Molecular Descriptors Property Prediction Using Transformer-Based Approach

In this study, we introduce semi-supervised machine learning models designed to predict molecular properties. Our model employs a two-stage approach, involving pre-training and fine-tuning. Particularly, our leverages substantial amount of labeled unlabeled data consisting SMILES strings, text representation system for molecules. During the stage, capitalizes on Masked Language Model, which is widely used in natural language processing, chemical space representations. fine-tuning trained smaller dataset tackle specific downstream tasks, such as classification or regression. Preliminary results indicate that demonstrates comparable performance state-of-the-art chosen tasks from MoleculeNet. Additionally, reduce computational overhead, propose new approach taking advantage 3D compound structures calculating attention score end-to-end transformer anti-malaria drug candidates. The show using proposed score, able have with pre-trained models.