Summary of Teddy: Trimming Edges with Degree-based Discrimination Strategy, by Hyunjin Seo et al.
TEDDY: Trimming Edges with Degree-based Discrimination strategY
by Hyunjin Seo, Jihun Yun, Eunho Yang
First submitted to arxiv on: 2 Feb 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 This paper introduces TEDDY, a novel framework for finding graph lottery tickets (GLT) in graph neural networks (GNNs). Building on previous work, the authors aim to discover sparse GLTs that achieve comparable performance to dense networks. To address limitations in current approaches, TEDDY incorporates edge-degree information and leverages structural pathways to sparsify edges and model parameters simultaneously. The framework is designed for efficient one-shot training, leveraging graph structures alone without feature information. Experimental results show that TEDDY outperforms conventional iterative approaches in generalization. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research paper is about finding a special type of neural network called a “graph lottery ticket”. This ticket can be trained quickly and still perform as well as more complex networks. The authors created a new way to find these tickets, called TEDDY, which uses information about the connections between data points in the graph. They tested TEDDY and found that it works better than other methods for finding these tickets. This is important because it could help make AI systems more efficient and effective. |
Keywords
* Artificial intelligence * Generalization * Neural network * One shot
