This repository contains code for the CAIAiPCP project, which focuses on using contestable large language models (LLMs) to generate care plans for aging-in-place scenarios. The project aims to enhance the quality and transparency of care plans through interactive and contestable AI systems.
To set up the project, follow these steps:
-
Install Ollama To support local model hosting, install Ollama on your machine. Once installed, run and serve the
gemmamodel:ollama run gemma4:latest
The application communicates with Ollama locally via
http://localhost:11434. -
Create your virtual environment and install the required Python packages:
pip install -r requirements.txt
-
Navigate to the frontend directory:
cd frontend -
Install the required npm packages:
npm install
-
Start the development server:
npm run dev
- "Tell" the development server to “open in browser” (e.g., type : o in the development server terminal window).
- This should open your default browser and open the frontend graphical user interface.
Note
The backend server may need to be running on http://localhost:8001 for the frontend to communicate with it.
- Setup the vector database with MedicalRAG dataset:
python rag/indexer.py- Start the application:
python main.py
If you use or reference this work in a scientific publication, we would appreciate that you use the following citations:
@article{nguyen2026position,
title={Position: Multi-Agent Algorithmic Care Systems Demand Contestability for Trustworthy AI},
author={Nguyen, Truong Thanh Hung and Fournier, H{\'e}l{\`e}ne and Jackson, Piper and Itoh, Makoto and Freeman, Shannon and Richard, Rene and Cao, Hung},
journal={arXiv preprint arXiv:2603.20595},
year={2026}
}
- After human review, the status is not updated.
- Streaming LLM is not real-time.
