1: \begin{abstract}

2: In a previous paper \cite{bib1}, the convergence performance of the proposed differential privacy algorithm in federated learning (FL), noising before model aggregation FL (NbAFL), was studied. However, the presented convergence upper bound of the NbAFL (Theorem 2) is incorrect.

3: A comment is made here to present the correct form of the convergence upper bound of NbAFL.

4: \end{abstract}

5: