Defending Federated Learning: Adaptive Integration of Differential Privacy, SMPC, and Byzantine Robustness

Syed Momin Naqvi, Faisal Iradat

Submission Type: I want my submission to be considered for both oral and poster presentation.

Keywords: federated, learning, smpc, dp, byzantine, aggregation, hybrid, multi-layer, layered, differential, privacy, secure, multi-party, computation, adaptive

TL;DR: Integrate differential privacy, secure multi-party computation, and Byzantine-robust aggregation into a single federated learning framework that adapts security parameters in real time based on threat levels and resource constraints

Confirmation Of Submission Requirements: I submit an abstract. It uses the template provided on the submission page and is no longer than 2 pages.

Submission Number: 236