cfeng783 / nsibf

Code for KDD' 21 paper: Time Series Anomaly Detection for Cyber-physical Systems via Neural System Identification and Bayesian Filtering

NSIBF

Time Series Anomaly Detection for Cyber-physical Systems via Neural System Identification and Bayesian Filtering

Cheng Feng and Pengwei Tian. 2021. Time Series Anomaly Detection for Cyber-physical Systems via Neural System Identification and Bayesian Filtering. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD ’21), August 14–18, 2021, Virtual Event, Singapore. ACM, New York, NY, USA, 10 pages. https://doi.org/10.1145/3447548.3467137

Getting Started

Install dependencies (with python 3.6)

pip install -r requirements.txt

Run the qualitative experiment

cd experiments
python qualitative_experiment.py

Run the PUMP experiment

cd experiments
python PUMP_experiement.py

Run the WADI experiment

cd experiments
python WADI_experiement.py

Run the SWAT experiment

cd experiments
python SWAT_experiement.py

nsibf's Issues

Related questions in bf_search

Hello,In bf_search I see that you have used _adjust_predicts function which has adjusted the predicted values based on the real values, I would like to ask what is the reason for this?

Using IP data to detect anomalies

Hi, does this model support IP data from packet captures such as addresses, ports, date and time stamps or is does it exclusively deal with signals from sensors and actuators? What modifications would need to be made to accomodate such data?

Reproducibility of other baselines

Thanks for your great work!

I just wonder how you reproduce the other baseline models' results.

I cannot reproduce the results of SWaT dataset in case of USAD.
(https://github.com/robustml-eurecom/usad)

Also, it seems that your paper has the exact same results as the USAD paeprs'.