Convergence Analysis of Proximal Gradient with Momentum for Nonconvex Optimization
This provides theoretical guarantees for efficient optimization in nonconvex machine learning applications, though it is incremental as it builds on existing proximal gradient methods.
The paper tackles nonconvex optimization problems by proposing an accelerated proximal gradient method (APGnc) that ensures monotonic decrease and proves convergence to critical points, with linear and sub-linear rates under the Kurdyka-Łojasiewicz property, and extends it to stochastic and inexact versions.
In many modern machine learning applications, structures of underlying mathematical models often yield nonconvex optimization problems. Due to the intractability of nonconvexity, there is a rising need to develop efficient methods for solving general nonconvex problems with certain performance guarantee. In this work, we investigate the accelerated proximal gradient method for nonconvex programming (APGnc). The method compares between a usual proximal gradient step and a linear extrapolation step, and accepts the one that has a lower function value to achieve a monotonic decrease. In specific, under a general nonsmooth and nonconvex setting, we provide a rigorous argument to show that the limit points of the sequence generated by APGnc are critical points of the objective function. Then, by exploiting the Kurdyka-Łojasiewicz (\KL) property for a broad class of functions, we establish the linear and sub-linear convergence rates of the function value sequence generated by APGnc. We further propose a stochastic variance reduced APGnc (SVRG-APGnc), and establish its linear convergence under a special case of the \KL property. We also extend the analysis to the inexact version of these methods and develop an adaptive momentum strategy that improves the numerical performance.