TY  - JOUR
T1  - Differential Privacy for Multiple Antenna Wireless Communication Assisted Federated Learning
AU  - Park, Junguk 
AU  - Yun, Sangseok 
JO  - The Journal of Korean Institute of Communications and Information Sciences
PY  - 2024
DA  - 2024/1/1
DO  - 10.7840/kics.2024.49.2.308
KW  - Dirferential privacy
KW  - federated learning
KW  - multiple antenna
KW  - wireless communication
AB  - In this work, we study differential privacy (DP) for federated learning over multiple antenna wireless communications. Specifically, we revisit DP from the perspective of wireless communications and propose an efficient beamforming scheme that can reduce inevitable degradation of inference accuracy due to intentional noises from DP. We also propose a power allocation scheme guaranteeing DP for the proposed beamforming scheme. Simulation results demonstrate that the proposed scheme can effectively prevent inference accuracy loss caused by DP.