par Cao, Zhengjun ;Markowitch, Olivier
Référence IEEE Transactions on Big Data, 7, 6, page (973-974)
Publication Publié, 2021-12-01
Article révisé par les pairs
Résumé : The scheme [IEEE Trans. Big Data, 4 (1), 2018, 26-39] is flawed because the sparsity of coefficient matrix is neglected. In the discussed scenario, we find it is unnecessary to outsource such a problem because the client can solve it locally. Even if the matrix is not sparse, the proposed paradigm which requires a great number of interactions, is rarely adopted because of the possibly incurred massive communication cost.