This research repurposes specialized neurosurgical publications into valuable clinical and educational tools using vision-language models (VLMs).

• Created NeuroPubs, a specialized database of neurosurgical literature
• Developed AI system generating high-quality board exam questions for neurosurgeons
• Built an AI diagnostic copilot for clinical neurosurgical service
• Combined specialized scientific knowledge with general-purpose AI models

This innovation bridges the gap between academic publications and practical medical applications, making specialized knowledge more accessible to practitioners while maintaining clinical accuracy and relevance.

Repurposing the scientific literature with vision-language models