Differentially Private Synthetic Tabular Data Generation

This project generates synthetic tabular data using deep learning models with Differential Privacy.

---

Models

• DP-CTGAN
• DP-DCF Diffusion

---

Project Structure

agents/
backend/
frontend/
models/
training/
data/
results/
requirements.txt

---

Requirements

• Python 3.9+
• PyTorch
• FastAPI
• Uvicorn
• NumPy
• Pandas
• Scikit-learn
• Opacus

---

Setup

python -m venv .venv
.venv\Scripts\activate (Windows)
source .venv/bin/activate (Linux/Mac)
pip install -r requirements.txt

---

Run Backend

uvicorn backend.api:app --reload --host 0.0.0.0 --port 8000
Open: http://127.0.0.1:8000

---

Run Frontend

Open frontend/index.html in browser

---

Train Models

python training/dp_ctgan_train.py
python training/dp_dcf_diffusion_train.py

---

Generate Data

Use frontend or API
Download CSV output

---

Evaluate

python training/evaluation.py

---

Author

Deval Jadav