Issue 13, 2022

Model agnostic generation of counterfactual explanations for molecules


An outstanding challenge in deep learning in chemistry is its lack of interpretability. The inability of explaining why a neural network makes a prediction is a major barrier to deployment of AI models. This not only dissuades chemists from using deep learning predictions, but also has led to neural networks learning spurious correlations that are difficult to notice. Counterfactuals are a category of explanations that provide a rationale behind a model prediction with satisfying properties like providing chemical structure insights. Yet, counterfactuals have been previously limited to specific model architectures or required reinforcement learning as a separate process. In this work, we show a universal model-agnostic approach that can explain any black-box model prediction. We demonstrate this method on random forest models, sequence models, and graph neural networks in both classification and regression.

Graphical abstract: Model agnostic generation of counterfactual explanations for molecules

Supplementary files

Article information

Article type
Edge Article
22 Sep 2021
06 Feb 2022
First published
16 Feb 2022
This article is Open Access

All publication charges for this article have been paid for by the Royal Society of Chemistry
Creative Commons BY license

Chem. Sci., 2022,13, 3697-3705

Model agnostic generation of counterfactual explanations for molecules

G. P. Wellawatte, A. Seshadri and A. D. White, Chem. Sci., 2022, 13, 3697 DOI: 10.1039/D1SC05259D

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. You can use material from this article in other publications without requesting further permissions from the RSC, provided that the correct acknowledgement is given.

Read more about how to correctly acknowledge RSC content.

Social activity