MEG: Generating Molecular Counterfactual Explanations for Deep Graph Networks

TL;DR

This paper introduces MEG, a reinforcement learning-based method for generating valid molecular counterfactual explanations to improve interpretability of deep graph networks in chemistry applications.

Contribution

The paper presents a novel approach combining reinforcement learning and validity constraints to generate meaningful counterfactual explanations for molecular property predictions.

Findings

MEG produces structurally similar molecules with different predicted properties.

The method offers insights into the model's focus areas around specific molecules.

MEG enhances trust and interpretability in deep graph networks for chemistry.

Abstract

Explainable AI (XAI) is a research area whose objective is to increase trustworthiness and to enlighten the hidden mechanism of opaque machine learning techniques. This becomes increasingly important in case such models are applied to the chemistry domain, for its potential impact on humans' health, e.g, toxicity analysis in pharmacology. In this paper, we present a novel approach to tackle explainability of deep graph networks in the context of molecule property prediction t asks, 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. Given a trained DGN, we train a reinforcement learning based generator to output counterfactual explanations. At each step, MEG feeds the current candidate counterfactual into the DGN, collects the prediction and uses it to reward the RL agent to guide the exploration. Furthermore, we restrict the action space of the agent in order to only keep actions that maintain the molecule in a valid state. We discuss the results showing how the model can convey non-ML experts with key insights into the learning model focus in the neighbourhood of a molecule.