Unsupervised Scalable Representation Learning for Multivariate Time Series

2019/01/31 by Jean-Yves Franceschi, Aymeric Dieuleveut, Martin Jaggi · 6 citations
Computer Science · Mathematics · #cs.LG #cs.NE #stat.ML

paper · pdf

published as Thirty-third Conference on Neural Information Processing Systems, Neural Information Processing Systems Foundation, Dec 2019, Vancouver, Canada

arxiv created 2020/01/03 · arxiv updated 2020/01/06

Abstract

Time series constitute a challenging data type for machine learning algorithms, due to their highly variable lengths and sparse labeling in practice. In this paper, we tackle this challenge by proposing an unsupervised method to learn universal embeddings of time series. Unlike previous works, it is scalable with respect to their length and we demonstrate the quality, transferability and practicability of the learned representations with thorough experiments and comparisons. To this end, we combine an encoder based on causal dilated convolutions with a novel triplet loss employing time-based negative sampling, obtaining general-purpose representations for variable length and multivariate time series.

Cited by

• Predicting the outcome of team movements -- Player time series analysis using fuzzy and deep methods for representation learning
• CANet: ChronoAdaptive Network for Enhanced Long-Term Time Series Forecasting under Non-Stationarity
• Dynamic Graph-Like Learning with Contrastive Clustering on Temporally-Factored Ship Motion Data for Imbalanced Sea State Estimation in Autonomous Vessel
• Energy-Aware Pattern Disentanglement: A Generalizable Pattern Assisted Architecture for Multi-task Time Series Analysis
• A synthetic dataset of French electric load curves with temperature conditioning
• AimTS: Augmented Series and Image Contrastive Learning for Time Series Classification
• STaRFormer: Semi-Supervised Task-Informed Representation Learning via Dynamic Attention-Based Regional Masking for Sequential Data