Machine Learning Applied in Small Molecule Drug Discovery: Models, Strategies, and Future Prospects

The application of machine learning (ML) in small-molecule drug discovery has expanded rapidly, garnering considerable attention in recent years. A thorough understanding of ML principles and their practical applications is increasingly essential for pharmacologists and drug researchers. Despite notable advancements, significant challenges persist, including limited data quality, difficulties in feature selection, and restricted model generalizability. This review systematically surveys the landscape of ML algorithms, categorizes them by model type, and highlights how various ML tools and techniques have been developed to address specific challenges at different stages of the drug discovery process.