views, i.e., structural multiscaling and semantic decomposition, our approach effectively isolates valid spatial-temporal signals from noise. Extensive experiments on real-world datasets demonstrate that $M^2Net$ outperforms state-of-the-art baselines on extensive empirical verifications.