Muddle Your Data
About the Project
The goal of this project is to study how users can defend their privacy against profiling attempts from different service providers like search engines, online social networks, etc, by obfuscating their own data. In particular, the research focuses on the investigation of techniques, like differential privacy, based on the idea to let the user artificially decrease the signal-to-noise ratio of the information he/she discloses.- Principal Investigator
-
- Giovanni Neglia, Inria
- Project's partner(s)
-
- Charles Bouveyron, LJAD
- Michela Chessa, GREDEG
- Bruno Ribeiro, Deparment of Computer Science, Purdue University, Indiana, USA
- Duration
-
- September 2018 - November 2020
- Total Amount
-
- 27 720 euros
- Publications
-
- Michela Chessa, “A Shapley-based Groves mechanism: When the mechanism designer plays the wise man,” Operations Research Letters, Vol. 47(6), 2019
https://www.sciencedirect.com/science/article/abs/pii/S0167637719302469 - Chuan Xu, Giovanni Neglia. What else is leaked when eavesdropping Federated Learning?. CCS workshop Privacy Preserving Machine Learning (PPML), Nov 2021, Online. ⟨10.1145/1122445.1122456⟩. ⟨hal-03364766v2⟩
- Oualid Zari, Chuan Xu, and Giovanni Neglia. “Efficient passive membership inference attack in federated learning”. In: NeurIPS workshop on Privacy in Machine Learning (PriML). Virtual., Dec. 2021.
https://hal.archives-ouvertes.fr/hal-03410152/
- Michela Chessa, “A Shapley-based Groves mechanism: When the mechanism designer plays the wise man,” Operations Research Letters, Vol. 47(6), 2019
- Related Documentation
-
- Project MYDATA presentation (Day of projects restitution of the Academy “Networks, Information and Digital Society", 14/03/2022)