Applications of Artificial Intelligence in Traditional Chinese Medicine

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Traditional Chinese medicine (TCM) represents a valuable part of Chinese cultural heritage. It is characterized by complex compositions and multiple therapeutic targets, which pose significant challenges to its modernization research. In recent years, the rapid advancement of artificial intelligence (AI) has enabled flexible, data-driven approaches that are now widely applied in TCM studies, including active ingredient analysis, quality control, and safety evaluation. This article reviews recent applications of AI in the TCM field, focusing on several key aspects: the advantages of graph neural network-based network pharmacology in analyzing multi-target interactions; computer vision and hyperspectral imaging for automatic medicinal material identification and quality grading; multimodal fusion techniques for component content detection and production process quality control; and graph attention networks and heterogeneous network representation learning for predicting TCM toxicity and assessing potential risks associated with combined use of Chinese and Western medicines. Based on a narrative review of the latest literature, this paper identifies current challenges such as data standardization and the lack of interpretability in model computation processes. Finally, it offers perspectives on how AI can support the standardization, precision, and global promotion of TCM.