STPFormer: A State-of-the-Art Pattern-Aware Spatio-Temporal Transformer for Traffic Forecasting

2025-08-18

AI Summary

Traffic spatiotemporal forecasting faces challenges including complex temporal patterns, dynamic spatial dependencies, and heterogeneous multi-source inputs. To address these, we propose the Pattern-Aware Spatiotemporal Transformer (MAST), which employs a Time Position Aggregator (TPA) for adaptive temporal encoding and a Spatial Sequence Aggregator (SSA) to serialize spatial structures. Furthermore, MAST introduces a Spatiotemporal Graph Matching (STGM) mechanism and an Attention Mixer module to enable fine-grained cross-domain alignment and multi-scale feature fusion. The framework unifies spatiotemporal dynamics modeling while preserving interpretability. Extensive experiments on five real-world traffic datasets consistently achieve state-of-the-art performance. Ablation studies and visualization analyses validate the effectiveness of each component and demonstrate the model’s strong generalization capability.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Spatio-temporal traffic forecasting is challenging due to complex temporal patterns, dynamic spatial structures, and diverse input formats. Although Transformer-based models offer strong global modeling, they often struggle with rigid temporal encoding and weak space-time fusion. We propose STPFormer, a Spatio-Temporal Pattern-Aware Transformer that achieves state-of-the-art performance via unified and interpretable representation learning. It integrates four modules: Temporal Position Aggregator (TPA) for pattern-aware temporal encoding, Spatial Sequence Aggregator (SSA) for sequential spatial learning, Spatial-Temporal Graph Matching (STGM) for cross-domain alignment, and an Attention Mixer for multi-scale fusion. Experiments on five real-world datasets show that STPFormer consistently sets new SOTA results, with ablation and visualizations confirming its effectiveness and generalizability.

Problem

• Handles complex temporal patterns in traffic forecasting
• Addresses dynamic spatial structures in traffic data
• Integrates diverse input formats for unified modeling

Innovation

• Pattern-aware temporal encoding via TPA
• Sequential spatial learning with SSA
• Cross-domain alignment through STGM