RWL-GNN: Robustifying Graph Neural Networks via Weighted Laplacian

¹ IIT Delhi,² Samsung Research Bangalore

A PyTorch implementation of "Robustifying Graph Neural Networks via Weighted Laplacian" Accepted at SPCOM 2022 (Best Student Paper Award).

Abstract

Graph neural network (GNN) is achieving remarkable performances in a variety of application domains. However, GNN is vulnerable to noise and adversarial attacks in input data. Making GNN robust against noises and adversarial attacks is an important problem. The existing defense methods for GNNs are computationally demanding, are not scalable, and are architecture dependent. In this paper, we propose a generic framework for robustifying GNN known as Weighted Laplacian GNN (RWL-GNN). The method combines Weighted Graph Laplacian learning with the GNN implementation. The proposed method benefits from the positive semi-definiteness property of Laplacian matrix, feature smoothness, and latent features via formulating a unified optimization framework, which ensures the ad- versarial/noisy edges are discarded and connections in the graph are appropriately weighted. For demonstration, the experiments are conducted with Graph convo- lutional neural network(GCNN) architecture, however, the proposed framework is easily amenable to any existing GNN architecture. The simulation results with benchmark dataset establish the efficacy of the proposed method over the state- of-the-art methods, both in accuracy and computational efficiency.

Requirements

See that in https://github.com/DSE-MSU/DeepRobust/blob/master/requirements.txt

Installation

To run the code, first you need to install DeepRobust:

pip install deeprobust

Or you can clone it and install from source code:

git clone https://github.com/DSE-MSU/DeepRobust.git
cd DeepRobust
python setup.py install

Run the code

After installation, you can clone this repository or you can use the demo notebook here COLAB

git clone https://github.com/Bharat-Runwal/RWL-GNN.git
cd RWL-GNN
python train.py --two_stage y --seed 10  --dataset cora --attack no --ptb_rate 0 --epochs 200  --epochs_pre 400 --alpha 1.0  --gamma 1.0 --beta 0.10 --lr_optim 1e-2 --lr 1e-3

Acknowledgements

The code is based on :

DeepRobust (https://github.com/DSE-MSU/DeepRobust)
Pro-GNN

Cite:

@article{runwal2022robust,
  title={Robust Graph Neural Networks using Weighted Graph Laplacian},
  author={Runwal, Bharat and Kumar, Sandeep and others},
  journal={arXiv preprint arXiv:2208.01853},
  year={2022}
}