Scientific Program

Plenary Lecture

Program -> Plenary and Semi-Plenary -> Fri.17:00.H 0105 title only | abstract | bio sketch

Friday, 17:00 - 17:50 h, H 0105

Jorge Nocedal
Second-Order Methods for Stochastic, Semi-Smooth and Nonlinear Programming


First-order methods have been advocated for solving optimization problems of large scale. Although they are sometimes the most appropriate techniques, we argue that in many applications it is advantageous to employ second-order information as an integral part of the iteration. This is particularly so when parallel computing environments are available. In this talk, we take a broad view of second-order methods, and center our discussion around three applications: convex L1 regularized optimization, inverse covariance estimation, and nonlinear programming. We note that many efficient methods for these problems can be derived using a semi-smooth Newton framework, which allows us to compare their manifold identification and subspace minimization properties. We propose an algorithm employing a novel active-set mechanism that is of interest in machine learning, PDE-constrained optimization, and other applications. We also discuss dynamic sampling techniques, illustrate their practical performance, and provide work complexity bounds. The talk concludes with some observations about the influence that parallel computing has on large scale optimization calculations.



Program -> Plenary and Semi-Plenary -> Fri.17:00.H 0105 title only | abstract | bio sketch

  There are three major facts that should be watched out for in all payday loans in the United States. The new drug with unique properties was developed to help men to get rid of all sexual disorders, and its name is Cialis Super Force. Now you do not have to buy two different medications to solve sexual problems.