Summary of An End-to-end Attention-based Approach For Learning on Graphs, by David Buterez et al.
An end-to-end attention-based approach for learning on graphs
by David Buterez, Jon Paul Janet, Dino Oglic, Pietro Lio
First submitted to arxiv on: 16 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 The paper proposes a purely attention-based approach for learning on graphs, which consists of an encoder and an attention pooling mechanism. The encoder uses vertically interleaved masked and vanilla self-attention modules to learn effective representations of edges. This simplicity-focused method outperforms fine-tuned message passing baselines and transformer-based methods on over 70 node and graph-level tasks, including long-range benchmarks. It also demonstrates state-of-the-art performance across different tasks, such as molecular and vision graphs, and heterophilous node classification. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to work with graphs using something called attention. Attention is like a highlighter that helps computers focus on important parts of the graph. The authors created a simple method that uses this attention to learn about edges in the graph. This method works well and beats other methods on many tasks, including ones where the graphs are very big or complex. |
Keywords
* Artificial intelligence * Attention * Classification * Encoder * Self attention * Transformer
