您好,欢迎来到华中科技大学自动化学院官网!

联系我们| English

当前位置: 首页 > 最新通知 > 科研 > 正文

学术报告第201744期

发布时间:2017-08-15 编辑:朱平 来源:

报告题目:Information Theoretic and Correntropic Learning

 

报告人:陈霸东 教授(西安交通大学)

报告时间:2017816 16:00

报告地点:南一楼中311

 

Abstract: Information theoretic measures, such as entropy, divergence and mutual information can be used as an efficient optimization cost in machine learning and signal processing since they can capture higher order statistics of the data. Many numerical examples have shown the superior performance of information theoretic learning (ITL). Particularly in recent years, a novel ITL measure, called correntropy, has been successfully applied to robust and sparsity-aware learning. Correntropy is a generalized correlation in high dimensional kernel space, directly related to the probability of how similar two random variables are in a neighborhood of the joint space. The correntropy induced metric (CIM) behaves like different norms (from L2 to L0) in different zones. This talk will give a brief overview of ITL and correntropy. Applications to robust regression, adaptive filtering, principal components analysis, compressive sensing, and deep learning and so on, will also be discussed.        

Biosketch陈霸东,西安交通大学教授,博导,陕西省百人计划特聘教授。2008年毕业于清华大学计算机专业获博士学位,20087月至20109月在清华大学精密仪器与机械学系做博士后研究,201010月至20129月在美国佛罗里达大学 (University of Florida) 电气与计算机工程系做博士后研究 (合作导师Jose C. Principe教授)20157月到8月在新加坡南洋理工大学(NTU)做访问科学家。研究兴趣包括信号处理、机器学习、人工智能、脑机接口等。目前发表学术论文180余篇,其中SCI期刊论文100余篇,发表在IEEE TSP, IEEE TNNLS, IEEE SPL, AUTOMATICA 等著名期刊。撰写学术章节4章,学术专著2部,其中以第一作者撰写的英文专著(Elsevier出版社)被国际计算评论(Computing Reviews)评选为2013Notable BookWeb of Science 中论文被引1000多次,其中4篇论文获“ESI高被引论文。陈教授是IEEE高级会员,担任IEEE TNNLSJournal of the Franklin InstituteEntropy 等著名国际刊物编委或副主编,并作为负责人承担了国家自然科学基金青年、面上、重点和973课题等多项重要科研项目。