Learning to Embed Time Series Patches Independently

TL;DR

This paper proposes a novel approach for time series representation learning that independently embeds patches using autoencoding and contrastive learning, outperforming existing Transformer-based methods in forecasting and classification tasks.

Contribution

It introduces a simple patch-wise embedding method with independent autoencoding and contrastive learning, challenging the dependency-based patch modeling in prior work.

Findings

Improves time series forecasting accuracy

Enhances classification performance

Reduces model complexity and training time

Abstract

Masked time series modeling has recently gained much attention as a self-supervised representation learning strategy for time series. Inspired by masked image modeling in computer vision, recent works first patchify and partially mask out time series, and then train Transformers to capture the dependencies between patches by predicting masked patches from unmasked patches. However, we argue that capturing such patch dependencies might not be an optimal strategy for time series representation learning; rather, learning to embed patches independently results in better time series representations. Specifically, we propose to use 1) the simple patch reconstruction task, which autoencode each patch without looking at other patches, and 2) the simple patch-wise MLP that embeds each patch independently. In addition, we introduce complementary contrastive learning to hierarchically capture adjacent time series information efficiently. Our proposed method improves time series forecasting and classification performance compared to state-of-the-art Transformer-based models, while it is more efficient in terms of the number of parameters and training/inference time. Code is available at this repository: https://github.com/seunghan96/pits.