1 paper across 1 session
A parameter-free regularized Newton-type algorithm that simultaneously achieves optimal global complexity and quadratic local convergence in nonconvex optimization.
