Kernel-U-Net: Symmetric and Hierarchical Architecture for Multivariate Time Series Forecasting

3 Jan 2024  ·  Jiang You, René Natowicz, Arben Cela, Jacob Ouanounou, Patrick Siarry ·

Time series forecasting task predicts future trends based on historical information. Transformer-based U-Net architectures, despite their success in medical image segmentation, have limitations in both expressiveness and computation efficiency in time series forecasting as evidenced in YFormer. To tackle these challenges, we introduce Kernel-U-Net, a symmetric and hierarchical U-shape neural network architecture. The kernel-U-Net encoder compresses gradually input series into latent vectors, and its symmetric decoder subsequently expands these vectors into output series. Specifically, Kernel-U-Net separates the procedure of partitioning input time series into patches from kernel manipulation, thereby providing the convenience of executing customized kernels. Our method offers two primary advantages: 1) Flexibility in kernel customization to adapt to specific datasets; 2) Enhanced computational efficiency, with the complexity of the Transformer layer reduced to linear. Experiments on seven real-world datasets, considering both multivariate and univariate settings, demonstrate that Kernel-U-Net's performance either exceeds or meets that of the existing state-of-the-art model PatchTST in the majority of cases and outperforms Yformer. The source code for Kernel-U-Net will be made publicly available for further research and application.

PDF Abstract

Datasets


Results from the Paper


  Submit results from this paper to get state-of-the-art GitHub badges and help the community compare results to other papers.

Methods