scholarly journals Privacy Amplification for Federated Learning via User Sampling and Wireless Aggregation

Author(s):  
Mohamed Seif ◽  
Wei-Ting Chang ◽  
Ravi Tandon
2021 ◽  
Vol 21 (3&4) ◽  
pp. 0181-0202
Author(s):  
Khodakhast Bibak ◽  
Robert Ritchie ◽  
Behrouz Zolfaghari

Quantum key distribution (QKD) offers a very strong property called everlasting security, which says if authentication is unbroken during the execution of QKD, the generated key remains information-theoretically secure indefinitely. For this purpose, we propose the use of certain universal hashing based MACs for use in QKD, which are fast, very efficient with key material, and are shown to be highly secure. Universal hash functions are ubiquitous in computer science with many applications ranging from quantum key distribution and information security to data structures and parallel computing. In QKD, they are used at least for authentication, error correction, and privacy amplification. Using results from Cohen [Duke Math. J., 1954], we also construct some new families of $\varepsilon$-almost-$\Delta$-universal hash function families which have much better collision bounds than the well-known Polynomial Hash. Then we propose a general method for converting any such family to an $\varepsilon$-almost-strongly universal hash function family, which makes them useful in a wide range of applications, including authentication in QKD.


2017 ◽  
Vol 37 (2) ◽  
pp. 0227002
Author(s):  
刘翼鹏 Liu Yipeng ◽  
郭建胜 Guo Jiansheng ◽  
崔竞一 Cui Jingyi

Sign in / Sign up

Export Citation Format

Share Document