Id: arxiv/2607.10741 Url: https://arxiv.org/abs/2607.10741 Tags: math.OC Score: 0 Title: Parameter-Free Cubic-Regularized Newton Method: Sharp Complexity and Generalized Smoothness Domain: math.OC Source: arxiv Authors: Shaoying Fang, Naoki Marumo, Akiko Takeda Reviews: 0 Subject: math.OC Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-07-12 Page: https://delphi.science/r/arxiv/2607.10741 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 parameters. Under the generalized smoothness condition $\|\nabla^3 f(x)\| \leq L_0 + L_1 \|\nabla f(x)\|$, we derive an oracle complexity bound for finding an $(\varepsilon, δ)$-second-order stationary point. This assumption is weaker than the generalized smoothness conditions used in existing analyses of second-order methods, while the complexity bound improves upon existing guarantees for parameter-free…