Summary of Cldg: Contrastive Learning on Dynamic Graphs, by Yiming Xu et al.
CLDG: Contrastive Learning on Dynamic Graphs
by Yiming Xu, Bin Shi, Teng Ma, Bo Dong, Haoyi Zhou, Qinghua Zheng
First submitted to arxiv on: 19 Dec 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 paper introduces a novel framework called CLDG to address the issue of performance drops in downstream tasks when using graph contrastive learning (GCL) on dynamic graphs. GCL constructs self-supervised signals by maximizing mutual information between graph augmentation views, but this approach is affected by changes in semantics and labels during the augmentation process, leading to significant performance drops. To mitigate this problem, CLDG extracts temporally-persistent signals through a sampling layer, encouraging nodes to maintain consistent local and global representations. The framework outperforms eight unsupervised state-of-the-art baselines and shows competitiveness against four semi-supervised methods on seven datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about finding a way to make graphs more useful for learning. Graphs are like maps that connect things, but they can be hard to work with because they change over time. The researchers created a new approach called CLDG to help solve this problem. They found a way to take the changes in the graph into account and make it easier to learn from the graph. This is important because graphs are used in many areas like social media, medicine, and more. |
Keywords
» Artificial intelligence » Self supervised » Semantics » Semi supervised » Unsupervised
