Print Email Facebook Twitter Improving privacy of Federated Learning Generative Adversarial Networks using Intel SGX Title Improving privacy of Federated Learning Generative Adversarial Networks using Intel SGX Author Jehee, Wouter (TU Delft Electrical Engineering, Mathematics and Computer Science) Contributor Liang, K. (mentor) Urbano, Julián (graduation committee) Wang, R. (mentor) Degree granting institution Delft University of Technology Programme Computer Science and Engineering Project CSE3000 Research Project Date 2022-06-22 Abstract Federated learning (FL), although a major privacy improvement over centralized learning, is still vulnerable to privacy leaks. The research presented in this paper provides an analysis of the threats to FL Generative Adversarial Networks. Furthermore, an implementation is provided to better protect the data of the participants with Trusted Execution Environments (TEEs), using Intel Software Guard Extensions. Lastly, the viability of it’s use in practice is evaluated and discussed. The results indicate that this approach protects the data, while not affecting the predicting capabilities of the model, with a noticeable but manageable impact on the training duration. Subject Federated learningPrivacy-preservingIntel SGXTrusted Execution EnvironmentsGenerative Adversarial Networks To reference this document use: http://resolver.tudelft.nl/uuid:d30df7e9-a945-4c29-adb3-2273430a58a2 Part of collection Student theses Document type bachelor thesis Rights © 2022 Wouter Jehee Files PDF final_paper_cse3000.pdf 361.43 KB Close viewer /islandora/object/uuid:d30df7e9-a945-4c29-adb3-2273430a58a2/datastream/OBJ/view