Fast Temporal Wavelet Graph Neural Networks

TL;DR

The paper introduces FTWGNN, a novel graph neural network architecture that leverages wavelet theory and multiresolution analysis to efficiently forecast spatio-temporal signals in neuroscience and transportation, achieving competitive performance with low computational cost.

Contribution

It presents a new wavelet-based GNN architecture that is both time- and memory-efficient for processing complex timeseries data on graphs.

Findings

Achieves competitive accuracy on traffic and neural datasets.

Maintains low computational footprint compared to state-of-the-art methods.

Utilizes multiresolution matrix factorization for efficient graph processing.

Abstract

Spatio-temporal signals forecasting plays an important role in numerous domains, especially in neuroscience and transportation. The task is challenging due to the highly intricate spatial structure, as well as the non-linear temporal dynamics of the network. To facilitate reliable and timely forecast for the human brain and traffic networks, we propose the Fast Temporal Wavelet Graph Neural Networks (FTWGNN) that is both time- and memory-efficient for learning tasks on timeseries data with the underlying graph structure, thanks to the theories of multiresolution analysis and wavelet theory on discrete spaces. We employ Multiresolution Matrix Factorization (MMF) (Kondor et al., 2014) to factorize the highly dense graph structure and compute the corresponding sparse wavelet basis that allows us to construct fast wavelet convolution as the backbone of our novel architecture. Experimental results on real-world PEMS-BAY, METR-LA traffic datasets and AJILE12 ECoG dataset show that FTWGNN is competitive with the state-of-the-arts while maintaining a low computational footprint. Our PyTorch implementation is publicly available at https://github.com/HySonLab/TWGNN