[NeurIPS '21] Adversarial Attacks on Graph Classification via Bayesian Optimisation (GRABNEL)

Related tags

Deep Learninggrabnel
Overview

Adversarial Attacks on Graph Classification via Bayesian Optimisation @ NeurIPS 2021

overall-pipeline

This repository contains the official implementation of GRABNEL, a Bayesian optimisation-based adversarial agent to conduct adversarial attacks on graph classification models. GRABNEL currently supports various topological attacks, such as via edge flipping (incl. both addition or deletion), node injection and edge swapping. We also include implementations of a number of baseline methods including random search, genetic algorithm [1] and a gradient-based white-box attacker (available on some victim model choices). We also implement a number of victim models, namely:

  • Graph convolution networks (GCN) [2]
  • Graph isomorphism networks (GIN) [3]
  • ChebyGIN [4] (only for MNIST-75sp task)
  • Graph U-Net [5]
  • S2V (only for the ER Graph task in [1])

For details please take a look at our paper: abstract / pdf.

The code repository also contains instructions for the TU datasets [6] in the DGL framework, as well as the MNIST-75sp dataset in [4]. For the Twitter dataset we used for node injection tasks, we are not authorised to redistribute the dataset and you have to ask for permission from the authors of [7] directly.

If you find our work to be useful for your research, please consider citing us:

Wan, Xingchen, Henry Kenlay, Binxin Ru, Arno Blaas, Michael A. Osborne, and Xiaowen Dong. "Adversarial Attacks on Graph Classifiers via Bayesian Optimisation." In Thirty-Fifth Conference on Neural Information Processing Systems. 2021.

Or in bibtex:

@inproceedings{wan2021adversarial,
  title={Adversarial Attacks on Graph Classifiers via Bayesian Optimisation},
  author={Wan, Xingchen and Kenlay, Henry and Ru, Binxin and Blaas, Arno and Osborne, Michael and Dong, Xiaowen},
  booktitle={Thirty-Fifth Conference on Neural Information Processing Systems},
  year={2021}
}

Instructions for use

  1. Install the required packages in requirements.txt

For TU Dataset(s):

  1. Train a selected architecture (GCN/GIN). Taking an example of GCN training on the PROTEINS dataset. By default DGL will download the requested dataset under ~/.dgl directory. If it throws an error, you might have to manually download the dataset and add to the appropriate directory.
python3 train_model.py --dataset PROTEINS --model gcn --seed $YOUR_SEED 

This by default deposits the trained victim model under src/output/models and the training log under src/output/training_logs.

  1. Evaluate the victim model on a separate test set. Run
python3 evaluate_model.py --dataset PROTEINS --seed $YOUR_SEED  --model gcn

This by default will create evaluation logs under src/output/evaluation_logs.

  1. Run the attack algorithm.
cd scripts && python3 run_bo_tu.py --dataset PROTEINS --save_path $YOUR_SAVE_PATH --model_path $YOUR_MODEL_PATH --seed $YOUR_SEED --model gcn

With no method specified, the script runs GRABNEL by default. You may use the -m to specify if, for example, you'd like to run one of the baseline methods mentioned above instead.

For the MNIST-75sp task For MNIST-75sp, we use the pre-trained model released by the authors of [4] as the victim model, so there is no need to train a victim model separately (unless you wish to).

  1. Generate the MNIST-75sp dataset. Here we use an adapted script from [4], but added a converter to ensure that the dataset generated complies with the rest of our code base (DGL-compliant, etc). You need to download the MNIST dataset beforehand (or use the torchvision download facility. Either is fine)
cd data && python3 build_mnist.py -D mnist -d $YOUR_DATA_PATH -o $YOUR_SAVE_PATH  

The output should be a pickle file mnist_75sp.p. Place it under $PROJECT_ROOT/src/data/

  1. Download the pretrained model from https://github.com/bknyaz/graph_attention_pool. The pretrained checkpointed model we use is checkpoint_mnist-75sp_139255_epoch30_seed0000111.pth.tar. Deposit the model under src/output/models

  2. Run attack algorithm.

cd scripts && python3 run_bo_image_classification.py --dataset mnist

References

[1] Dai, Hanjun, Hui Li, Tian Tian, Xin Huang, Lin Wang, Jun Zhu, and Le Song. "Adversarial attack on graph structured data." In International conference on machine learning, pp. 1115-1124. PMLR, 2018.

[2] Kipf, Thomas N., and Max Welling. "Semi-supervised classification with graph convolutional networks." arXiv preprint arXiv:1609.02907 (2016).

[3] Xu, Keyulu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. "How powerful are graph neural networks?." arXiv preprint arXiv:1810.00826 (2018).

[4] Knyazev, Boris, Graham W. Taylor, and Mohamed R. Amer. "Understanding attention and generalization in graph neural networks." NeurIPS (2019).

[5] Gao, Hongyang, and Shuiwang Ji. "Graph u-nets." In international conference on machine learning, pp. 2083-2092. PMLR, 2019.

[6] Morris, Christopher, Nils M. Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, and Marion Neumann. "Tudataset: A collection of benchmark datasets for learning with graphs." arXiv preprint arXiv:2007.08663 (2020).

[7] Vosoughi, Soroush, Deb Roy, and Sinan Aral. "The spread of true and false news online." Science 359, no. 6380 (2018): 1146-1151.

