We propose a differentially private algorithm that approximates the second moment matrix of any subsamplable dataset, and show its applicability to heavy-tailed distributions and input with outliers.

We consider the privacy guarantees of a new sampling method, which combines aspects of Poisson subsamplig and shuffling, and prove it can be approximated as Poisson subsampling.

We propose the vertical federated feature screening (VFS) algorithm, which effectively reduces computational, communication, and encryption costs.

Our paper shows that ensemble learning via majority voting achieves exponentially faster risk decay, improving base learners with slow rates.