A pioneering approach to developing trustworthy AI has emerged, harnessing the power of Knowledge Graphs as the foundation of truth and leveraging RAG audits to ensure answer accuracy. This innovative method has already shown promising results in the medical field, enabling the creation of an AI capable of clinical diagnosis and supporting medical students in their training.

The AI model is built upon a comprehensive Knowledge Graph, comprising 5,000 nodes representing medical terms and 25,000 relationships, allowing it to provide verifiable answers through rigorous RAG audits of the Knowledge Graph. This cutting-edge technology holds immense potential for application in various specialized domains of human knowledge, revolutionizing the way we interact with AI.

A test model is currently available for evaluation at https://huggingface.co/spaces/cmtopbas/medical-slm-testing, boasting response times of under 3 seconds on a dedicated GPU. The developer is actively seeking partnerships with medical schools and clinics for complimentary test runs, as well as co-founders with a medical background and marketing expertise to further propel this groundbreaking initiative.

Photo by Thirdman on Pexels
Photos provided by Pexels