ADGC: Awesome Deep Graph Clustering
Introduction
Awesome-Deep-Graph-Clustering (ADGC) is an extensive collection of state-of-the-art (SOTA) methods, papers, code, and datasets dedicated to deep graph clustering. This project is a hub for those who are delving into the world of graph clustering and wish to explore the latest advancements and methodologies. It not only serves as a repository of novel techniques but also opens doors for the contribution of new and interesting papers and codes. Researchers and developers who find this repository useful are encouraged to star it on GitHub, serving as a token of appreciation and community support.
What is Deep Graph Clustering?
Deep graph clustering involves uncovering the hidden structure within graphs and categorizing nodes into various groups. This area of research has gained significant attention due to its potential applications in numerous fields such as social network analysis, bioinformatics, and recommendation systems. It focuses on leveraging deep learning to enhance the effectiveness and efficiency of clustering tasks on graph-structured data. For those interested in a detailed survey of this field, ADGC provides links to comprehensive papers.
Important Survey Papers
ADGC compiles influential survey papers that provide a broad understanding of the developments in deep graph clustering. Here are a few notable mentions:
- An Overview of Advanced Deep Graph Node Clustering (2023, TCSS): This paper presents a detailed overview of deep graph node clustering.
- A Survey of Deep Graph Clustering: Taxonomy, Challenge, and Application (2022, arXiv): This survey discusses the categorical challenges and applications of deep graph clustering.
- A Comprehensive Survey on Community Detection with Deep Learning (2022, TNNLS): It explores the landscape of community detection through deep learning techniques.
Key Contributions and Code Repositories
ADGC hosts a variety of pioneering deep graph clustering methodologies across different categories:
New-architecture Deep Graph Clustering:
- Kolmogorov-Arnold Network (KAN) for Graphs (2024): A fresh approach to graph clustering. [Code].
Temporal Deep Graph Clustering:
- Deep Temporal Graph Clustering (TGC) (2024, ICLR): Focuses on clustering in dynamic graphs. Paper, Code.
Deep Graph Clustering with Unknown Cluster Number:
- Reinforcement Graph Clustering with Unknown Cluster Number (RGC) (2023, ACM MM): Tackles the challenge of clustering without a predefined number of clusters. Paper, Code.
Reconstructive Deep Graph Clustering:
- Embedding-Induced Graph Refinement Clustering Network (EGRC-Net) (2023, TIP): Enhances clustering by refining graph embeddings. Paper, Code.
Adversarial Methods
ADGC also explores adversarial methods in deep graph clustering, which involve leveraging adversarial approaches to improve the robustness and generalization of clustering models. This area continues to evolve, offering exciting opportunities for research.
Conclusion
ADGC is a valuable resource for anyone working in the field of deep graph clustering. It provides a comprehensive insight into existing methodologies, backed up by scholarly papers and practical code implementations. The community aspect of the project welcomes new contributions, fostering a collaborative environment for continuous development and innovation in deep graph clustering techniques. Whether you are a researcher seeking to enhance your work with the latest technologies or someone entering the realm of graph clustering, ADGC offers a wealth of information and inspiration.