README.md

metadata

title: AI Biomedical Assistant
emoji: 🧬
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 6.14.0
python_version: '3.13'
app_file: app.py
pinned: false
license: mit

🧬 AI Biomedical Assistant using LLMs and RAG

📌 Project Overview

AI Biomedical Assistant is a Retrieval-Augmented Generation (RAG)-based biomedical intelligence system developed for biomedical symptom analysis and drug recommendation.

The system combines:

• Large Language Models (LLMs)
• PubMed biomedical retrieval
• Semantic search
• Audio symptom transcription
• Molecular visualization

to generate evidence-grounded biomedical reports.

---

🚀 Key Features

• ✅ Biomedical symptom analysis
• ✅ Retrieval-Augmented Generation (RAG)
• ✅ PubMed biomedical literature retrieval
• ✅ Audio + text symptom input
• ✅ Drug recommendation generation
• ✅ 2D molecular structure visualization
• ✅ Animated 3D molecular visualization
• ✅ Follow-up clinical reasoning
• ✅ Multi-domain biomedical support

---

🧠 Technologies Used

• Gradio
• BioGPT / Transformers
• Sentence Transformers
• FAISS Vector Search
• PubMed API
• RDKit
• 3Dmol.js
• Whisper Speech Recognition

---

⚙️ System Workflow

1. User enters symptoms through text or audio
2. Audio is transcribed using Whisper
3. PubMed biomedical papers are retrieved
4. RAG retrieves relevant biomedical context
5. LLM generates biomedical clinical report
6. Drug recommendation is generated
7. 2D and 3D molecular structures are visualized

---

📄 Biomedical Report Includes

• Possible Conditions
• Biological Causes
• Recommended Medications
• Common Side Effects
• Urgency Assessment
• Recommended Next Steps
• Follow-Up Questions

---

🧪 Example Symptoms Supported

• Fever
• Cough
• Chest Pain
• Nausea
• Stomach Pain
• Headache
• Skin Irritation
• Allergy
• Breathing Difficulty
• Vomiting
• Migraine
• Fatigue
• Joint Pain
• And many more...

---

🎓 Academic Context

Developed as part of:

CS3235 — Working with Large Language Models

The project demonstrates:

• Retrieval-Augmented Generation
• Biomedical semantic retrieval
• Multimodal AI pipelines
• Molecular visualization
• LLM-powered biomedical reasoning

---

👩‍💻 Author

Anagha Nagesh