Motivation:
When you want to build a medical agent system,
do you know is there any existing relevant tool or model that you can directly use?
are these models reliable or ready to use?
do you need lots of efforts to build a framework to incorparate them?
What this system would be look like:
Model Market: Collections of public medical models which are ready to use.
Agent Support: Easy to load and offload models (agents).
LLM & MLLM: Support both public or self-built Large Language Model.
Why should we do this?
This is a obstacle hinder in development medical agent system now. Because most users have on idea about which model is ready to use in medicine.
Members of our lab work on almost every area in medical imaging. We can easily collect models in different domain.
The difficulties lie in:
Collection of the models and make them ready to use.
Developing a robust and scalable framework.