Towards Expressive Spectral-Temporal Graph Neural Networks for Time Series Forecasting
This work addresses a theoretical gap for researchers in time series forecasting, providing a foundation for designing more expressive GNN-based models, though it is incremental in building on existing spectral-temporal methods.
The paper tackles the lack of theoretical understanding of spectral-temporal graph neural networks for time series forecasting by establishing a framework that proves their universality and bounded expressive power, and proposes TGGC, which outperforms existing models with better efficiency.
Time series forecasting has remained a focal point due to its vital applications in sectors such as energy management and transportation planning. Spectral-temporal graph neural network is a promising abstraction underlying most time series forecasting models that are based on graph neural networks (GNNs). However, more is needed to know about the underpinnings of this branch of methods. In this paper, we establish a theoretical framework that unravels the expressive power of spectral-temporal GNNs. Our results show that linear spectral-temporal GNNs are universal under mild assumptions, and their expressive power is bounded by our extended first-order Weisfeiler-Leman algorithm on discrete-time dynamic graphs. To make our findings useful in practice on valid instantiations, we discuss related constraints in detail and outline a theoretical blueprint for designing spatial and temporal modules in spectral domains. Building on these insights and to demonstrate how powerful spectral-temporal GNNs are based on our framework, we propose a simple instantiation named Temporal Graph Gegenbauer Convolution (TGGC), which significantly outperforms most existing models with only linear components and shows better model efficiency. Our findings pave the way for devising a broader array of provably expressive GNN-based models for time series.