Unsupervised Scalable Representation Learning for Multivariate Time Series - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2019

Unsupervised Scalable Representation Learning for Multivariate Time Series

Résumé

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.
Fichier principal
Vignette du fichier
article-supplementary-NeurIPS19.pdf (1001.04 Ko) Télécharger le fichier
code.zip (2.91 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01998101 , version 1 (29-01-2019)
hal-01998101 , version 2 (11-06-2019)
hal-01998101 , version 3 (01-11-2019)
hal-01998101 , version 4 (18-12-2019)

Licence

Paternité

Identifiants

Citer

Jean-Yves Franceschi, Aymeric Dieuleveut, Martin Jaggi. Unsupervised Scalable Representation Learning for Multivariate Time Series. Thirty-third Conference on Neural Information Processing Systems, Neural Information Processing Systems Foundation, Dec 2019, Vancouver, Canada. pp.4650--4661. ⟨hal-01998101v4⟩

Relations

696 Consultations
356 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More