Dispute over reaction prediction puts machine learning’s pitfalls in spotlight

A picture showing corrupted data

Source: © Shutterstock

Two research teams’ argument over a reaction-predicting algorithm show that there is still a lot to understand when applying machine learning to chemistry

An algorithm designed to predict reaction outcomes is now disputed by another research team who says it fails classical controls in machine learning. Although the study’s authors disagree with this assessment, the controversy highlights the potential pitfalls of machine learning methods as they gain traction. A decade ago, there were only a few hundred chemistry machine learning publications. In 2018, almost 8000 articles in Web of Science’s chemistry collection contained the keyword.