TABLE 3.

Application of AI in quality standardization of TCM.

Aim of study	AI methods	Results	REF
Verify the advantages of two-dimensional correlation spectral images (2D-COS) combined with deep learning in identifying herbs	Deep learning, residual convolutional neural network	DL model based on digital image processing is more suitable for identification of different habitats and parts of Panax notoginseng	Liu et al. (2022c)
Predict Quality Markers of Atractylodis Rhizoma [Asteraceae; Atractylodis macrocephalae Rhizoma]	Network pharmacology	Identified four active constituents which can be used as Q-markers of Atractylodis Rhizoma	Zhao et al. (2023)
Explore effective mechanism and quality control of Tongsaimai tablet (TSMT) for anti-atherosclerosis benefit	Network pharmacology	Screened out the potential anti-AS mechanisms and chemical quality markers	Cheng et al. (2022b)
Developed a new strategy of multi-dimensional “radar chart” mode to stablish the holistic quality control of Qiliqiangxin Capsule (QLQX)	Network pharmacology	Discovered the Q-markers of QLQX	He et al. (2021)
Develop a processing-associated quality marker (Q-marker) discovery strategy	Systems pharmacology, in vivo high-throughput screening model	Developed a processing-associated Q-marker discovery strategy for carbonized TCM	Gao et al. (2022)
Establish the quality maker evaluation system	Molecular docking	Selected 7 compounds as quality markers of Mume Fructus that could be used for the process quality control	Liu et al. (2022d)
Identify more reasonable markers for quality control of TCM formulas	Network pharmacology analysis, visualization, molecular docking	Successfully established a novel strategy combining intestinal absorption with network pharmacology analysis	Duan et al. (2020)