新世纪优秀人才校内答辩.ppt

面交通大浮 Xian Jiaotong University Mining Preorder Relation between Knowledge Units from Text JunLiu(liukeen@mail.xjtu.edu.cn Lu Jiang, Zhaohui wu Qinghua Zheng, Yanan Qian March 10. 2011 SAC 2010, Sierre, Switzerland  Outline e Motivation and Challenges e Two features of the preorder relation .. Process of Mining Preorder Relations Experimental evaluation ● Conclusions 步交通大浮 XIan Jiaotong University  Motivation earning is an incremental process. To understand a new knowledge unit often relies on the understanding of certain existing knowledge units Preorder relations among the knowledge units help th antenor Angles a striThebrems disorientation problem In learning ■ Manua| pefinitiomnotating the p自正 le preorder relations is very time consuming, andr requires the annotat efinition ofe domain experts Exterior Angle P Preorder Relation 》步通大浮 XIan Jiaotong University  Problem Definition Given a text document set T=t dsksm and a knowledge unit set U={u1}(l≤i≤n) extracted from T as the input, the preorder relation mining process will output a set ACUXU Each u, can be further represented as a triplet of (name, type, content) Name: such as definition of subnet mask v Type: such as definition, property or method Y Content: the text content of the knowledge unit 》步通大浮 XIan Jiaotong University  Challenges There has been no previous work on mining the relation among the knowledge units a Ontology learning, KAT and rDC can hardly be applied to mine the preorder relations ■ Challenges v Knowledge units expressed in natural language are ambiguous or ill-formed v Knowledge units have far more complex structures than the concepts and named entities v Preorder relations have the characteristic of long distance dependency 面步通大浮 XIan Jiaotong University  Outline e Motivation and Challenges e Two features of the Preorder relation .. Process of Mining Preorder Relations Experimental evaluation ● Conclusions 步交通大浮 XIan Jiaotong