报告题目：Several Augmented Lagrangian-Based Methods

报告时间：2022年11月11日（周五）下午13:30-15:30

报告方式：腾讯会议（会议号：873-653-272）

主讲人：白建超 副教授

摘要：In this talk, a new exact augmented Lagrangian method and an inexact version will be presented for solving a family of convex optimization problems subject to equality or inequality constraint. We show their global convergence and sublinear convergence rate for the general convex case. Performance of the proposed methods are verified by testing the popular support vector machine problem and some image processing problems. In addition, an inexact alternating direction method of multipliers is also discussed for solving the separable nonconvex and nonsmooth minimization problem.

白建超，西北工业大学数学与统计学院副教授，硕士生导师。主要从事机器学习和统计学习等领域的大规模优化方法、理论与应用研究；在CSIAM Transactions on Applied Mathematics、IEEE Transactions on Medical Imaging、Computational Optimization and Applications、Numerical Linear Algebra with Applications等期刊上发表学术论文20余篇；主研科技部重点研发计划项目、国家自然科学基金青年项目、中国博士后科学基金项目等；担任CSIAM大数据与人工智能专委会委员、陕西省运筹学会理事、西安市科技计划项目会评专家。