Summary of Graphxain: Narratives to Explain Graph Neural Networks, by Mateusz Cedro et al.
GraphXAIN: Narratives to Explain Graph Neural Networks
by Mateusz Cedro, David Martens
First submitted to arxiv on: 4 Nov 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 Graph Neural Networks (GNNs) are a powerful machine learning technique for processing graph-structured data but struggle with interpretability. Existing GNN explanation methods provide technical outputs that are difficult to comprehend, violating the purpose of explanations. To address this challenge, we propose GraphXAIN, a model-agnostic method that leverages Large Language Models (LLMs) to translate explanatory subgraphs and feature importance scores into natural language narratives explaining GNN predictions. Our evaluations on real-world datasets demonstrate GraphXAIN’s ability to improve graph explanations. Furthermore, our survey of machine learning researchers and practitioners reveals that GraphXAIN enhances multiple explainability dimensions, including understandability, satisfaction, convincingness, suitability for communicating model predictions, trustworthiness, insightfulness, confidence, and usability. By providing natural language narratives, our approach benefits both GNN experts and non-experts by offering clearer and more effective explanations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you’re trying to understand why a machine learning model made a certain prediction on a graph. Existing methods make it hard for people without technical backgrounds to understand the explanation. To fix this, we created GraphXAIN, which uses special language models to translate complex data into easy-to-understand stories about how the model works. We tested GraphXAIN on real-world data and found that it improves explanations. We also asked machine learning experts what they thought of GraphXAIN and found that it makes explanations clearer, more convincing, and easier to use. |
Keywords
* Artificial intelligence * Gnn * Machine learning
