Cross-City Transfer Learning for Deep Spatio-Temporal Prediction

TL;DR

This paper introduces RegionTrans, a transfer learning method that leverages data from one city to improve deep spatio-temporal predictions in another city with limited data, demonstrating significant accuracy improvements.

Contribution

The paper proposes a novel cross-city transfer learning framework, including region matching and feature transfer, for deep spatio-temporal prediction tasks in urban computing.

Findings

RegionTrans outperforms existing models by up to 10.7% in prediction accuracy.

Effective transfer of knowledge reduces data requirements for target cities.

RegionTrans demonstrates robustness across different urban scenarios.

Abstract

Spatio-temporal prediction is a key type of tasks in urban computing, e.g., traffic flow and air quality. Adequate data is usually a prerequisite, especially when deep learning is adopted. However, the development levels of different cities are unbalanced, and still many cities suffer from data scarcity. To address the problem, we propose a novel cross-city transfer learning method for deep spatio-temporal prediction tasks, called RegionTrans. RegionTrans aims to effectively transfer knowledge from a data-rich source city to a data-scarce target city. More specifically, we first learn an inter-city region matching function to match each target city region to a similar source city region. A neural network is designed to effectively extract region-level representation for spatio-temporal prediction. Finally, an optimization algorithm is proposed to transfer learned features from the source city to the target city with the region matching function. Using citywide crowd flow prediction as a demonstration experiment, we verify the effectiveness of RegionTrans. Results show that RegionTrans can outperform the state-of-the-art fine-tuning deep spatio-temporal prediction models by reducing up to 10.7% prediction error.