SpoT-Mamba: Learning Long-Range Dependency on Spatio-Temporal Graphs with Selective State Spaces

TL;DR

SpoT-Mamba is a novel framework for spatio-temporal graph forecasting that effectively captures long-range dependencies using node-specific walk sequences and state space modeling, improving traffic prediction accuracy.

Contribution

It introduces SpoT-Mamba, a new method combining node-specific walks and state space models to better capture long-range spatio-temporal dependencies in forecasting.

Findings

Outperforms existing methods on traffic datasets

Effectively captures long-range dependencies

Improves forecasting accuracy

Abstract

Spatio-temporal graph (STG) forecasting is a critical task with extensive applications in the real world, including traffic and weather forecasting. Although several recent methods have been proposed to model complex dynamics in STGs, addressing long-range spatio-temporal dependencies remains a significant challenge, leading to limited performance gains. Inspired by a recently proposed state space model named Mamba, which has shown remarkable capability of capturing long-range dependency, we propose a new STG forecasting framework named SpoT-Mamba. SpoT-Mamba generates node embeddings by scanning various node-specific walk sequences. Based on the node embeddings, it conducts temporal scans to capture long-range spatio-temporal dependencies. Experimental results on the real-world traffic forecasting dataset demonstrate the effectiveness of SpoT-Mamba.