Acknowledgements

The repository builds, directly or indirectly, on multiple open-sourced code bases available online. The authors would like to express their gratitudes towards the maintainers of the following repos:

  1. https://github.com/Hanjun-Dai/graph_adversarial_attack
  2. https://github.com/DSE-MSU/DeepRobust
  3. https://github.com/HongyangGao/Graph-U-Nets
  4. https://github.com/xingchenwan/nasbowl
  5. The Deep graph library team
  6. The grakel team (https://ysig.github.io/GraKeL/0.1a8/)
Owner
Xingchen Wan
PhD Student in Machine Learning @ University of Oxford
Xingchen Wan
Source Code for DialogBERT: Discourse-Aware Response Generation via Learning to Recover and Rank Utterances (https://arxiv.org/pdf/2012.01775.pdf)

DialogBERT This is a PyTorch implementation of the DialogBERT model described in DialogBERT: Neural Response Generation via Hierarchical BERT with Dis

Xiaodong Gu 67 Jan 06, 2023
PyTorch original implementation of Cross-lingual Language Model Pretraining.

XLM NEW: Added XLM-R model. PyTorch original implementation of Cross-lingual Language Model Pretraining. Includes: Monolingual language model pretrain

Facebook Research 2.7k Dec 27, 2022
Conceptual 12M is a dataset containing (image-URL, caption) pairs collected for vision-and-language pre-training.

Conceptual 12M We introduce the Conceptual 12M (CC12M), a dataset with ~12 million image-text pairs meant to be used for vision-and-language pre-train

Google Research Datasets 226 Dec 07, 2022
Generating images from caption and vice versa via CLIP-Guided Generative Latent Space Search

CLIP-GLaSS Repository for the paper Generating images from caption and vice versa via CLIP-Guided Generative Latent Space Search An in-browser demo is

Federico Galatolo 172 Dec 22, 2022
Semantic Segmentation in Pytorch

PyTorch Semantic Segmentation Introduction This repository is a PyTorch implementation for semantic segmentation / scene parsing. The code is easy to

Hengshuang Zhao 1.2k Jan 01, 2023
Fast Neural Style for Image Style Transform by Pytorch

FastNeuralStyle by Pytorch Fast Neural Style for Image Style Transform by Pytorch This is famous Fast Neural Style of Paper Perceptual Losses for Real

Bengxy 81 Sep 03, 2022
Suite of 500 procedurally-generated NLP tasks to study language model adaptability

TaskBench500 The TaskBench500 dataset and code for generating tasks. Data The TaskBench dataset is available under wget http://web.mit.edu/bzl/www/Tas

Belinda Li 20 May 17, 2022
Reinforcement Learning Theory Book (rus)

Reinforcement Learning Theory Book (rus)

qbrick 206 Nov 27, 2022
Temporal Segment Networks (TSN) in PyTorch

TSN-Pytorch We have released MMAction, a full-fledged action understanding toolbox based on PyTorch. It includes implementation for TSN as well as oth

1k Jan 03, 2023
Fully Automatic Page Turning on Real Scores

Fully Automatic Page Turning on Real Scores This repository contains the corresponding code for our extended abstract Henkel F., Schwaiger S. and Widm

Florian Henkel 7 Jan 02, 2022
Solve a Rubiks Cube using Python Opencv and Kociemba module

Rubiks_Cube_Solver Solve a Rubiks Cube using Python Opencv and Kociemba module Main Steps Get the countours of the cube check whether there are tota

Adarsh Badagala 176 Jan 01, 2023
Code for our CVPR2021 paper coordinate attention

Coordinate Attention for Efficient Mobile Network Design (preprint) This repository is a PyTorch implementation of our coordinate attention (will appe

Qibin (Andrew) Hou 726 Jan 05, 2023
Simple tutorials using Google's TensorFlow Framework

TensorFlow-Tutorials Introduction to deep learning based on Google's TensorFlow framework. These tutorials are direct ports of Newmu's Theano Tutorial

Nathan Lintz 6k Jan 06, 2023
[CVPR'21] FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space

FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space by Quande Liu, Cheng Chen, Ji

Quande Liu 178 Jan 06, 2023
Code accompanying the paper "Knowledge Base Completion Meets Transfer Learning"

Knowledge Base Completion Meets Transfer Learning This code accompanies the paper Knowledge Base Completion Meets Transfer Learning published at EMNLP

14 Nov 27, 2022
auto-tuning momentum SGD optimizer

YellowFin YellowFin is an auto-tuning optimizer based on momentum SGD which requires no manual specification of learning rate and momentum. It measure

Jian Zhang 288 Nov 19, 2022
Massively parallel Monte Carlo diffusion MR simulator written in Python.

Disimpy Disimpy is a Python package for generating simulated diffusion-weighted MR signals that can be useful in the development and validation of dat

Leevi 16 Nov 11, 2022
Rainbow is all you need! A step-by-step tutorial from DQN to Rainbow

Do you want a RL agent nicely moving on Atari? Rainbow is all you need! This is a step-by-step tutorial from DQN to Rainbow. Every chapter contains bo

Jinwoo Park (Curt) 1.4k Dec 29, 2022
TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-Captured Scenarios

TPH-YOLOv5 This repo is the implementation of "TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-Captured

cv516Buaa 439 Dec 22, 2022
This repository is to support contributions for tools for the Project CodeNet dataset hosted in DAX

The goal of Project CodeNet is to provide the AI-for-Code research community with a large scale, diverse, and high quality curated dataset to drive innovation in AI techniques.

International Business Machines 1.2k Jan 04, 2023