| Title | Federated Learning with Bayesian Differential Privacy |
| Publication Type | Conference Paper |
| Year of Publication | 2019 |
| Authors | Triastcyn, Aleksei, Faltings, Boi |
| Conference Name | 2019 IEEE International Conference on Big Data (Big Data) |
| Keywords | Bayes methods, Bayesian differential privacy, Bayesian privacy accounting method, composability, Computational modeling, data privacy, Deep Learning, Differential privacy, federated learning, federated setting, formal privacy guarantees, Human Behavior, image classification, image classification tasks, learning (artificial intelligence), machine learning, privacy, privacy accounting, privacy budget, pubcrawl, Resiliency, Scalability, Servers, sharper privacy loss |
| Abstract | We consider the problem of reinforcing federated learning with formal privacy guarantees. We propose to employ Bayesian differential privacy, a relaxation of differential privacy for similarly distributed data, to provide sharper privacy loss bounds. We adapt the Bayesian privacy accounting method to the federated setting and suggest multiple improvements for more efficient privacy budgeting at different levels. Our experiments show significant advantage over the state-of-the-art differential privacy bounds for federated learning on image classification tasks, including a medical application, bringing the privacy budget below e = 1 at the client level, and below e = 0.1 at the instance level. Lower amounts of noise also benefit the model accuracy and reduce the number of communication rounds. |
| DOI | 10.1109/BigData47090.2019.9005465 |
| Citation Key | triastcyn_federated_2019 |
Federated Learning with Bayesian Differential Privacy