Privacy-preserving Deep Learning models

Primary supervisor

Chunyang Chen

Modern machine learning is increasingly applied to create amazing new technologies and user experiences, many of which involve training machines to learn responsibly from sensitive data, such as personal photos or email. Ideally, the parameters of trained machine-learning models should encode general patterns rather than facts about specific training examples.

Aim/outline

To ensure the privacy, we are working on a series of works including developing privacy preserving deep learning models, detecting potential privacy leaking, fixing privacy issues with deep learning models. In this work, we are specifically targeting at the security and privacy on-device deep learning models in Android Apps.

Required knowledge

  • Deep Learning
  • Adversarial Attack
  • Android App Development