Synthetic Series-Symbol Data Generation for Time Series Foundation Models

Wenxuan Wang, Kai Wu, Yujian Betterest Li, Dan Wang, Xiaoyu Zhang

Keywords: Time Series Analysis, Foundation Model, Data Scarcity

TL;DR: Based on the complex dynamics perspective, this paper proposes a bimodal time series generation mechanism to mitigate the current data shortage problem in time series analysis.

Abstract: Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as training data scarcity and imbalance continue to hinder their development. Inspired by complex dynamic system theories, we design a series-symbol data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic expressions. To leverage series-symbol data pairs with strong correlations, we develop SymTime, a pre-trained foundation model for enhancing time series representation using symbolic information. SymTime demonstrates competitive performance across five major TSA tasks when fine-tunes with downstream tasks, rivaling foundation models pre-trained on real-world datasets. This approach underscores the potential of series-symbol data generation and pretraining mechanisms in overcoming data scarcity and enhancing task performance. The code is available at https://github.com/wwhenxuan/SymTime.

Primary Area: Deep learning (e.g., architectures, generative models, optimization for deep networks, foundation models, LLMs)

Submission Number: 5193