Negation Focus Identification with Contextual Discourse Information

Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, pages 522–530, Baltimore, Maryland, USA, June 23-25 2014. c⃝2014 Association for Computational Linguistics Negation Focus Identification with Contextual Discourse Information Bowei Zou Qiaoming Zhu Guodong Zhou* Natural Language Processing Lab, School of Computer Science and Technology Soochow University, Suzhou, 215006, China zoubowei@gmail.com, {qmzhu, gdzhou}@suda.edu.cn Abstract Negative expressions are common in natural language text and play a critical role in in- formation extraction. However, the perfor- mances of current systems are far from satis- faction, largely due to its focus on intra- sentence information and its failure to con- sider inter-sentence information. In this paper, we propose a graph model to enrich intra- sentence features with inter-sentence features from both lexical and topic perspectives. Evaluation on the *SEM 2012 shared task corpus indicates the usefulness of contextual discourse information in negation focus iden- tification and justifies the effectiveness of our graph model in capturing such global infor- mation. * 1 Introduction Negation is a grammatical category which com- prises various kinds of devices to reverse the truth value of a proposition (Morante and Sporleder, 2012). For example, sentence (1) could be interpreted as it is not the case that he stopped. (1) He didn't stop. Negation expressions are common in natural language text. According to the statistics on bio- medical literature genre (Vincze et al., 2008), 19.44% of sentences contain negative expres- sions. The percentage rises to 22.5% on Conan Doyle stories (Morante and Daelemans, 2012). It is interesting that a negative sentence may have both negative and positive meanings. For exam- ple, sentence (2) could be interpreted as He stopped, but not until he got to Jackson Hole with positive part he stopped and negative part until he got to Jackson Hole. Moreover, a nega- * Corresponding author tive expression normally interacts with some special part in the sentence, referred as negation focus in linguistics. Formally, negation focus is defined as the special part in the sentence, which is most prominently or explicitly negated by a negative expression. Hereafter, we denote nega- tive expression in boldface and negation focus underlined. (2) He didn't stop until he got to Jackson Hole. While people tend to employ stress or intona- tion in speech to emphasize negation focus and thus it is easy to identify negation focus in speech corpora, such stress or intonation infor- mation often misses in the dominating text cor- pora. This poses serious challenges on negation focus identification. Current studies (e.g., Blanco and Moldovan, 2011; Rosenberg and Bergler, 2012) sort to various kinds of intra-sentence in- formation, such as lexical features, syntactic fea- tures, semantic role features and so on, ignoring less-obvious inter-sentence information. This largely defers the performance of negation focus identification and its wide applications, since such contextual discourse information plays a critical role on negation focus identification. Take following sentence as an example. (3) Helen didn’t allow her youngest son to play the violin. In sentence (3), there are several scenarios on identification of …

Content truncated to 500 words