We studied the subsampled Newton (SSN) method and its variants, GIANT and SSPN, and established stronger convergence guarantees than the prior works. In particular, we showed that a sample size of s=~Θ(dγeff) suffices, where γ is the ℓ2 regularizati…
We now return to the task of computing the residue in Equation (5.3). We begin by reinterpreting the residue in terms of a non-equivariant limit of twisted invariants.
There aren't more papers