Do We Really Need Transformers for Global Spatial Information Extraction in Traffic Forecasting?

arXiv cs.AI Papers

Summary

This paper investigates whether complex transformer-based attention is necessary for global spatial information extraction in traffic forecasting, finding that simple global aggregation operators achieve comparable performance with lower computational complexity.

arXiv:2607.12462v1 Announce Type: new Abstract: Existing traffic forecasting models commonly focus on extracting spatial dependencies, particularly global spatial information, which characterizes the representations obtained through interactions between each individual node and all nodes across the traffic network. However, the underlying mechanism by which such global information is modeled and extracted remains insufficiently investigated. Whether global information must be extracted by high-degree-of-freedom adaptive attention or can be captured by a simple global aggregation operator remains unclear. For this purpose, we design a controlled ablation framework that replaces only the spatial mixing module to test attention-based global interaction. Across six traffic benchmarks, uniform full-range mixing and standard spatial attention each achieve lower MAE on three datasets, with only a 0.14% difference in mean MAE, while the former reduces node-scale spatial mixing complexity from O(N2) to O(N). Mechanism analysis further decomposes spatial attention into a row-uniform global background and a non-uniform residual. The residual shows dataset-dependent marginal value, suggesting that spatial attention should be justified by stable gains beyond a row-uniform global background. The corresponding source code is publicly available at: https://github.com/uuesti/U-Trans
Original Article
View Cached Full Text

Cached at: 07/15/26, 04:20 AM

# Do We Really Need Transformers for Global Spatial Information Extraction in Traffic Forecasting?
Source: [https://arxiv.org/abs/2607.12462](https://arxiv.org/abs/2607.12462)
[View PDF](https://arxiv.org/pdf/2607.12462)

> Abstract:Existing traffic forecasting models commonly focus on extracting spatial dependencies, particularly global spatial information, which characterizes the representations obtained through interactions between each individual node and all nodes across the traffic network\. However, the underlying mechanism by which such global information is modeled and extracted remains insufficiently investigated\. Whether global information must be extracted by high\-degree\-of\-freedom adaptive attention or can be captured by a simple global aggregation operator remains unclear\. For this purpose, we design a controlled ablation framework that replaces only the spatial mixing module to test attention\-based global interaction\. Across six traffic benchmarks, uniform full\-range mixing and standard spatial attention each achieve lower MAE on three datasets, with only a 0\.14% difference in mean MAE, while the former reduces node\-scale spatial mixing complexity from O\(N2\) to O\(N\)\. Mechanism analysis further decomposes spatial attention into a row\-uniform global background and a non\-uniform residual\. The residual shows dataset\-dependent marginal value, suggesting that spatial attention should be justified by stable gains beyond a row\-uniform global background\. The corresponding source code is publicly available at:[this https URL](https://github.com/uuesti/U-Trans)

## Submission history

From: Siyao Zhang \[[view email](https://arxiv.org/show-email/b754a013/2607.12462)\] **\[v1\]**Tue, 14 Jul 2026 07:44:01 UTC \(813 KB\)

Similar Articles

A Global-Local Graph Attention Network for Traffic Forecasting

arXiv cs.AI

Proposes a Global-Local Graph Attention Network (GLGAT) with pairwise encoding and event-based adjacency matrix for traffic forecasting, effectively capturing spatio-temporal correlations and achieving competitive performance on real-world datasets.