Repository files navigation

This repository is the official implementation of NeurIPS'23 paper "Truncated Affinity Maximization: One-classHomophily Modeling for Graph Anomaly Detection"

The full paper can be found atNeurIPS Portal orarXiv.

📋 We explore the property one class homophily to introduce a novel unsupervised anomaly scoring measure for GAD -- local node affinity -- that assigns a larger anomaly score to nodes that are less affiliated with their neighbors, with the affinity defined as similarity on node attributes/representations.We further propose Truncated Affinity Maximization (TAM) that learns tailored node representations for our anomaly measure by maximizing the local affinity of nodes to their neighbors.TAM is instead optimized on truncated graphs where non-homophily edges are removed iteratively to mitigate this bias. Extensive empirical results on six real-world GAD datasets show that TAM substantially outperforms seven competing models

To install requirements:

pip install -r requirements.txt

📋 TAM is implemented in Pytorch 1.6.0 with Python 3.7 and all the experiments are run on an NVIDIA GeForce RTX 3090 24GB GPU. For large-scale graph datasets, like T-Finance and YelpChi-all, the experiments are run on NVIDIA A100 GPU for quickness. The experiments on OGB-Protein are run on the CPU.

BlogCatalog and ACM were downloaded fromhttps://github.com/yixinliu233/CoLA
Amazon and Yelpchi were downloaded fromhttps://github.com/YingtongDou/CARE-GNN
Amazon-all and Yelpchi-all were downloaded fromhttps://github.com/YingtongDou/CARE-GNN
Facebook is obtained fromhttps://github.com/zhiming-xu/conad
Reddit is downloaded fromhttps://github.com/pygod-team/data
T-finance is downloaded fromhttps://drive.google.com/drive/folders/1PpNwvZx_YRSCDiHaBUmRIS3x1rZR7fMr
OGB-Protein is downloaded fromhttps://ogb.stanford.edu/docs/nodeprop/#ogbn-arxiv
For convenience, all the datasets can be obtained fromhttps://drive.google.com/drive/folders/1qcDBcVdcfAr_q5VOXBYagtnhA_r3Mm3Z?usp=drive_link

To train the model(s) in the paper, run this command:

python train.py

📋 In TAM, each LAMNet is implemented by a two-layer GCN, and its weight parameters are optimized using Adam optimizer with 500 epochs and a learning rate of 1e-5 by default. T=3 and K=4 are used for all datasets. Datasets with injected anomalies, such as BlogCatalog, ACM and Facebook, require strong regularization, so$\lambda=1$ is used by default; whereas$\lambda=0$ is used for the four real-world datasets.For the larger datasets like Amazon-all, YelpChi-all, and OGB-Protein, they require larger truncation times due to the large number of edges. So we set K = 7.

To evaluate our model on datasets, run:

python train.py

Our model achieves the following performance on :

We reveal an important anomaly-discriminative property, the one-class homophily, in GAD datasets with either injected or real anomalies. We utilize this property to introduce a novel unsupervised GAD measure, local node affinity, and further introduce a truncated affinity maximization (TAM) approach that end-to-end optimizes the proposed anomaly measure on truncated adjacency matrix with the non-homophily edges eliminated

If you use this package and find it useful, please cite our paper using the following BibTeX. Thanks! :)

@inproceedings{qiao2023truncated,title={Truncated Affinity Maximization: One-class Homophily Modeling for Graph Anomaly Detection},author={Qiao, Hezhe and Pang, Guansong},booktitle={Advances in Neural Information Processing Systems},year={2023}}