作者位次 #4/9ACL 2026Poster· Natural Language Processing
On the (In-)Security of the Shuffling Defense in the Transformer Secure Inference
Zhengyi Li · Yakai Wang · Jingwen Leng · Kang Yang · Yu Yu · Jiaping Gui · Yu Feng · Ning Liu · Minyi Guo
摘要
For Transformer models, cryptographically secure inference ensures that the client learns only the final output, while the server learns nothing about the client’s input. However, securely computing nonlinear layers remains a major efficiency bottleneck due to the substantial communication rounds...