STAMImputer: Spatio-Temporal Attention MoE for Traffic Data Imputation

TL;DR

STAMImputer introduces a spatio-temporal attention mixture of experts model that effectively imputes missing traffic data by capturing dynamic spatial-temporal correlations and addressing nonstationarity in traffic patterns.

Contribution

The paper proposes a novel MoE-based framework with a dynamic graph attention mechanism for improved traffic data imputation in block missing scenarios.

Findings

Significant performance improvements over state-of-the-art methods.

Effective modeling of real-time spatial correlations.

Robustness in nonstationary traffic data conditions.

Abstract

Traffic data imputation is fundamentally important to support various applications in intelligent transportation systems such as traffic flow prediction. However, existing time-to-space sequential methods often fail to effectively extract features in block-wise missing data scenarios. Meanwhile, the static graph structure for spatial feature propagation significantly constrains the models flexibility in handling the distribution shift issue for the nonstationary traffic data. To address these issues, this paper proposes a SpatioTemporal Attention Mixture of experts network named STAMImputer for traffic data imputation. Specifically, we introduce a Mixture of Experts (MoE) framework to capture latent spatio-temporal features and their influence weights, effectively imputing block missing. A novel Low-rank guided Sampling Graph ATtention (LrSGAT) mechanism is designed to dynamically balance the local and global correlations across road networks. The sampled attention vectors are utilized to generate dynamic graphs that capture real-time spatial correlations. Extensive experiments are conducted on four traffic datasets for evaluation. The result shows STAMImputer achieves significantly performance improvement compared with existing SOTA approaches. Our codes are available at https://github.com/RingBDStack/STAMImupter.