What role does the Hessian matrix play in second$-$order optimization methods for machine learning? $q,$

It determines the learning rate adaptively

It provides curvature information to adjust the learning step size more effectively

It reduces the overall computational time

It simplifies the optimization process to linear components