Summary of Hc-glad: Dual Hyperbolic Contrastive Learning For Unsupervised Graph-level Anomaly Detection, by Yali Fu et al.
HC-GLAD: Dual Hyperbolic Contrastive Learning for Unsupervised Graph-Level Anomaly Detection
by Yali Fu, Jindong Li, Jiahong Liu, Qianli Xing, Qi Wang, Irwin King
First submitted to arxiv on: 2 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Social and Information Networks (cs.SI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel approach to unsupervised graph-level anomaly detection (UGAD) called Dual Hyperbolic Contrastive Learning for Unsupervised Graph-Level Anomaly Detection (HC-GLAD). Existing UGAD methods mainly focus on pairwise relationships between first-order neighbors, neglecting high-order node interactions and underlying properties like hierarchy and power-law structure. HC-GLAD constructs hypergraphs based on gold motifs to capture high-order node group information and performs hypergraph convolution. Additionally, it introduces hyperbolic geometry to preserve the hierarchy of real-world graphs through graph and hypergraph embedding learning in hyperbolic space with the hyperboloid model. This approach outperforms existing methods on 13 real-world datasets from various fields. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine a special kind of detective that can find unusual patterns in complex networks like social media or biological systems. This paper proposes a new way for this detective to work, called HC-GLAD. The current methods are limited because they only look at close friends and ignore the bigger picture. HC-GLAD is different because it looks at groups of people who might be connected in unexpected ways. It also takes into account the underlying structure of these networks, which can help it find anomalies more accurately. The authors tested their method on 13 real-world datasets and found that it outperformed existing methods. |
Keywords
* Artificial intelligence * Anomaly detection * Embedding * Unsupervised
