(ICLR 2024) Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting

This code is a PyTorch implementation of our ICLR'24 paper "Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting". [arXiv]

🌟 Pathformer代码在阿里云仓库也进行同步更新：阿里云Pathformer代码链接

Citing Pathformer

If you find this resource helpful, please consider to cite our research:

@inproceedings{chen2024pathformer,
  author       = {Peng Chen and Yingying Zhang and Yunyao Cheng and Yang Shu and Yihang Wang and Qingsong Wen and Bin Yang and Chenjuan Guo},
  title        = {Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting},
  booktitle    = {International Conference on Learning Representations (ICLR)},
  year         = {2024}
}

Introduction

Pathformer, a Multi-Scale Transformer with Adaptive Pathways for time series forecasting. It integrates multi-scale temporal resolutions and temporal distances by introducing patch division with multiple patch sizes and dual attention on the divided patches, enabling the comprehensive modeling of multi-scale characteristics. Furthermore, adaptive pathways dynamically select and aggregate scale-specific characteristics based on the different temporal dynamics.

The important components of Pathformer: Multi-Scale Transformer Block and Multi-Scale Router.

Requirements

To install all dependencies:

pip install -r requirements.txt

Datasets

You can access the well pre-processed datasets from Google Drive, then place the downloaded contents under ./dataset

Quick Demos

1. Download datasets and place them under ./dataset
2. Run each script in scripts/, for example

bash scripts/multivariate/ETTm2.sh

Further Reading

1, Transformers in Time Series: A Survey, in IJCAI 2023. [GitHub Repo]

@inproceedings{wen2023transformers,
  title={Transformers in time series: A survey},
  author={Wen, Qingsong and Zhou, Tian and Zhang, Chaoli and Chen, Weiqi and Ma, Ziqing and Yan, Junchi and Sun, Liang},
  booktitle={International Joint Conference on Artificial Intelligence(IJCAI)},
  year={2023}
}