Truncated Affinity Maximization: One-class Homophily Modeling for Graph Anomaly Detection

Authors: Hezhe Qiao, Guansong Pang

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive empirical results on 10 real-world GAD datasets show that TAM substantially outperforms seven competing models, achieving over 10% increase in AUROC/AUPRC compared to the best contenders on challenging datasets.
Researcher Affiliation Academia Hezhe Qiao, Guansong Pang School of Computing and Information Systems, Singapore Management University hezheqiao.2022@phdcs.smu.edu.sg, gspang@smu.edu.sg
Pseudocode Yes The training algorithms of TAM are summarized in Algorithm 1 and Algorithm 2. Algorithm 1 describes the process of NSGT. Algorithm 2 describes the training process of TAM.
Open Source Code Yes Our code is available at https://github.com/mala-lab/TAM-master/.
Open Datasets Yes We conduct the experiments on six commonly-used publicly-available real-world GAD datasets from diverse online shopping services and social networks, and citation networks, including Blog Catalog [49], ACM [48], Amazon [10], Facebook [56], Reddit, and Yelp Chi [21]. The first two datasets contain two types of injected anomalies contextual and structural anomalies [8,34] that are nodes with significantly deviated graph structure and node attributes respectively. The other four datasets contain real anomalies. Detailed information about the datasets can be found in App. B.
Dataset Splits No The paper mentions running experiments multiple times with different random seeds and using datasets for evaluation, but it does not specify explicit train/validation/test splits (e.g., 80/10/10 percentages or specific sample counts for each split) for its experiments. It refers to standard datasets but not their specific splits used.
Hardware Specification Yes All the experiments are run on an NVIDIA Ge Force RTX 3090 24GB GPU.
Software Dependencies Yes TAM is implemented in Pytorch 1.6.0 with python 3.7 and all the experiments are run on an NVIDIA Ge Force RTX 3090 24GB GPU.
Experiment Setup Yes In TAM, each LAMNet is implemented by a two-layer GCN, and its weight parameters are optimized using Adam [19] 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 Blog Catalog and ACM, require strong regularization, so λ = 1 is used by default; whereas λ = 0 is used for the four real-world datasets.