Time-Series Representation Learning via Temporal and Contextual Contrasting

26 Jun 2021  ·  Emadeldeen Eldele, Mohamed Ragab, Zhenghua Chen, Min Wu, Chee Keong Kwoh, XiaoLi Li, Cuntai Guan ·

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.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Epilepsy Prediction Epilepsy seizure prediction TS-TCC 1:1 Accuracy 97.23 # 1
Recognizing And Localizing Human Actions HAR TS-TCC 1:1 Accuracy 90.37 # 1
Automatic Sleep Stage Classification Sleep-EDF TS-TCC Accuracy 83.0 # 3


No methods listed for this paper. Add relevant methods here