TimesBERT: A BERT-Style Foundation Model for Time Series Understanding
It addresses the problem of versatile time series analysis for domains like classification and anomaly detection, positioning it as a foundational model, though it builds incrementally on BERT-style architectures.
The paper tackles the lack of a BERT-style foundation model for time series understanding by introducing TimesBERT, which learns generic representations from 260 billion time points and achieves state-of-the-art performance across four downstream tasks, outperforming task-specific models and language pre-trained backbones.
Time series analysis is crucial in diverse scenarios. Beyond forecasting, considerable real-world tasks are categorized into classification, imputation, and anomaly detection, underscoring different capabilities termed time series understanding in this paper. While GPT-style models have been positioned as foundation models for time series forecasting, the BERT-style architecture, which has made significant advances in natural language understanding, has not been fully unlocked for time series understanding, possibly attributed to the undesirable dropout of essential elements of BERT. In this paper, inspired by the shared multi-granularity structure between multivariate time series and multisentence documents, we design TimesBERT to learn generic representations of time series including temporal patterns and variate-centric characteristics. In addition to a natural adaptation of masked modeling, we propose a parallel task of functional token prediction to embody vital multi-granularity structures. Our model is pre-trained on 260 billion time points across diverse domains. Leveraging multi-granularity representations, TimesBERT achieves state-of-the-art performance across four typical downstream understanding tasks, outperforming task-specific models and language pre-trained backbones, positioning it as a versatile foundation model for time series understanding.