AutoSTF: Decoupled Neural Architecture Search for Cost-Effective Automated Spatio-Temporal Forecasting

TL;DR

AutoSTF introduces a decoupled neural architecture search framework that significantly reduces computational costs and enhances accuracy in automated spatio-temporal forecasting for smart city applications.

Contribution

It proposes a novel decoupled search strategy and multi-patch transfer module, enabling cost-effective and more effective spatio-temporal dependency modeling.

Findings

Achieves up to 13.48x speed-up over existing methods.

Maintains superior forecasting accuracy across eight datasets.

Demonstrates effectiveness in both efficiency and accuracy improvements.

Abstract

Spatio-temporal forecasting is a critical component of various smart city applications, such as transportation optimization, energy management, and socio-economic analysis. Recently, several automated spatio-temporal forecasting methods have been proposed to automatically search the optimal neural network architecture for capturing complex spatio-temporal dependencies. However, the existing automated approaches suffer from expensive neural architecture search overhead, which hinders their practical use and the further exploration of diverse spatio-temporal operators in a finer granularity. In this paper, we propose AutoSTF, a decoupled automatic neural architecture search framework for cost-effective automated spatio-temporal forecasting. From the efficiency perspective, we first decouple the mixed search space into temporal space and spatial space and respectively devise representation compression and parameter-sharing schemes to mitigate the parameter explosion. The decoupled spatio-temporal search not only expedites the model optimization process but also leaves new room for more effective spatio-temporal dependency modeling. From the effectiveness perspective, we propose a multi-patch transfer module to jointly capture multi-granularity temporal dependencies and extend the spatial search space to enable finer-grained layer-wise spatial dependency search. Extensive experiments on eight datasets demonstrate the superiority of AutoSTF in terms of both accuracy and efficiency. Specifically, our proposed method achieves up to 13.48x speed-up compared to state-of-the-art automatic spatio-temporal forecasting methods while maintaining the best forecasting accuracy.