SemEval-2016 Task 9: Chinese Semantic Dependency Parsing

Proceedings of SemEval-2016, pages 1074–1080, San Diego, California, June 16-17, 2016. c⃝2016 Association for Computational Linguistics SemEval-2016 Task 9: Chinese Semantic Dependency Parsing Wanxiang Che† Yanqiu Shao‡ Ting Liu† Yu Ding† †{car, tliu, yding}@ir.hit.edu.cn School of Computer Science and Technology Harbin Institute of Technology Harbin, China, 150001 ‡yqshao@blcu.edu.cn School of Information Science Beijing Language and Culture University Beijing, China, 100083 Abstract This paper describes the SemEval-2016 Shared Task 9: Chinese semantic Depen- dency Parsing. We extend the traditional tree- structured representation of Chinese sentence to directed acyclic graphs that can capture richer latent semantics, and the goal of this task is to identify such semantic structures from a corpus of Chinese sentences. We pro- vide two distinguished corpora in the NEWS domain with 10,068 sentences and the TEXT- BOOKS domain with 14,793 sentences re- spectively. We will first introduce the moti- vation for this task, and then present the task in detail including data preparation, data for- mat, task evaluation and so on. At last, we briefly describe the submitted systems and an- alyze these results. 1 Introduction This task is a rerun of the task 5 at SemEval 2012 (Che et al., 2012), named Chinese seman- tic dependency parsing (SDP). In the previous task, we aimed at investigating “deep” semantic relations within sentences through tree-structured dependen- cies. As traditionally defined, syntactic dependency parsing results are connected trees defined over all words of a sentence and language-specific gram- matical functions. On the contrary, in semantic de- pendency parsing, each head-dependent arc instead bears a semantic relation, rather than grammatical relation. In this way, semantic dependency parsing results can be used to answer questions directly, like who did what to whom when and where. However, according to the meaning-text linguis- tic theory (Mel’ˇcuk and ˇZolkovskij, 1965), a theo- retical framework for the description of natural lan- guages, it is said that trees are not sufficient to ex- press the complete meaning of sentences in some cases, which has been proven undoubted in our prac- tice of corpus annotation. This time, not only do we refine the easy-to-understand meaning represen- tation in Chinese in order to decrease ambiguity or fuzzy boundary of semantic relations on the basis of Chinese linguistic knowledge, we extend the depen- dency structure to directed acyclic graphs that con- form to the characteristics of Chinese, because Chi- nese is an parataxis language with flexible word or- ders, and rich latent information is hidden in facial words. Figure 1 illustrates an example of semantic de- pendency graph. Here, “她 (she)” is the argument of “脸色 (face)” and at the same time it is an argu- ment of “病 (disease)”. Researchers in dependency parsing community realized dependency parsing re- stricted in a tree structure is still too shallow, so they explored semantic information beyond tree structure in task 8 at SemEval 2014 (Oepen et al., 2014) and task 18 at SemEval 2015 (Oepen et al., 2015). They provided data in similar structure with what we are going to provide, but …

Content truncated to 500 words