Deep Probabilistic Learning for Car Hacking Analysis

Please cite the following paper if you use our code.

Laha Ale, Scott A. King, and Ning Zhang, "Deep Bayesian Learning for Car Hacking Detection", Bayesian Deep Learning Workshop, 35th Conference on Neural Information Processing Systems

bibtex for citing the paper:

@InProceedings{aleneurips2021, title = {Deep Bayesian Learning for Car Hacking Detection, author = {Laha Ale, Scott King, and Ning Zhang}, booktitle = {Bayesian Deep Learning Workshop, 35th Conference on Neural Information Processing Systems}, year = {2021}, month = {6--12 Dec} }

1. Download Data

• Request data by access Car-Hacking Dataset for the intrusion detection
• Download the data files and save them into data folder

2. install required libraries

1) for ENV 1

• pip install pandas
• pip install numpy
• pip install matplotlib
• pip install sklearn
• pip install tensorflow-gpu
• pip install --upgrade tensorflow-probability

• 2) for ENV 2

• pip install pandas
• pip install numpy
• pip install matplotlib
• pip install sklearn
• conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch -c nvidia
• pip install pyro-ppl
• pip install gpytorch

3. Covert Attack-free(normal) text file to CSV

One way to process is: On linux run the follow code to convert txt file format to csv file

Original format:

Timestamp: 1479121500.969313 ID: 0140 000 DLC: 8 00 00 00 00 1a 00 24 ee

Expected format:

1479121500.969313,0140,8,00,00,00,00,1a,00,24,ee

• sed -i 's/Timestamp: //g' normal_run_data.txt
• sed -i 's/ ID: /,/g' normal_run_data.txt
• sed -i 's/ 000 DLC: /,/g' normal_run_data.txt
• sed -i 's/ /,/g' normal_run_data.txt
• sed -i 's/ /,/g' normal_run_data.txt

4. Play with notebooks

• int_data_small.ipynb
• bayesian_net.ipynb
• .... more