ritwikkanodia / Privacy-Preserving-Machine-Learning

Maintaining the privacy of local server data in a federated learning framework using differential privacy by TensorFlow Privacy Library.

URECA

Undergraduate Student Research

Dataset

creditcard.csv

The dataset can be found at the following link: https://www.kaggle.com/mlg-ulb/creditcardfraud

Pre-processing

1. Remove rows with NaN values

2. Balance the data set unsing SMOTE:

Federated Learning Framework

Before Adding Privacy:

After Adding Tensorflow Privacy:

Authors

1. Numair Fazili
2. Ritwik Kanodia

Special thanks to Dr. Saurav Sen gupta, our supervisor for the research project.