Explainable Uncertainty Quantifications for Deep Learning-Based Molecular Property Prediction

12 September 2022, Version 1
This content is a preprint and has not undergone peer review at the time of posting.

Abstract

Quantifying uncertainty in machine learning is important in new research areas with scarce high-quality data. In this work, we develop an explainable uncertainty quantification method for deep learning-based molecular property prediction. This method can capture aleatoric and epistemic uncertainties separately and attribute the uncertainties to atoms present in the molecule. The atom-based uncertainty method provides an extra layer of explainability to the estimated uncertainties, i.e., one can analyze individual atomic uncertainty values to diagnose the chemical component that introduces uncertainty to the prediction. Our experiments demonstrate that atomic uncertainty can detect unseen chemical structures and identify chemical species whose data are potentially associated with significant noise. Furthermore, we propose a post-hoc calibration method to refine the uncertainty quantified by ensemble models for better confidence interval estimates. This work improves uncertainty calibration and provides a framework for assessing whether and why a prediction should be considered unreliable.

Keywords

uncertainty quantifications
deep learning
explainable AI
molecular property prediction

Supplementary materials

Title
Description
Actions
Title
Supporting Information
Description
Additional information as noted in the text, including confidence- and error-based calibration curves for the datasets listed in Table 2, correlation coefficient matrixes for atomic uncertainties, and complete lists of atom and bond features, are provided in the Supporting Information.
Actions

Comments

Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.