Summary of Explaining Graph Neural Networks Via Structure-aware Interaction Index, by Ngoc Bui et al.
Explaining Graph Neural Networks via Structure-aware Interaction Index
by Ngoc Bui, Hieu Trung Nguyen, Viet Anh Nguyen, Rex Ying
First submitted to arxiv on: 23 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 the Myerson-Taylor interaction index, a novel framework for interpreting black-box machine learning models with structured inputs, such as graph neural networks. Unlike existing Shapley-based approaches, this method internalizes the graph structure and decomposes coalitions into components satisfying a pre-chosen connectivity criterion. The authors propose MAGE (Myerson-Taylor Structure-Aware Graph Explainer), which uses the second-order Myerson-Taylor index to identify key motifs influencing model predictions. The paper proves that the Myerson-Taylor index satisfies five natural axioms accounting for graph structure and high-order interactions among nodes. Extensive experiments on various graph datasets and models demonstrate MAGE’s superior subgraph explanations compared to state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine trying to understand why a computer program made a certain decision, but it’s too complicated and hard to see how the pieces fit together. This paper introduces a new way to help explain these decisions by taking into account the relationships between different parts of the information. Instead of just looking at individual pieces, this method looks at how they work together to make a prediction. The authors test their idea on several datasets and show that it works better than other methods for explaining these kinds of predictions. |
Keywords
» Artificial intelligence » Machine learning
