A Clustering Approach for Nearly Unsupervised Recognition of Nonliteral Language

A Clustering Approach for the Nearly Unsupervised Recognition of Nonliteral Language∗ Julia Birke and Anoop Sarkar School of Computing Science, Simon Fraser University Burnaby, BC, V5A 1S6, Canada jbirke@alumni.sfu.ca, anoop@cs.sfu.ca Abstract In this paper we present TroFi (Trope Finder), a system for automatically classi- fying literal and nonliteral usages of verbs through nearly unsupervised word-sense disambiguation and clustering techniques. TroFi uses sentential context instead of selectional constraint violations or paths in semantic hierarchies. It also uses lit- eral and nonliteral seed sets acquired and cleaned without human supervision in or- der to bootstrap learning. We adapt a word-sense disambiguation algorithm to our task and augment it with multiple seed set learners, a voting schema, and addi- tional features like SuperTags and extra- sentential context. Detailed experiments on hand-annotated data show that our en- hanced algorithm outperforms the base- line by 24.4%. Using the TroFi algo- rithm, we also build the TroFi Example Base, an extensible resource of annotated literal/nonliteral examples which is freely available to the NLP research community. 1 Introduction In this paper, we propose TroFi (Trope Finder), a nearly unsupervised clustering method for sep- arating literal and nonliteral usages of verbs. For example, given the target verb “pour”, we would expect TroFi to cluster the sentence “Custom demands that cognac be poured from a freshly opened bottle” as literal, and the sentence “Salsa and rap music pour out of the windows” as nonlit- eral, which, indeed, it does. We call our method nearly unsupervised. See Section 3.1 for why we use this terminology. We reduce the problem of nonliteral language recognition to one of word-sense disambiguation ∗ This research was partially supported by NSERC, Canada (RGPIN: 264905). We would like to thank Bill Dolan, Fred Popowich, Dan Fass, Katja Markert, Yudong Liu, and the anonymous reviewers for their comments. by redefining literal and nonliteral as two differ- ent senses of the same word, and we adapt an ex- isting similarity-based word-sense disambiguation method to the task of separating usages of verbs into literal and nonliteral clusters. This paper fo- cuses on the algorithmic enhancements necessary to facilitate this transformation from word-sense disambiguation to nonliteral language recognition. The output of TroFi is an expandable example base of literal/nonliteral clusters which is freely avail- able to the research community. Many systems that use NLP methods – such as dialogue systems, paraphrasing and summariza- tion, language generation, information extraction, machine translation, etc. – would benefit from be- ing able to recognize nonliteral language. Con- sider an example based on a similar example from an automated medical claims processing system. We must determine that the sentence “she hit the ceiling” is meant literally before it can be marked up as an ACCIDENT claim. Note that the typical use of “hit the ceiling” stored in a list of idioms cannot help us. Only using the context, “She broke her thumb while she was cheering for the Patriots and, in her excitement, she hit the ceiling,” can we decide. We further motivate the usefulness …

Content truncated to 500 words