English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The prediction of material properties plays a crucial role in the developmentand discovery of materials in diverse applications, such as batteries,semiconductors, catalysts, and pharmaceuticals. Recently, there has been agrowing interest in employing data-driven approaches by using machine learningtechnologies, in combination with conventional theoretical calculations. Inmaterial science, the prediction of unobserved values, commonly referred to asextrapolation, is particularly critical for property prediction as it enablesresearchers to gain insight into materials beyond the limits of available data.However, even with the recent advancements in powerful machine learning models,accurate extrapolation is still widely recognized as a significantlychallenging problem. On the other hand, self-supervised pretraining is amachine learning technique where a model is first trained on unlabeled datausing relatively simple pretext tasks before being trained on labeled data fortarget tasks. As self-supervised pretraining can effectively utilize materialdata without observed property values, it has the potential to improve themodel's extrapolation ability. In this paper, we clarify how suchself-supervised pretraining can enhance extrapolation performance.We propose anexperimental framework for the demonstration and empirically reveal that whilemodels were unable to accurately extrapolate absolute property values,self-supervised pretraining enables them to learn relative tendencies ofunobserved property values and improve extrapolation performance.

Is Self-Supervised Pretraining Good for Extrapolation in Molecular Property Prediction?