Reimagining GNN Explanations with ideas from Tabular Data

Anjali Singh; Shamanth R Nayak K; Balaji Ganesan

Reimagining GNN Explanations with ideas from Tabular Data

Machine Learning 2021-06-25 v1 Artificial Intelligence

Authors: Anjali Singh , Shamanth R Nayak K , Balaji Ganesan

Abstract

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.

Keywords

graph neural network knowledge graph graph generation

Cite

@article{arxiv.2106.12665,
  title  = {Reimagining GNN Explanations with ideas from Tabular Data},
  author = {Anjali Singh and Shamanth R Nayak K and Balaji Ganesan},
  journal= {arXiv preprint arXiv:2106.12665},
  year   = {2021}
}

Comments

4 pages, 8 figures, XAI Workshop at ICML 2021

Related papers

View all related →