English

Time-Series Representation Learning via Temporal and Contextual Contrasting

Machine Learning 2021-06-29 v1 Artificial Intelligence

Abstract

Learning decent representations from unlabeled time-series data with temporal dynamics is a very challenging task. In this paper, we propose an unsupervised Time-Series representation learning framework via Temporal and Contextual Contrasting (TS-TCC), to learn time-series representation from unlabeled data. First, the raw time-series data are transformed into two different yet correlated views by using weak and strong augmentations. Second, we propose a novel temporal contrasting module to learn robust temporal representations by designing a tough cross-view prediction task. Last, to further learn discriminative representations, we propose a contextual contrasting module built upon the contexts from the temporal contrasting module. It attempts to maximize the similarity among different contexts of the same sample while minimizing similarity among contexts of different samples. Experiments have been carried out on three real-world time-series datasets. The results manifest that training a linear classifier on top of the features learned by our proposed TS-TCC performs comparably with the supervised training. Additionally, our proposed TS-TCC shows high efficiency in few-labeled data and transfer learning scenarios. The code is publicly available at https://github.com/emadeldeen24/TS-TCC.

Keywords

Cite

@article{arxiv.2106.14112,
  title  = {Time-Series Representation Learning via Temporal and Contextual Contrasting},
  author = {Emadeldeen Eldele and Mohamed Ragab and Zhenghua Chen and Min Wu and Chee Keong Kwoh and Xiaoli Li and Cuntai Guan},
  journal= {arXiv preprint arXiv:2106.14112},
  year   = {2021}
}

Comments

Accepted in IJCAI-21 conference ... please cite the conference version

R2 v1 2026-06-24T03:37:56.500Z