Empowering Medical Research with Verified AI Insights.
VeriMed-GenAI is an advanced AI-driven system designed to enhance the reliability of medical research. By leveraging a Multi-Agent Orchestration (using LangGraph), the system processes complex medical queries through a pipeline of specialized agents to minimize AI hallucinations and ensure scientific accuracy.
In the era of Generative AI, "hallucinations" in medical data can be critical. This project implements a Research-Verify-Edit workflow, ensuring that every piece of information is cross-checked by a dedicated Fact-Checker agent before being presented to the user.
The system uses LangGraph to manage the state and transitions between three specialized agents:
- 🔍 Medical Researcher: Scans the medical domain to provide a detailed, evidence-based initial answer.
- 🛡️ Fact-Checker: Critically analyzes the researcher's output, flagging potential inaccuracies, fake drugs, or outdated clinical data.
- 🏆 Final Editor: Synthesizes the verified data and the critique into a polished, professional medical report.
- Framework: LangGraph (Stateful Multi-Agent Orchestration).
- LLM: Google Gemini (Dynamic Model Selection: 1.5 Pro/Flash).
- Interface: Streamlit (Interactive Medical Dashboard).
- Language: Python 3.12+.
- Agentic Reasoning: Unlike standard chatbots, this system "thinks" in steps.
- Hallucination Mitigation: Dedicated node for critical fact-checking.
- Dynamic Model Discovery: Automatically detects and utilizes the best available Gemini model in your region.
- Pharmacogenomics Support: Optimized for complex queries involving drug interactions and genomic markers.
git clone [https://github.com/amiragamalyassin/VeriMed-GenAI.git](https://github.com/amiragamalyassin/VeriMed-GenAI.git)
cd VeriMed-GenAI