MediPedia: AI-Powered Medical Knowledge Assistant

MediPedia is an AI-powered application that provides accurate medical information based on trusted medical resources. It uses Retrieval Augmented Generation (RAG) to answer medical questions by searching through a database of medical knowledge.

Features

• Medical Q&A: Ask any medical question and get accurate information sourced from medical literature
• Source Attribution: Responses include references to the source documents and page numbers
• Chat Interface: User-friendly chat interface built with Streamlit
• Secure Knowledge Base: Information is retrieved from trusted medical resources

Technology Stack

• Frontend: Streamlit for the interactive web interface
• NLP & ML:
  • LangChain for orchestrating LLM workflows
  • Hugging Face for embeddings and the LLM model (Mistral-7B)
  • FAISS for efficient vector search
• Document Processing:
  • LangChain for document loading and text splitting
  • PDF processing for medical documents

Project Structure

medipedia/
├── data/                   # Medical PDF documents
│   └── GaleEncyclopediaOfMedicine.pdf
├── vector_store/           # Vector embeddings database
│   └── db_faiss/           # FAISS vector store
├── create_memory_for_llm.py   # Script to create embeddings from PDFs
├── connect_memory_with_llm.py # Script to test LLM with vector store
├── medipedia.py            # Main Streamlit application
├── requirements.txt        # Dependencies
└── README.md               # Project documentation

How It Works

1. Data Ingestion: Medical PDFs are loaded and processed into chunks
2. Embedding Generation: Text chunks are converted into vector embeddings
3. Knowledge Storage: Embeddings are stored in a FAISS vector database
4. User Query: User enters a medical question through the Streamlit interface
5. Semantic Search: The system finds relevant information from the knowledge base
6. Response Generation: The LLM generates a comprehensive answer based on retrieved context
7. Source Attribution: References to source documents are provided for transparency

Installation

1. Clone the repository:

git clone https://github.com/yourusername/medipedia.git
cd medipedia

2. Create a virtual environment and activate it:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Set up your Hugging Face API token: Create a .env file in the project root with your Hugging Face API token:

HF_TOKEN=your_huggingface_token_here

Usage

Creating the Knowledge Base

1. Place your medical PDFs in the data/ directory.

2. Generate the vector database:

python create_memory_for_llm.py

Running the Application

Start the Streamlit application:

streamlit run medipedia.py

The application will be available at http://localhost:8501

Future Improvements

• Add more medical resources to expand the knowledge base
• Implement multi-modal support for medical images
• Add user authentication for personalized medical information
• Implement search history and favorite responses
• Optimize for mobile devices

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

• The Gale Encyclopedia of Medicine for the medical knowledge base
• Hugging Face for providing access to state-of-the-art NLP models
• LangChain for the powerful RAG framework

About the Author

This project was created as a demonstration of building AI-powered knowledge systems using retrieval augmented generation techniques. It showcases skills in natural language processing, vector embeddings, and building interactive AI applications.

---

Note: This application is for educational purposes only and should not be used as a substitute for professional medical advice, diagnosis, or treatment.