Optimizing Time Series Forecasting Architectures: A Hierarchical Neural Architecture Search Approach

Difan Deng, Marius Lindauer

Keywords: Neural Architecture Search, Time Series Forecasting

TL;DR: We propose a general differentiable neural architecture search space for time series forecasting that fits most of the forecasting models, finding architectures comparable to hand-designed models but with much less computational resources required.

Abstract: The rapid development of time series forecasting research has brought many deep learning-based modules in this field. However, despite the increasing amount of new forecasting architectures, it is still unclear if we have leveraged the full potential of these existing modules within a properly designed architecture. In this work, we propose a novel hierarchical neural architecture search approach for time series forecasting tasks. With the design of a hierarchical search space, we incorporate many architecture types designed for forecasting tasks and allow for the efficient combination of different forecasting architecture modules. Results on long-term-time-series-forecasting tasks show that our approach can search for lightweight high-performing forecasting architectures across different forecasting tasks.

Submission Number: 53