Self-supervised molecular pretraining strategy for low-resource reaction prediction scenarios

27 July 2021, Version 2
This content is a preprint and has not undergone peer review at the time of posting.

Abstract

In the face of low-resource reaction training samples, we construct a chemical platform for addressing small-scale reaction prediction problem. By using a self-supervised pretraining strategy called MASS, the transformer model can absorb the chemical information about 1 billion molecules and then finetunes on small-scale reaction prediction, which is different from previous works that only rely on reaction samples. To demonstrate the broad applicability of our approach, we adopt three dif-ferent name reactions in our work. In the Baeyer-Villiger, Heck and Sharpless asymmetric epoxidation reaction prediction tasks, the average accuracies increase by 5.7%, 10.8%, 4.8% respectively, marking an important step to low-resource reaction prediction.

Keywords

Deep Learning
Self-supervised pretraining
Organic Chemistry
Reaction Prediction
Molecules
MASS
Transformer
SMILES string representation

Supplementary materials

Title
Description
Actions
Title
Self-supervised molecular pretraining strategy for low-resource reaction prediction scenarios
Description
Supplementary Materials for self-supervised molecular pretraining strategy for low-resource reaction prediction scenarios
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.