PeFAD: A Parameter-Efficient Federated Framework for Time Series Anomaly Detection

Train and Evaluate

You can reproduce the experiment results by running each script in ./scripts/:

bash ./scripts/SMD.sh
bash ./scripts/PSM.sh
bash ./scripts/SWAT.sh
bash ./scripts/MSL.sh

Citation

Please cite the following paper if this paper/repository is useful for your research.

@inproceedings{xu2024pefad,
  title={PeFAD: A Parameter-Efficient Federated Framework for Time Series Anomaly Detection},
  author={Xu, Ronghui and Miao, Hao and Wang, Senzhang and Yu, Philip S and Wang, Jianxin},
  booktitle={SIGKDD},
  pages={3621--3632},
  year={2024}
}

Datasets

SMD, PSM, SWaT, and MSL can be downloaded from Google Drive.

Biref Dataset Details

Server Machine Dataset (SMD) is a 5-week-long dataset collected from a large Internet company with a time granularity of 1 minute. Please refer to Robust Anomaly Detection for Multivariate Time Series through Stochastic Recurrent Neural Network.
Pooled Server Metrics (PSM) is a public dateset collected internally from multiple application server nodes at eBay. Please refer to Practical Approach to Asynchronous Multivariate Time Series Anomaly Detection and Localization.
Secure Water Treatment (SWaT) is obtained from 51 sensors of the critical infrastructure system under continuous operations. Please refer to SWaT: a water treatment testbed for research and training on ICS security.
NASA's Mars Science Laboratory (MSL) dataset, collected during the spacecraft's journey to Mars, is a valuable resource accessible through NASA-designated data centers. Please refer to Detecting Spacecraft Anomalies Using LSTMs and Nonparametric Dynamic Thresholding.

Environment

Python Version: 3.8
PyTorch Version: 1.7.1
Run the following script for environment configuration.
```
pip install -r requirements.txt
```

xu737 / PeFAD