Herbal knowledge and artificial intelligence may unlock smarter disease inference
Herbal knowledge and artificial intelligence may unlock smarter disease inference
https://www.eurekalert.org/news-releases/1129291
Publish Date: 2026-05-22 10:23:00
Source Domain: www.eurekalert.org
- MediHerb is an innovative multi-modal artificial intelligence framework developed to advance disease inference by integrating Traditional Chinese Medicine (TCM) knowledge.
- The framework combines five complementary datasets: molecular sequences, chemical fingerprints, physicochemical properties, graphical prescription representations, and textual TCM prescription descriptions for richer insights.
- Using an attention-based fusion mechanism, MediHerb aligns biological, herbal, and diagnostic information in a shared latent space, enabling multi-granular reasoning across different biomedical layers.
- Experimental results show MediHerb’s superior performance over existing methods in herb-disease association prediction, along with biologically meaningful embedding patterns that enhance interpretability.
- The framework provides interpretable embeddings and attention mechanisms, explaining the pharmacological basis of herb-disease relationships which supports future mechanistic studies in TCM research.
- To increase accessibility, the researchers developed a graphical user interface for easier interaction with the model and datasets, promoting broader adoption among researchers and clinicians.
- MediHerb aims to bridge molecular biology, herbal medicine, and clinical semantics by integrating diverse biomedical signals into one holistic framework.
- The study underscores the growing importance of AI-driven multi-modal data integration in modern biomedical research, particularly in modernizing Traditional Chinese Medicine and developing intelligent healthcare systems.
