Summary of Thesaurus: Contrastive Graph Clustering by Swapping Fused Gromov-wasserstein Couplings, By Bowen Deng et al.
THESAURUS: Contrastive Graph Clustering by Swapping Fused Gromov-Wasserstein Couplings
by Bowen Deng, Tong Wang, Lele Fu, Sheng Huang, Chuan Chen, Tao Zhang
First submitted to arxiv on: 16 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 The proposed method, conTrastive grapH clustEring by SwApping fUsed gRomov-wasserstein coUplingS (THESAURUS), aims to improve graph node clustering by introducing semantic prototypes, cross-view assignment prediction pretext tasks, and Gromov-Wasserstein Optimal Transport. The method addresses drawbacks in existing methods, such as the Uniform Effect and Cluster Assimilation, which stem from a lack of contextual information, misaligned training tasks, and underutilization of graph structure cluster information. THESAURUS updates prototype graphs and marginal distributions using momentum to adapt to diverse real-world data. Experimental results demonstrate higher cluster separability compared to prior art. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary THESAURUS is a new way to group similar things on social media networks or other types of connections. Right now, this task is hard because current methods don’t use enough information about the connections between things. THESAURUS solves this problem by giving each group its own special description, and using that to figure out where things belong. It also uses a new way to measure how similar things are based on their connections. This helps to make sure that groups are well-defined and clear, even when there’s not much difference between them. |
Keywords
» Artificial intelligence » Clustering
