ST-UNet: A Spatio-Temporal U-Network for Graph-structured Time Series Modeling.

The spatio- temporalgraph learning is becoming an increasingly important object of graph study. Many application domains involve highly dynamic graphs where temporalinformation is crucial, e.g. traffic networks and financial transactiongraphs. Despite the constant progress made on learning structured data, there is still a lack of effective means to extract dynamic complex features from spatio- temporalstructures. Particularly, conventional models such as convolutional networks or recurrent neural networks are incapable of revealing the temporalpatterns in short or long terms and exploring the spatial properties in local or global scope from spatio- temporalgraphs simultaneously. To tackle this problem, we design a novel multi-scale architecture, Spatio- TemporalU-Net (ST-UNet), for graph-structured time series modeling. In this U-shaped network, a paired sampling operation is proposed in spacetime domain accordingly: the pooling (ST-Pool) coarsens the input graph in spatial from its deterministic partition while abstracts multi-resolution temporaldependencies through dilated recurrent skip connections; based on previous settings in the downsampling, the unpooling (ST-Unpool) restores the original structure of spatio- temporalgraphs and resumes regular intervals within graph sequences. Experiments on spatio- temporalprediction tasks demonstrate that our model effectively captures comprehensive features in multiple scales and achieves substantial improvements over mainstream methods on several real-world datasets.