Decoding Herbal Medicine: Machine Learning-Driven Insights into Structural Identification and Pharmacological...

Herbal medicine is indispensable and crucial to many healthcare systems. Its therapeutic value arises from diverse plant-derived compounds with potential pharmacological activities. This review summarizes recent advances in applying artificial intelligence (AI) and machine learning (ML) to elucidate the pharmacological mechanisms of herbal medicines, and offers a particular focus on revealing their bioactive compounds and pharmacological actions. It highlights how AI-driven analytical frameworks are enhancing the precision and efficiency of mechanism elucidation. Information relevant to herbal medicine and related computational studies was collected via keyword searching in common scientific databases including Google Scholar, PubMed, ACS Publications, Wiley Online Library, ScienceDirect, SpringerLink, Scopus, IEEE Xplore, ACM Digital Library, and Web of Science by using keywords "herbal medicine", "machine learning", "deep learning", "natural compounds", "docking", "QSAR", "toxicity", "mass spectrometry", "nuclear magnetic resonance", "feature extraction", "Absorption", "Distribution", "Metabolism", and "Excretion". The review finds that AI, particularly ML and deep learning approaches, has been applied to identify bioactive compounds, predict potential targets and mechanistic pathways, and thereby enable the identification of hidden patterns in complex chemical-biological datasets. Recent advances in AI have improved the understanding of the complex components and biological effects of herbal preparations. However, the variability of plant materials and the lack of systematic pharmacological studies remain the major challenges. Approaches that combine experimental data with computational analysis show promise for identifying bioactive compounds, predicting their targets, and clarifying the underlying mechanisms.