Safe AI Systems with Pre-trained Hardcoded responsibilities and Symantic search capabilities to be Used as (Aid-Only) in the Medical Field.
An experimental modular architecture for a medically safe AI system integrating Retrieval-Augmented Generation (RAG), RLHF, explainability modules, and dedicated image models for radiology/histology.
⚠️ WARNING:
This project is highly experimental and should be treated with the utmost caution. The field is rapidly evolving and many critical components and safety steps may still need to be incorporated. Use this as a starting point and update regularly.