Recently, various contrastive learning techniques have been developed to categorize time series data and have exhibited promising performance for real-world applications. A general paradigm is to ...