Summary of Better Not to Propagate: Understanding Edge Uncertainty and Over-smoothing in Signed Graph Neural Networks, by Yoonhyuk Choi et al.
Better Not to Propagate: Understanding Edge Uncertainty and Over-smoothing in Signed Graph Neural Networks
by Yoonhyuk Choi, Jiho Choi, Taewook Ko, Chong-Kwon Kim
First submitted to arxiv on: 9 Aug 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 addresses the over-smoothing issue in traditional Graph Neural Networks (GNNs) by introducing a novel approach for estimating homophily and edge error ratio. This allows for dynamic selection between blocked and signed propagation during training, which is more effective under high edge error ratios. Theoretical analysis and experiments demonstrate improved performance in both homophilic and heterophilic graphs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you’re trying to understand relationships between people or things that are different from each other. Traditional ways of looking at these relationships can get lost when there’s a lot of noise or mistakes. This paper suggests a new way to look at this problem by combining two techniques: one that helps eliminate errors and another that chooses the best approach based on how similar the data is. The results show that this method works better than others in certain situations. |
