Image Home
Image People
Image Publication
Image Applications
Image Data & Code
Image Library
Image Seminar
Image Link
Image Album



Search LAMDA
»

Seminar abstract

Using Relations among Attributes/Classes to Boost Learning Performance

Songcan Chen
Professor
Nanjing University of Aeronautics and Astronautics

Abstract: In machine learning community, using as much (relation) information among attributes/classes as possible is a line to boost learning performance, in this talk, we learn or/and formulate relations respectively for the attribute-attribute or attribute-class in zero-shot learning and the class-class in multiclass/multilabel learning, and then exploit such relationship to boost the performance of attribute/class predictors. Experimental results on a series of real benchmark data sets demonstrate the efficacy of the proposed methods.

Bio: 陈松灿,南京航空航天大学教授、博导。独立主持9项国家自然科学基金,12项省部级基金并参与1项国家自然科学重点基金等项目。在包括IEEE Transactions等在内的国际主流学术期刊上已发表130多篇SCIE论文,其中3篇发表在国际权威期刊《Pattern Recognition》上的论文获2年一评的年度最佳论文提名奖(Best Paper Awards: Honorable Mentions)。1篇《计算机学报》论文2015在合肥计算机大会上获颁2010-2014年5年度的3篇优秀论文奖之一。所发论文据Google Scholar统计,已被引7900多次,H-指数为41。2014和2015连续2年入选Elsevier中国高引学者榜。2011年作为南京大学的合作者获教育部自然科学1等奖1项,并进而获2013年国家自然科学2等奖,排名第2。已培养毕业博士生35位,有5位获江苏省优博,2位进一步获全国百篇优博论文提名奖。
  Name Size

Image
PoweredBy
(for FireFox 3+ and IE 7+)
Contact LAMDA: (email) contact@lamda.nju.edu.cn (tel) +86-25-89685926. © LAMDA, 2016