federated learning Naive Bayes ？

[bultiful]

Hi, If you don't use differential privacy, could Naive Bayes be combined with federated learning? What are the parameters shared with the aggregator?
and differential privacy is "local differential privacy"?

thank u very much