Limitations of machine learning models when predicting compounds with completely new chemistries: possible improvements applied to the discovery of new non-fullerene acceptors.



Zhao, Zhi-Wen, Del Cueto, Marcos ORCID: 0000-0001-7744-8530 and Troisi, Alessandro ORCID: 0000-0002-5447-5648
(2022) Limitations of machine learning models when predicting compounds with completely new chemistries: possible improvements applied to the discovery of new non-fullerene acceptors. Digital discovery, 1 (3). pp. 266-276.

Access the full-text of this item by clicking on the Open Access link.

Abstract

We try to determine if machine learning (ML) methods, applied to the discovery of new materials on the basis of existing data sets, have the power to predict completely new classes of compounds (extrapolating) or perform well only when interpolating between known materials. We introduce the leave-one-group-out cross-validation, in which the ML model is trained to explicitly perform extrapolations of unseen chemical families. This approach can be used across materials science and chemistry problems to improve the added value of ML predictions, instead of using extrapolative ML models that were trained with a regular cross-validation. We consider as a case study the problem of the discovery of non-fullerene acceptors because novel classes of acceptors are naturally classified into distinct chemical families. We show that conventional ML methods are not useful in practice when attempting to predict the efficiency of a completely novel class of materials. The approach proposed in this work increases the accuracy of the predictions to enable at least the categorization of materials with a performance above and below the median value.

Item Type: Article
Uncontrolled Keywords: 4613 Theory Of Computation, 46 Information and Computing Sciences, 34 Chemical Sciences, Machine Learning and Artificial Intelligence
Divisions: Faculty of Science and Engineering > School of Physical Sciences
Depositing User: Symplectic Admin
Date Deposited: 13 Sep 2022 13:50
Last Modified: 21 Jun 2024 08:27
DOI: 10.1039/d2dd00004k
Open Access URL: https://doi.org/10.1039/D2DD00004K
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3164544