English

Nonlinear Acceleration of Momentum and Primal-Dual Algorithms

Optimization and Control 2019-10-18 v2

Abstract

We describe convergence acceleration schemes for multistep optimization algorithms. The extrapolated solution is written as a nonlinear average of the iterates produced by the original optimization method. Our analysis does not need the underlying fixed-point operator to be symmetric, hence handles e.g. algorithms with momentum terms such as Nesterov's accelerated method, or primal-dual methods. The weights are computed via a simple linear system and we analyze performance in both online and offline modes. We use Crouzeix's conjecture to show that acceleration performance is controlled by the solution of a Chebyshev problem on the numerical range of a non-symmetric operator modeling the behavior of iterates near the optimum. Numerical experiments are detailed on logistic regression problems.

Keywords

Cite

@article{arxiv.1810.04539,
  title  = {Nonlinear Acceleration of Momentum and Primal-Dual Algorithms},
  author = {Raghu Bollapragada and Damien Scieur and Alexandre d'Aspremont},
  journal= {arXiv preprint arXiv:1810.04539},
  year   = {2019}
}