Negation and Speculation Identification in Chinese Language

Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 656–665, Beijing, China, July 26-31, 2015. c⃝2015 Association for Computational Linguistics Negation and Speculation Identification in Chinese Language 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 Identifying negative or speculative narra- tive fragments from fact is crucial for natural language processing (NLP) appli- cations. Previous studies on negation and speculation identification in Chinese lan- guage suffers much from two problems: corpus scarcity and the bottleneck in fun- damental Chinese information processing. To resolve these problems, this paper constructs a Chinese corpus which con- sists of three sub-corpora from different resources. In order to detect the negative and speculative cues, a sequence labeling model is proposed. Moreover, a bilingual cue expansion method is proposed to in- crease the coverage in cue detection. In addition, this paper presents a new syn- tactic structure-based framework to iden- tify the linguistic scope of a cue, instead of the traditional chunking-based frame- work. Experimental results justify the usefulness of our Chinese corpus and the appropriateness of our syntactic struc- ture-based framework which obtained significant improvement over the state- of-the-art on negation and speculation identification in Chinese language. * 1 Introduction Negation and speculation are ubiquitous phe- nomena in natural language. While negation is a grammatical category which comprises various kinds of devices to reverse the truth value of a proposition, speculation is a grammatical catego- ry which expresses the attitude of a speaker to- wards a statement in terms of degree of certainty, * Corresponding author reliability, subjectivity, sources of information, and perspective (Morante and Sporleder, 2012). Current studies on negation and speculation identification mainly focus on two tasks: 1) cue detection, which aims to detect the signal of a negative or speculative expression, and 2) scope resolution, which aims to determine the linguistic coverage of a cue in sentence, in distinguishing unreliable or uncertain information from facts. For example, (E1) and (E2) include a negative cue and a speculative cue respectively, both de- noted in boldface with their linguistic scopes denoted in square brackets (adopted hereinafter). In sentence (E1), the negative cue “不(not)” trig- gers the scope of “不会追究酒店的这次管理失 职(would not investigate the dereliction of ho- tel)”, within which the fragment “investigate the dereliction of hotel” is the part that is repudiated; While the speculative cue “有望(expected)” in sentence (E2) triggers the scope “后期仍有望反 弹(is still expected to rebound in the late)”, with- in which the fragment “the benchmark Shanghai Composite Index will rebound in the late” is the speculative part. (E1) 所有住客均表示[不会追究酒店的这次管 理失职]. (All of guests said that they [would not in- vestigate the dereliction of hotel].) (E2) 尽管上周五沪指盘中还受创业板的下跌 所拖累,但[后期仍有望反弹]. (Although dragged down by GEM last Fri- day, the benchmark Shanghai Composite In- dex [is still expected to rebound in the late].) Negation and speculation identification is very relevant for almost all NLP …

Content truncated to 500 words