Utilizing Expert Features for Contrastive Learning of Time-Series Representations
This addresses representation learning for time-series data in industrial or medical fields, where expert knowledge is available, and is incremental by building on contrastive learning with a novel objective.
The paper tackled the problem of learning time-series representations by incorporating expert features into contrastive learning, as data transformations are often elusive for such data. The result was that ExpCLR outperformed state-of-the-art methods on three real-world datasets for unsupervised and semi-supervised learning.
We present an approach that incorporates expert knowledge for time-series representation learning. Our method employs expert features to replace the commonly used data transformations in previous contrastive learning approaches. We do this since time-series data frequently stems from the industrial or medical field where expert features are often available from domain experts, while transformations are generally elusive for time-series data. We start by proposing two properties that useful time-series representations should fulfill and show that current representation learning approaches do not ensure these properties. We therefore devise ExpCLR, a novel contrastive learning approach built on an objective that utilizes expert features to encourage both properties for the learned representation. Finally, we demonstrate on three real-world time-series datasets that ExpCLR surpasses several state-of-the-art methods for both unsupervised and semi-supervised representation learning.