Forecaster: A Graph Transformer for Forecasting Spatial and Time-Dependent Data
This work addresses forecasting challenges in applications like urban mobility, but it appears incremental as it builds on existing Transformer and graph-based methods.
The paper tackles the problem of forecasting spatial and time-dependent data, such as taxi ride-hailing demand, by proposing Forecaster, a graph Transformer architecture that addresses complex spatial and temporal dependencies, non-stationarity, and heterogeneity, and it significantly outperforms state-of-the-art baselines.
Spatial and time-dependent data is of interest in many applications. This task is difficult due to its complex spatial dependency, long-range temporal dependency, data non-stationarity, and data heterogeneity. To address these challenges, we propose Forecaster, a graph Transformer architecture. Specifically, we start by learning the structure of the graph that parsimoniously represents the spatial dependency between the data at different locations. Based on the topology of the graph, we sparsify the Transformer to account for the strength of spatial dependency, long-range temporal dependency, data non-stationarity, and data heterogeneity. We evaluate Forecaster in the problem of forecasting taxi ride-hailing demand and show that our proposed architecture significantly outperforms the state-of-the-art baselines.