A Unified Framework for Scope Learning via Simplified Shallow Semantic Parsing

Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, pages 714–724, MIT, Massachusetts, USA, 9-11 October 2010. c⃝2010 Association for Computational Linguistics A Unified Framework for Scope Learning via Simplified Shallow Seman- tic Parsing Qiaoming Zhu Junhui Li Hongling Wang Guodong Zhou∗ School of Computer Science and Technology Soochow University, Suzhou, China 215006 {qmzhu, lijunhui, hlwang, gdzhou}@suda.edu.cn ∗ Corresponding author Abstract This paper approaches the scope learning problem via simplified shallow semantic pars- ing. This is done by regarding the cue as the predicate and mapping its scope into several constituents as the arguments of the cue. Evaluation on the BioScope corpus shows that the structural information plays a critical role in capturing the relationship between a cue and its dominated arguments. It also shows that our parsing approach significantly outper- forms the state-of-the-art chunking ones. Al- though our parsing approach is only evaluated on negation and speculation scope learning here, it is portable to other kinds of scope learning. 1 Introduction Recent years have witnessed an increasing interest in the analysis of linguistic scope in natural lan- guage. The task of scope learning deals with the syntactic analysis of what part of a given sentence is under user’s special interest. For example, of negation assertion concerned, a negation cue (e.g., not, no) usually dominates a fragment of the given sentence, rather than the whole sentence, especially when the sentence is long. Generally, scope learn- ing involves two subtasks: cue recognition and its scope identification. The former decides whether a word or phrase in a sentence is a cue of a special interest, where the semantic information of the word or phrase, rather than the syntactic informa- tion, plays a critical role. The latter determines the sequences of words in the sentence which are dominated by the given cue. Recognizing the scope of a special interest (e.g., negative assertion and speculative assertion) is es- sential in information extraction (IE), whose aim is to derive factual knowledge from free text. For example, Vincze et al. (2008) pointed out that the extracted information within the scope of a nega- tion or speculation cue should either be discarded or presented separately from factual information. This is especially important in the biomedical and scientific domains, where various linguistic forms are used extensively to express impressions, hy- pothesized explanations of experimental results or negative findings. Besides, Vincez et al. (2008) reported that 13.45% and 17.70% of the sentences in the abstracts subcorpus of the BioScope corpus contain negative and speculative assertions, respec- tively, while 12.70% and 19.44% of the sentences in the full papers subcorpus contain negative and speculative assertions, respectively. In addition to the IE tasks in the biomedical domain, negation scope learning has attracted increasing attention in some natural language processing (NLP) tasks, such as sentiment classification (Turney, 2002). For example, in the sentence “The chair is not comfortable but cheap”, although both the polari- ties of the words “comfortable” and “cheap” are positive, the polarity of “the chair” regarding …

Content truncated to 500 words