« search calendars« DIMACS/TRIPODS Workshop on Optimization and Machine Learning

« Direct Runge-Kutta Discretization Achieves Acceleration

Direct Runge-Kutta Discretization Achieves Acceleration

August 14, 2018, 12:00 PM - 12:30 PM

Location:

Iacocca Hall

Lehigh University

Bethlehem PA

Click here for map.

Aryan Mokhtari, Massachusetts Institute of Technology

In this talk, we study gradient-based optimization methods obtained by directly discretizing a second-order ordinary differential equation (ODE) related to the continuous limit of Nesterov’s accelerated gradient method. When the function is smooth enough, we show that acceleration can be achieved by a stable discretization of this ODE using standard Runge-Kutta integrators. Specifically, we prove that under Lipschitz-gradient, convexity and order-($s+2$) differentiability assumptions, the sequence of iterates generated by discretizing the proposed second-order ODE converges to the optimal solution at a rate of $O(N^{-2s/(s+1)})$, where s is the order of the Runge-Kutta numerical integrator. Furthermore, we introduce a new local flatness condition on the objective, under which rates even faster than $O(N^{-2})$ can be achieved with low-order integrators and only gradient information. Notably, this flatness condition is satisfied by several standard loss functions used in machine learning.