Matrix Sketching for Secure Collaborative Machine Learning

                        Mengjiao Zhang 1 Shusen Wang 1

              Abstract                clients, each client uses a batch of local data to compute a
                                 stochastic gradient, and the server aggregates the stochas-
    Collaborative learning allows participants to
                                 tic gradients and updates the model parameters. Based on
    jointly train a model without data sharing. To
  ...