Foundation models are revolutionizing healthcare by leveraging large-scale pre-trained AI architectures that adapt to diverse biomedical applications.

• Pre-trained on vast unlabeled datasets using unsupervised methods
• Outperform traditional task-specific AI models across medical applications
• Enable advances in clinical informatics, medical imaging, and public health
• Support computational biology and complex biological analysis

These versatile models are fundamentally changing medical research and practice by providing powerful tools for diagnosis, treatment planning, and biomedical discovery that were previously impossible with conventional AI approaches.

Biomedical Foundation Model: A Survey