On the Inclusion of Spatial Information for Spatio-Temporal Neural Networks

TL;DR

This paper challenges the common practice of incorporating prior spatial information into spatio-temporal neural networks, showing that models without such information often perform just as well across various real-world datasets.

Contribution

It provides a comprehensive empirical comparison demonstrating that spatial agnostic neural networks can be as effective as models using prior spatial knowledge.

Findings

Spatial agnostic models perform comparably to spatially informed models.

Prior spatial information is often unnecessary for effective spatio-temporal modeling.

Results are validated on ten real-world datasets related to mobility and air quality.

Abstract

When confronting a spatio-temporal regression, it is sensible to feed the model with any available prior information about the spatial dimension. For example, it is common to define the architecture of neural networks based on spatial closeness, adjacency, or correlation. A common alternative, if spatial information is not available or is too costly to introduce it in the model, is to learn it as an extra step of the model. While the use of prior spatial knowledge, given or learnt, might be beneficial, in this work we question this principle by comparing spatial agnostic neural networks with state of the art models. Our results show that the typical inclusion of prior spatial information is not really needed in most cases. In order to validate this counterintuitive result, we perform thorough experiments over ten different datasets related to sustainable mobility and air quality, substantiating our conclusions on real world problems with direct implications for public health and economy.