Empirically, we conduct extensive experiments to verify our theories under a range of poisoning attacks on different datasets. We find that increasing the level of privacy protection in DPFL results in stronger certified attack inefficacy; however, it does not necessarily lead to a stronger ...