LLM-AMT system integrates authoritative medical textbooks with general-purpose language models to dramatically improve biomedical question answering capabilities.

• Creates plug-and-play knowledge augmentation without requiring model retraining
• Leverages retrieval-augmented generation to provide context-specific medical information
• Achieves performance comparable to specialized medical AI models while using general-purpose LLMs
• Demonstrates improved factual accuracy and reliability for healthcare applications

This research bridges the gap between general AI and specialized medical knowledge, enabling more reliable AI-assisted healthcare solutions without building domain-specific models from scratch.

Augmenting Black-box LLMs with Medical Textbooks for Biomedical Question Answering