A Systematic Evaluation of Generated Time Series and Their Effects in Self-Supervised Pretraining

TL;DR

This paper systematically evaluates the impact of generated time series data on self-supervised pretraining, finding that synthetic data can enhance classification performance when real data is scarce.

Contribution

It introduces a comprehensive assessment of time series generation methods and demonstrates that synthetic data can improve self-supervised pretraining effectiveness.

Findings

Generated data improves classification accuracy in low-data scenarios.

Replacing real data with synthetic data can outperform traditional pretraining.

Synthetic data volume correlates positively with performance gains.

Abstract

Self-supervised Pretrained Models (PTMs) have demonstrated remarkable performance in computer vision and natural language processing tasks. These successes have prompted researchers to design PTMs for time series data. In our experiments, most self-supervised time series PTMs were surpassed by simple supervised models. We hypothesize this undesired phenomenon may be caused by data scarcity. In response, we test six time series generation methods, use the generated data in pretraining in lieu of the real data, and examine the effects on classification performance. Our results indicate that replacing a real-data pretraining set with a greater volume of only generated samples produces noticeable improvement.