Interpreting Graph Neural Networks with Myerson Values for Cheminformatics Approaches
Preprint
- 8 May 2023
- preprint
- Published by American Chemical Society (ACS) in ChemRxiv
Abstract
Here we introduce a novel method to interpret the predictions of graph neural networks (GNNs) based on Myerson values from cooperative game theory. Myerson values are closely related to Shapley values and thus provide an interpretability approach similar to the SHAP values. We developed the technique for applications in drug discovery, but it can be used with any graph. Using the GNN as a coalition game and the interpreted graph as the cooperation structure, the Myerson values determine the worth of each node of the graph. The worth of all nodes of the graph adds up to the predicted value of the model, allowing for a simple and intuitive interpretation of the prediction. To interpret predictions on molecular graphs we show visual explanations on molecular structures using two molecular datasets.Keywords
All Related Versions
This publication has 13 references indexed in Scilit:
- Everything is connected: Graph neural networksCurrent Opinion in Structural Biology, 2023
- EdgeSHAPer: Bond-centric Shapley value-based explanation method for graph neural networksiScience, 2022
- Benchmarking Molecular Feature Attribution Methods with Activity CliffsJournal of Chemical Information and Modeling, 2022
- GraphSVX: Shapley Value Explanations for Graph Neural Networks2021
- Interpretation of Compound Activity Predictions from Complex Machine Learning Models Using Local Approximations and Shapley ValuesJournal of Medicinal Chemistry, 2019
- Unmasking Clever Hans predictors and assessing what machines really learnNature Communications, 2019
- Molecular fingerprint similarity search in virtual screeningMethods, 2015
- ZINC − A Free Database of Commercially Available Compounds for Virtual ScreeningJournal of Chemical Information and Modeling, 2004
- Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings 1PII of original article: S0169-409X(96)00423-1. The article was originally published in Advanced Drug Delivery Reviews 23 (1997) 3–25. 1Advanced Drug Delivery Reviews, 2001
- Prediction of Physicochemical Parameters by Atomic ContributionsJournal of Chemical Information and Computer Sciences, 1999