Scale-MIA: A Scalable Model Inversion Attack against Secure Federated Learning via Latent Space Reconstruction

Shanghao Shi, Ning Wang 0022, Yang Xiao 0010, Chaoyu Zhang, Yi Shi 0001, Y. Thomas Hou 0001, Wenjing Lou. Scale-MIA: A Scalable Model Inversion Attack against Secure Federated Learning via Latent Space Reconstruction. In 32nd Annual Network and Distributed System Security Symposium, NDSS 2025, San Diego, California, USA, February 24-28, 2025. The Internet Society, 2025. [doi]

Abstract

Abstract is missing.