LGAIJun 23, 2021

Reimagining GNN Explanations with ideas from Tabular Data

arXiv:2106.12665v12 citations
Originality Synthesis-oriented
AI Analysis

This work identifies shortcomings in GNN explainability for researchers and practitioners, but it is incremental as it primarily comments on existing issues without introducing new methods.

The paper addresses the gap in explainability techniques for Graph Neural Networks compared to those for tabular data models, using Entity Matching as a case study to highlight missing aspects in GNN explanations.

Explainability techniques for Graph Neural Networks still have a long way to go compared to explanations available for both neural and decision decision tree-based models trained on tabular data. Using a task that straddles both graphs and tabular data, namely Entity Matching, we comment on key aspects of explainability that are missing in GNN model explanations.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes