This paper gives a unified and succinct approach to the $O(1/\sqrt{k}), O(1/k),$ and $O(1/k^2)$ convergence rates of the subgradient, gradient, and accelerated gradient methods for unconstrained convex minimization. In the three cases the proof of convergence follows from a generic bound defined by the convex conjugate of the objective function. Citation Working Paper, Carnegie Mellon … Read more
This paper describes an implementation of the L-BFGS method designed to deal with two adversarial situations. The first occurs in distributed computing environments where some of the computational nodes devoted to the evaluation of the function and gradient are unable to return results on time. A similar challenge occurs in a multi-batch approach in which … Read more
For a type of multi-block separable convex programming raised in machine learning and statistical inference, we propose a proximal alternating direction method of multiplier (PADMM) with partially parallel splitting, which has the following nice properties: (1) To alleviate the weight of the proximal terms, the restrictions imposed on the proximal parameters are relaxed substantively; (2) … Read more
Recently, a generalized version of Peaceman-Rachford splitting method (GPRSM) for solving a convex minmization model with a general separable structure has been proposed by \textbf{Sun} et al and its global convergence has been proved. In this paper, we further study theoretical aspects of the generalized Peaceman-Rachford splitting method. We first establish the worst-case $\mathcal{O}(1/t)$ convergence … Read more
We introduce an extension of Dual Dynamic Programming (DDP) to solve convex nonlinear dynamic programming equations. We call Inexact DDP (IDDP) this extension which applies to situations where some or all primal and dual subproblems to be solved along the iterations of the method are solved with a bounded error. We show that any accumulation … Read more
Dearing and Zeck presented a dual algorithm for the problem of the minimum covering ball in $\mathbb{R}^n$. Each iteration of their algorithm has a computational complexity of at least $\mathcal O(n^3)$. In this paper we propose a modification to their algorithm that, together with an implementation that uses updates to the QR factorization of a … Read more
In the first part of the paper we focus on two problems: (a) regularized least squares and (b) nonsmooth minimization over an affine subspace. For these problems we establish the connection between the primal-dual method of Chambolle-Pock and Tseng’s proximal gradient method. For problem (a) it allows us to derive a nonergodic $O(1/k^2)$ convergence rate … Read more
The alternating direction method of multipliers (ADMM) was proposed by Glowinski and Marrocco in 1975; and it has been widely used in a broad spectrum of areas, especially in some sparsity-driven application domains. In 1982, Fortin and Glowinski suggested to enlarge the range of the step size for updating the dual variable in ADMM from … Read more
Designing distributed algorithms for empirical risk minimization (ERM) has become an active research topic in recent years because of the practical need to deal with the huge volume of data. In this paper, we propose a general framework for training an ERM model via solving its dual problem in parallel over multiple machines. Our method … Read more
The alternating direction method of multipliers is a powerful operator splitting technique for solving structured optimization problems. For convex optimization problems, it is well-known that the algorithm generates iterates that converge to a solution, provided that it exists. If a solution does not exist, then the iterates diverge. Nevertheless, we show that they yield conclusive … Read more