1: \begin{abstract}

2: This note considers the inexact cubic-regularized Newton's method (CR), which has been shown in \cite{Cartis2011a} to achieve the same order-level convergence rate to a secondary stationary point as the exact CR \citep{Nesterov2006}. However, the inexactness condition in \cite{Cartis2011a} is not implementable due to its dependence on future iterates variable. This note fixes such an issue by proving the same convergence rate for nonconvex optimization under an inexact adaptive condition that depends on only the current iterate. Our proof controls the sufficient decrease of the function value over the total iterations rather than each iteration as used in the previous studies, which can be of independent interest in other contexts. 

3: % is a popular second-order method that has guaranteed convergence to a secondary stationary point. Inexact CR method, which approximates the Hessian in CR, has been shown to achieve the same convergence rate as long as an inexact condition is satisfies. However, such However, the high computation complexity of Hessian makes the CR impractical for solving large finite-sum problems. To overcome this issue, inexact Hessian techniques have been proposed in several variants of CR method, in which the inexact condition either is adapted to unknown future information or is too conservative. In this paper, we proposed a new inexact condition for evaluating the Hessian in CR that is adapted to past information and is much less conservative. Under the proposed inexact condition, we show that such inexact CR achieves the same order of convergence rate to a second-order stationary point as that of vanilla CR in nonconvex optimization, and the inexact condition further saves much computation in practical scenarios. 

4: \end{abstract}

5: