Exploiting Constituent Dependencies for Tree Kernel-Based Semantic Relation Extraction

Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 697–704 Manchester, August 2008 Exploiting Constituent Dependencies for Tree Kernel-based Semantic Relation Extraction Longhua Qian Guodong Zhou Fang Kong Qiaoming Zhu Peide Qian Jiangsu Provincial Key Lab for Computer Information Processing Technology School of Computer Science and Technology, Soochow University 1 Shizi Street, Suzhou, China 215006 {qianlonghua,gdzhou,kongfang,qmzhu,pdqian}@suda.edu.cn Abstract This paper proposes a new approach to dynamically determine the tree span for tree kernel-based semantic relation ex- traction. It exploits constituent dependen- cies to keep the nodes and their head children along the path connecting the two entities, while removing the noisy in- formation from the syntactic parse tree, eventually leading to a dynamic syntactic parse tree. This paper also explores entity features and their combined features in a unified parse and semantic tree, which in- tegrates both structured syntactic parse information and entity-related semantic information. Evaluation on the ACE RDC 2004 corpus shows that our dy- namic syntactic parse tree outperforms all previous tree spans, and the composite kernel combining this tree kernel with a linear state-of-the-art feature-based ker- nel, achieves the so far best performance. 1 Introduction Information extraction is one of the key tasks in natural language processing. It attempts to iden- tify relevant information from a large amount of natural language text documents. Of three sub- tasks defined by the ACE program1, this paper focuses exclusively on Relation Detection and Characterization (RDC) task, which detects and classifies semantic relationships between prede- fined types of entities in the ACE corpus. For © 2008. Licensed under the Creative Commons Attribution- Noncommercial-Share Alike 3.0 Unported (http://creativecommons.org/licenses/by-nc-sa/3.0/). license Some rights reserved. 1 http://www.ldc.upenn.edu/Projects/ACE/ example, the sentence “Microsoft Corp. is based in Redmond, WA” conveys the relation “GPE- AFF.Based” between “Microsoft Corp.” [ORG] and “Redmond” [GPE]. Due to limited accuracy in state-of-the-art syntactic and semantic parsing, reliably extracting semantic relationships be- tween named entities in natural language docu- ments is still a difficult, unresolved problem. In the literature, feature-based methods have dominated the research in semantic relation ex- traction. Featured-based methods achieve prom- ising performance and competitive efficiency by transforming a relation example into a set of syn- tactic and semantic features, such as lexical knowledge, entity-related information, syntactic parse trees and deep semantic information. How- ever, detailed research (Zhou et al., 2005) shows that it’s difficult to extract new effective features to further improve the extraction accuracy. Therefore, researchers turn to kernel-based methods, which avoids the burden of feature en- gineering through computing the similarity of two discrete objects (e.g. parse trees) directly. From prior work (Zelenko et al., 2003; Culotta and Sorensen, 2004; Bunescu and Mooney, 2005) to current research (Zhang et al., 2006; Zhou et al., 2007), kernel methods have been showing more and more potential in relation extraction. The key problem for kernel methods on rela- tion extraction is how to represent and capture the structured syntactic information inherent in relation instances. While kernel methods using the dependency tree (Culotta and Sorensen, 2004) and the shortest …

Content truncated to 500 words