Summary of Spectral Graph Pruning Against Over-squashing and Over-smoothing, by Adarsh Jamadandi et al.
Spectral Graph Pruning Against Over-Squashing and Over-Smoothing
by Adarsh Jamadandi, Celia Rubio-Madrigal, Rebekka Burkholz
First submitted to arxiv on: 6 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Signal Processing (eess.SP); Machine Learning (stat.ML)
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 message passing graph neural network addresses the issues of over-squashing and over-smoothing by introducing edge deletions inspired by the Braess phenomenon. This approach improves generalization while reducing computational resources, connecting spectral gap optimization to pruning graphs for lottery tickets. The framework adds or deletes edges to optimize the spectral gap, demonstrating its effectiveness on large heterophilic datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Inspired by real-life traffic patterns, a new approach is proposed to improve message passing graph neural networks. These networks often struggle with over-squashing and over-smoothing, but deleting edges can help solve both problems simultaneously. This idea connects the concepts of spectral gap optimization and pruning graphs for lottery tickets. The result is a more effective framework that adds or deletes edges to optimize the spectral gap, showing its power on large datasets. |
Keywords
* Artificial intelligence * Generalization * Graph neural network * Optimization * Pruning
