报告人:Binyan Jiang
题目:On the sparsity of signals in a random sample 摘要:This article proposes a method of moments technique for estimating the sparsity of signals in a random sample. This involves estimating the largest eigenvalue of a large Hermitian trigonometric matrix under mild conditions. As illustration, the method is applied to two well-known problems. The first focuses on the sparsity of a large covariance matrix and the second investigates the sparsity of a sequence of signals observed with stationary, weakly dependent noise. Simulation shows that the proposed estimators can have significantly smaller mean absolute errors than their main competitors. Some key words: Large covariance matrix; Method of moments; Signal sequence; Sparsity; Trigonometric matrix. 时间:2014年12月22日14:30-15:30
报告人:Yongjia Song 题目:Risk Averse Stochastic Optimization 摘要:In this talk, we will first give an overall introduction to risk averse stochastic optimization, and then discuss some recent progress on chance-constrained stochastic programs. Risk averse stochastic optimization dates back to Markowitz's groundbreaking work on portfolio investment optimization, where risk is addressed in the decision making via a mean-risk objective function. Chance-constrained stochastic program (CCSP) is a convenient risk averse optimization model that controls the probability of bad outcomes. Despite its popularity, CCSP is notoriously challenging to solve because its feasible region is in general non convex. We will focus on integer programming techniques based on various mathematical programming formulations to solve CCSP more efficiently. Numerical examples will be provided to illustrate the effectiveness of these approaches. 时间:2014年12月22日15:30-16:30 报告人:Qiong Zhang 题目:Statistical Designs for Model Assessment 摘要:In this talk, I will present space-filling design approaches to reduce the variability in assessing approximation models for a black-box system. The key of this approach is to generate a structured cross-validation sample such that the input values in each fold achieve uniformity. The advantage of the proposed method will be demonstrated by theoretical and numerical results. 时间:2014年12月22日16:30-17:30 |