R-GAP

Recursive Gradient Attack on Privacy

Description:

We provide a closed-form recursive procedure to recover data from gradients in deep neural networks. We name it Recursive Gradient Attack on Privacy (R-GAP)

More Information:

Originates from the AI-Driven Data Science research challenge (FAIR 2019–2023).
Responsible Research Lead: Matthew Blaschko
Authors: Junyi Zhu
Programming Language: Python
Link to the repository
Categories: Images, Networks and graphs

Publications:

J. Zhu, M. Blaschko, "R-GAP: Recursive Gradient Attack on Privacy", in Proceedings of the International Conference on Learning Representations, 2021.