Computer Vision Self-supervised Learning Methods on Time Series
This work addresses the problem of applying computer vision SSL methods to time series data for researchers in machine learning, but it is incremental as it builds on existing frameworks with specific enhancements.
The study evaluated if computer vision self-supervised learning frameworks, primarily based on Siamese networks, are effective on time series data using UCR and UEA archives, showing they can be effective, and proposed a new method that improves on VICReg by modifying a covariance term and adding an iterative normalization layer to accelerate convergence.
Self-supervised learning (SSL) has had great success in both computer vision. Most of the current mainstream computer vision SSL frameworks are based on Siamese network architecture. These approaches often rely on cleverly crafted loss functions and training setups to avoid feature collapse. In this study, we evaluate if those computer-vision SSL frameworks are also effective on a different modality (\textit{i.e.,} time series). The effectiveness is experimented and evaluated on the UCR and UEA archives, and we show that the computer vision SSL frameworks can be effective even for time series. In addition, we propose a new method that improves on the recently proposed VICReg method. Our method improves on a \textit{covariance} term proposed in VICReg, and in addition we augment the head of the architecture by an iterative normalization layer that accelerates the convergence of the model.