New paper at ICML 2024!

Authors:
Ivan Marisca, Cesare Alippo, Filippo Maria Bianchi

Arxiv:
https://arxiv.org/abs/2402.10634

Repo:
https://github.com/marshka/hdtts

Description:
Spatiotemporal graph neural networks achieve striking results by representing the relationships across time series as a graph. Nevertheless, most existing methods rely on the often unrealistic assumption that inputs are always available and fail to capture hidden spatiotemporal dynamics when part of the data is missing. In this work, we tackle this problem through hierarchical spatiotemporal downsampling. The input time series are progressively coarsened over time and space, obtaining a pool of representations that capture heterogeneous temporal and spatial dynamics. Conditioned on observations and missing data patterns, such representations are combined by an interpretable attention mechanism to generate the forecasts.