Stochastic Newton Proximal Extragradient Method

However, these methods typically reach superlinear convergence only when the stochastic Hessian noise diminishes, increasing per-iteration costs over time.