Explaining Deep Graph Networks with Molecular Counterfactuals

2020/11/09 by Danilo Numeroso, Numeroso, Danilo, Davide Bacciu +1 · 1 citation
Computer Science · Materials Science · #Computational Drug Discovery Methods #Machine Learning in Materials Science #Explainable Artificial Intelligence (XAI)

paper · pdf · doi:10.48550/arxiv.2011.05134

Abstract

We present a novel approach to tackle explainability of deep graph networks in the context of molecule property prediction tasks, named MEG (Molecular Explanation Generator). We generate informative counterfactual explanations for a specific prediction under the form of (valid) compounds with high structural similarity and different predicted properties. We discuss preliminary results showing how the model can convey non-ML experts with key insights into the learning model focus in the neighborhood of a molecule.

Cited by

• Class-Aware Reinforcement Learning for Counterfactual Explanation Generation

Related

• MEG: Generating Molecular Counterfactual Explanations for Deep Graph Networks
• Global Concept Explanations for Graphs by Contrastive Learning
• Global Human-guided Counterfactual Explanations for Molecular Properties via Reinforcement Learning
• Fragment-Wise Interpretability in Graph Neural Networks via Molecule Decomposition and Contribution Analysis
• Bi-level Contrastive Learning for Knowledge-Enhanced Molecule Representations