Summary of Gnnanatomy: Systematic Generation and Evaluation Of Multi-level Explanations For Graph Neural Networks, by Hsiao-ying Lu et al.
GNNAnatomy: Systematic Generation and Evaluation of Multi-Level Explanations for Graph Neural Networks
by Hsiao-Ying Lu, Yiran Li, Ujwal Pratap Krishna Kaluvakolanu Thyagarajan, Kwan-Liu Ma
First submitted to arxiv on: 6 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Information Retrieval (cs.IR); Social and Information Networks (cs.SI)
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 research paper introduces GNNAnatomy, a visual analytics system designed to generate and evaluate multi-level explanations for Graph Neural Networks (GNNs) in graph classification tasks. The proposed method uses graphlets, primitive graph substructures, to identify the most critical substructures in a graph class by analyzing the correlation between GNN predictions and graphlet frequencies. These correlations are presented interactively through the visual analytics system. To further validate top-ranked graphlets, the paper measures the change in classification confidence after removing each graphlet from the original graph. The effectiveness of GNNAnatomy is demonstrated through case studies on synthetic and real-world graph datasets from sociology and biology domains. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research helps us understand how Graph Neural Networks (GNNs) make decisions by breaking them down into smaller parts called “graphlets”. These graphlets are like building blocks that help GNNs figure out which group a new graph belongs to. The researchers created a tool, called GNNAnatomy, that shows these graphlets and helps us see how important each one is for making predictions. They tested this tool on real-world data from sociology and biology fields and compared it to other similar tools. |
Keywords
* Artificial intelligence * Classification * Gnn
