Improving Nominal SRL in Chinese Language with Verbal SRL Information and Automatic Predicate Recognition

Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, pages 1280–1288, Singapore, 6-7 August 2009. c⃝2009 ACL and AFNLP Improving Nominal SRL in Chinese Language with Verbal SRL In- formation and Automatic Predicate Recognition Junhui Li† Guodong Zhou†∗ Hai Zhao†‡ Qiaoming Zhu† Peide Qian† † Jiangsu Provincial Key Lab for Computer Information Processing Technologies School of Computer Science and Technology Soochow University, Suzhou, China 215006 ‡ Department of Chinese, Translation and Linguistics City University of HongKong, China Email: {lijunhui,gdzhou,hzhao,qmzhu,pdqian}@suda.edu.cn ∗ Corresponding author Abstract This paper explores Chinese semantic role la- beling (SRL) for nominal predicates. Besides those widely used features in verbal SRL, various nominal SRL-specific features are first included. Then, we improve the perform- ance of nominal SRL by integrating useful features derived from a state-of-the-art verbal SRL system. Finally, we address the issue of automatic predicate recognition, which is es- sential for a nominal SRL system. Evaluation on Chinese NomBank shows that our research in integrating various features derived from verbal SRL significantly improves the per- formance. It also shows that our nominal SRL system much outperforms the state-of-the-art ones. 1. Introduction Semantic parsing maps a natural language sen- tence into a formal representation of its meaning. Due to the difficulty in deep semantic parsing, most of previous work focuses on shallow se- mantic parsing, which assigns a simple structure (such as WHO did WHAT to WHOM, WHEN, WHERE, WHY, HOW) to each predicate in a sentence. In particular, the well-defined seman- tic role labeling (SRL) task has been drawing more and more attention in recent years due to its importance in deep NLP applications, such as question answering (Narayanan and Harabagiu, 2004), information extraction (Surdeanu et al., 2003), and co-reference resolution (Ponzetto and Strube, 2006). Given a sentence and a predicate (either a verb or a noun) in it, SRL recognizes and maps all the constituents in the sentence into their corresponding semantic arguments (roles) of the predicate. According to the predicate types, SRL could be divided into SRL for verbal predicates (verbal SRL, in short) and SRL for nominal predicates (nominal SRL, in short). During the past few years, verbal SRL has dominated the research on SRL with the avail- ability of FrameNet (Baker et al., 1998), Prop- Bank (Palmer et al., 2005), and the consecutive CoNLL shared tasks (Carreras and Màrquez, 2004 & 2005) in English language. As a com- plement to PropBank on verbal predicates, NomBank (Meyers et al., 2004) annotates nomi- nal predicates and their corresponding semantic roles using similar semantic framework as PropBank. As a representative, Jiang and Ng (2006) pioneered the exploration of various nominal SRL-specific features besides the tradi- tional verbal SRL-related features on NomBank. They achieved the performance of 72.73 and 69.14 in F1-measure on golden and automatic syntactic parse trees, respectively, given golden nominal predicates. For SRL in Chinese, Sun and Jurafsky (2004) and Pradhan et al. (2004) pioneered the research on Chinese verbal and nominal SRLs, respec- tively, on small private datasets. Taking the …

Content truncated to 500 words