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

Pre-trained Language Models have emerged as promising tools for predictingmolecular properties, yet their development is in its early stages,necessitating further research to enhance their efficacy and address challengessuch as generalization and sample efficiency. In this paper, we present amulti-view approach that combines latent spaces derived from state-of-the-artchemical models. Our approach relies on two pivotal elements: the embeddingsderived from MHG-GNN, which represent molecular structures as graphs, andMoLFormer embeddings rooted in chemical language. The attention mechanism ofMoLFormer is able to identify relations between two atoms even when theirdistance is far apart, while the GNN of MHG-GNN can more precisely capturerelations among multiple atoms closely located. In this work, we demonstratethe superior performance of our proposed multi-view approach compared toexisting state-of-the-art methods, including MoLFormer-XL, which was trained on1.1 billion molecules, particularly in intricate tasks such as predictingclinical trial drug toxicity and inhibiting HIV replication. We assessed ourapproach using six benchmark datasets from MoleculeNet, where it outperformedcompetitors in five of them. Our study highlights the potential of latent spacefusion and feature integration for advancing molecular property prediction. Inthis work, we use small versions of MHG-GNN and MoLFormer, which opens up anopportunity for further improvement when our approach uses a larger-scaledataset.

Improving Molecular Properties Prediction Through Latent Space Fusion