UGroningen: Negation detection with Discourse Representation Structures

First Joint Conference on Lexical and Computational Semantics (*SEM), pages 301–309, Montr´eal, Canada, June 7-8, 2012. c⃝2012 Association for Computational Linguistics UGroningen: Negation detection with Discourse Representation Structures Valerio Basile and Johan Bos and Kilian Evang and Noortje Venhuizen {v.basile,johan.bos,k.evang,n.j.venhuizen}@rug.nl Center for Language and Cognition Groningen (CLCG) University of Groningen, The Netherlands Abstract We use the NLP toolchain that is used to con- struct the Groningen Meaning Bank to address the task of detecting negation cue and scope, as defined in the shared task “Resolving the Scope and Focus of Negation”. This toolchain applies the C&C tools for parsing, using the formalism of Combinatory Categorial Gram- mar, and applies Boxer to produce seman- tic representations in the form of Discourse Representation Structures (DRSs). For nega- tion cue detection, the DRSs are converted to flat, non-recursive structures, called Dis- course Representation Graphs (DRGs). DRGs simplify cue detection by means of edge la- bels representing relations. Scope detection is done by gathering the tokens that occur within the scope of a negated DRS. The re- sult is a system that is fairly reliable for cue detection and scope detection. Furthermore, it provides a fairly robust algorithm for detect- ing the negated event or property within the scope. 1 Introduction Nothing is more home to semantics than the phe- nomenon of negation. In classical theories of mean- ing all states of affairs are divided in two truth val- ues, and negation plays a central role to determine which truth value is at stake for a given sentence. Negation lies at the heart of deductive inference, of which consistency checking (searching for contra- dictions in texts) is a prime example in natural lan- guage understanding. It shouldn’t therefore come as a surprise that detecting negation and adequately representing its scope is of utmost importance in computational se- mantics. In this paper we present and evaluate a sys- tem that transforms texts into logical formulas – us- ing the C&C tools and Boxer (Bos, 2008) – in the context of the shared task on recognising negation in English texts (Morante and Blanco, 2012). We will first sketch the background and the basics of the formalism that we employ in our analysis of negation (Section 2). In Section 3 we explain how we detect negation cues and scope. Finally, in Sec- tion 4 we present the results obtained in the shared task, and we discuss them in Section 5. 2 Background The semantic representations that are used in this shared task on detecting negation in texts are con- structed by means of a pipeline of natural language processing components, of which the backbone is provided by the C&C tools and Boxer (Curran et al., 2007). This tool chain is currently in use semi- automatically for constructing a large semantically annotated corpus, the Groningen Meaning Bank (Basile et al., 2012). The C&C tools are applied for tagging the data with part-of-speech and super tags and for syntactic parsing, using the formalism of Combinatory Cate- gorial Grammar, CCG …

Content truncated to 500 words