PARROT-Python

Python implementation of "PARROT: Position-Aware Regularized Optimal Transport for Network Alignment". The official implementation is here.

Prerequisites

• numpy
• scipy
• pytorch
• tqdm

Datasets

You can run main.py using one of the following datasets

• ACM-DBLP-A
• ACM-DBLP-P
• cora
• foursquare-twitter
• phone-email

Efficiency

Dataset / Runtime	MATLAB	PyTorch (CPU)	PyTorch (GPU, V100)
ACM_DBLP_A	12.32s + 53.95s	110.55s + 241.52s	28.48s + 24.64s
ACM_DBLP_P	15.44s + 59.52s	89.29s + 241.11s	21.67s + 24.48s
cora	7.20s + 8.07s	3.11s + 7.78s	3.36s + 0.77s
foursquare-twitter	6.89s + 17.59s	23.92s + 46.93s	8.53s + 4.03s
phone-email	0.32s + 1.18s	0.38s + 1.60s	1.36s + 0.33s

* Run on Apple M1 Pro 16GB

Usage

1. Clone the repository to your local machine:

git clone https://github.com/yq-leo/PARROT-Python.git

2. Navigate to the project directory:

cd PARROT-Python

3. Install the required dependencies:

pip install -r requirements.txt

4. To run PARROT, execute the following command in the terminal:

python main.py --dataset={dataset}

Reference

Official Code

PARROT-WWW23

Paper

Zeng, Z., Zhang, S., Xia, Y., & Tong, H. (2023, April). Parrot: Position-aware regularized optimal transport for network alignment. In Proceedings of the ACM Web Conference 2023 (pp. 372-382). DOI.