DT+GNN: A Fully Explainable Graph Neural Network using Decision TreesDownload PDF

Published: 01 Feb 2023, Last Modified: 17 Sept 2023Submitted to ICLR 2023Readers: Everyone
TL;DR: A new GNN architecture that allows for full explanation not only of the important imputs but also the full decision making process how the inputs are used.
Abstract: We propose a new Decision Tree Graph Neural Network (DT+GNN) architecture for Graph Neural Network (GNN) explanation. Existing post-hoc explanation methods highlight important inputs but fail to reveal how a GNN uses these inputs. In contrast DT+GNN is fully explainable: Humans can inspect and understand the decision making of DT+GNN at every step. DT+GNN internally uses a novel GNN layer that is restricted to categorical state spaces for nodes and messages. After training with gradient descent we can easily distill these layers into decision trees. These trees are further pruned using our newly proposed method to ensure they are small and easy to interpret. DT+GNN can also compute node-level importance scores like the existing explanation methods. We demonstrate on real-world GNN benchmarks that DT+GNN has competitive classification accuracy and computes competitive explanations. Furthermore, we leverage DT+GNN's full explainability to inspect the decision processes in synthetic and real-world datasets with surprising results. We make this inspection accessible through an interactive web tool.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Supplementary Material: zip
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics
Submission Guidelines: Yes
Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 5 code implementations](https://www.catalyzex.com/paper/arxiv:2205.13234/code)
11 Replies