Self-Supervised Time Series Representation Learning with Temporal-Instance Similarity Distillation

Ainaz Hajimoradlou, Leila Pishdad, Frederick Tung, Maryna Karpusha

Keywords: self-supervised learning, time series, classification, anomaly detection, forecasting, similarity distillation, teacher-student architecture

Abstract: We propose a self-supervised method for pre-training universal time series representations in which we learn contrastive representations using similarity distillation along the temporal and instance dimensions. We analyze the effectiveness of both dimensions, and evaluate our pre-trained representations on three downstream tasks: time series classification, anomaly detection, and forecasting.

TL;DR: We propose a self-supervised methods for pre-training universal time series representations using similarity distillation.

Community Implementations: [ 8 code implementations](https://www.catalyzex.com/paper/self-supervised-time-series-representation/code)