Implicit Runge-Kutta Methods for Accelerated Unconstrained Convex Optimization
Implicit Runge-Kutta Methods for Accelerated Unconstrained Convex Optimization
Accelerated gradient methods have the potential of achieving optimal convergence rates and have successfully been used in many practical applications.Despite this fact, the rationale underlying these accelerated methods remain elusive.In this work, we study gradient-based accelerated optimization methods obtained by directly discretizing a second-order ordinary differential equation (ODE) related to …