太阳成集团tyc7111cc学术报告
Exact and Inexact Augmented Lagrangian Methods
白建超
(西北工业大学)
报告时间:2023年5月23日 星期二 16:00-17:00
报告地点:沙河校区E405
报告摘要:
In this talk, an exact augmented Lagrangian method (ALM) and an inexact ALM will be presented for solving a family of linearly constrained convex optimization problems. By the variational analysis, the generated iterates are characterized as a variational inequality. Then, we establish the global convergence and sublinear convergence rate of these algorithms. Performance of the algorithms are tested on solving the support vector machine problem and two image processing problems.
报告人简介:白建超,男,博士 (后),西北工业大学数学与统计学院副教授,硕士生导师。他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等期刊上发表论文30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和陕西省科协科技人才奖励评审专家。
邀请人: 谢家新