1 paper
Shaoying Fang, Naoki Marumo, Akiko Takeda
We analyze a variant of the cubic-regularized Newton method for nonconvex optimization. This variant is parameter-free in that it requires no prior knowledge of problem-dependent p…