AttentionRX: Reflection Type Agent for Medical Symptom Identification Program 💊🩺(UNDER DEVELOPEMENT)
AttentionRX is an innovative software solution designed to enhance the analysis and interpretation of medical patient records by cross-referencing them with scholarly journal articles. By leveraging the latest advancements in artificial intelligence, AttentionX identifies symptoms from patient records and provides evidence-based prescription suggestions. The core technology stack includes Retrieval Augmented Generation (RAG), and Reflection type agents powered by Langchain, Llama3-OpenBioLLM-70B, Qdrant, DSPy and Langsmith, facilitating a robust and insightful analysis.
- Symptom Identification: Automated identification of symptoms and getting information about microbe based diseases from patient records using advanced LLMs.
- Scholarly Journal Integration: Cross-referencing symptoms with the latest scholarly articles and research for evidence-based diagnosis and prescription.
- Evidence-Based Prescriptions: Utilizes cutting-edge AI to suggest prescriptions based on the most current research and data.
- Advanced Tech Stack: Incorporates Retrieval Augmented Generation (RAG), Langchain, Llama3-OpenBioLLM-70B, Qdrant, DSPy and Langsmith for comprehensive data analysis and retrieval as well as for optimization and evaluation.
| Feature | Tech Stack |
|---|---|
| Data Collection Tools | Arxiv, Scholar, Tavily |
| VectorDB, RAG | Qdrant, Cohore Reranker |
| System Building | Langchain, Langgraph |
| Fine-tuning | RAFT + QLoRa |
| Optimization | DSPy |
| Evaluation | Langsmith, DSPy |
| Deployement | Azure, Chainlit |
data/: Contains the dataset of medical patient records and scholarly journal articlesReflection_Agents.ipynb: Jupyter notebook containing samples and demonstrations of the project.
- Clone the repository:
git clone https://github.com/yourusername/AttentionX.git - Navigate to the project directory:
cd AttentionX - Install the required packages:
pip install -r requirements.txt
For samples and demonstrations, open the notebooks folder.
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.