Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

TimesURL: Self-supervised Contrastive Learning for Universal Time Series Representation Learning

About

Learning universal time series representations applicable to various types of downstream tasks is challenging but valuable in real applications. Recently, researchers have attempted to leverage the success of self-supervised contrastive learning (SSCL) in Computer Vision(CV) and Natural Language Processing(NLP) to tackle time series representation. Nevertheless, due to the special temporal characteristics, relying solely on empirical guidance from other domains may be ineffective for time series and difficult to adapt to multiple downstream tasks. To this end, we review three parts involved in SSCL including 1) designing augmentation methods for positive pairs, 2) constructing (hard) negative pairs, and 3) designing SSCL loss. For 1) and 2), we find that unsuitable positive and negative pair construction may introduce inappropriate inductive biases, which neither preserve temporal properties nor provide sufficient discriminative features. For 3), just exploring segment- or instance-level semantics information is not enough for learning universal representation. To remedy the above issues, we propose a novel self-supervised framework named TimesURL. Specifically, we first introduce a frequency-temporal-based augmentation to keep the temporal property unchanged. And then, we construct double Universums as a special kind of hard negative to guide better contrastive learning. Additionally, we introduce time reconstruction as a joint optimization objective with contrastive learning to capture both segment-level and instance-level information. As a result, TimesURL can learn high-quality universal representations and achieve state-of-the-art performance in 6 different downstream tasks, including short- and long-term forecasting, imputation, classification, anomaly detection and transfer learning.

Jiexi Liu, Songcan Chen• 2023

Related benchmarks

TaskDatasetResultRank
Multivariate ForecastingETTh1
MSE0.79
909
Multivariate Time-series ForecastingETTm1
MSE0.59
742
Multivariate Time-series ForecastingETTm2
MSE0.8
593
Multivariate Time-series ForecastingWeather
MSE0.3
466
Multivariate Time-series ForecastingETTh2
MSE1.83
219
Time-series classificationSelfRegulationSCP2
Accuracy54.4
148
Time-series classificationEthanolConcentration
Accuracy34.2
63
Multivariate Time Series ClassificationStandWalkJump
Accuracy43.3
45
Multivariate Time Series ClassificationMotorImagery
Accuracy46
41
Multivariate Time Series ClassificationHandMovementDirection
Accuracy36.1
36
Showing 10 of 25 rows

Other info

Follow for update