PI: Andy Li, Dapeng Wu

The goal of this project is to explore potential vulnerabilities in federated learning applications. Federated learning is a new kind of distributed machine learning with decentralized data. There is no need for data sharing for federated learning. Federated learning can distribute trained model to devices, such as IoT devices and mobile devices. However, there are potential vulnerabilities in federated learning. In this project, we first investigate two privacy attacks: membership inference attacks and model extraction attacks and two integrity attacks: adversarial examples and data poisoning attacks. We plan to propose an approach to preserve privacy in federated learning. Also, we are going to propose an approach to defend integrity attacks.