Exploiting variational inequalities for generalized change detection on graphs

This repository contains relevant datasets and Python code to implement and reproduce methods and results presented in our paper titled:

Exploiting variational inequalities for generalized change detection on graphs.

The workflow and main elements of our framework are illustrated in the following figure:

Usage

To quickly familiarize yourself with the main functionality of our framework, we provide a Jupyter notebook demo_cd_with_proposed_models.ipynb. This notebook walks you through the steps to utilize our framework effectively.

If you're interested in replicating the experimental results, you can run main_experiments.py. This script executes the necessary computations and generates the same results as presented in our research paper.

Please ensure that you have the required dependencies installed before running the code. If you come across any bugs, have suggestions or questions for enhancements, I would greatly appreciate it if you could contact me at [email protected]. Your feedback is valuable in improving the repo's quality.

Installation

Follow these steps to install and set up the project:

1. Download our GitHub repository

2. Open an Anaconda Prompt (Anaconda3) as administrator, and set the current directory to the path of the project's folder.

3. Create the project's environment vi_gcd_env by running the following command:

conda env create -f environment.yml

4. Activate the created environment by running the appropriate command based on your preferred Python IDE or terminal:

• Jupyter Notebook/Lab: When starting a new notebook, select the vi_gcd_env environment from the kernel options.

• Visual Studio Code (VS Code): After opening your project in VS Code, click on the Python interpreter in the status bar and choose the vi_gcd_env environment.

Datasets

Datasets available here are provided to facilitate reproducible results. However, please note that they have not been collected by us, and proper attribution should be given if used for academic and research purposes. The datasets were downloaded from:

• Alaska, Atlantico, Mulargia~¹
• California ²
• Toulouse ³
• Shuguang ⁴

Citation

If you find our work insightful and our code useful, kindly cite the following paper:

@article{florez2023exploiting,
  title={Exploiting variational inequalities for generalized change detection on graphs},
  author={Florez-Ospina, Juan F and Jimenez-Sierra, David A and Benitez-Restrepo, Hernan D and Arce, Gonzalo R},
  volume={??},
  pages={1--16},
  year={2023},
  publisher={TechRxiv}
}

Footnotes

1. Jimenez, et al. "Dataset citation for Alaska, Atlantico, Mulargia." (2022). Link to citation ↩

2. Luppino, M. "Dataset citation for California." (2019). Link to citation ↩

3. Mignotte, M. "Dataset citation for Toulouse." (2020). Link to citation ↩

4. Sun, Y. "Dataset citation for Shuguang." (2022). Link to citation ↩