Classifying Idiomatic and Literal Expressions Using Vector Space Representations

Proceedings of Recent Advances in Natural Language Processing, pages 507–511, Hissar, Bulgaria, Sep 7–9 2015. Classifying Idiomatic and Literal Expressions Using Vector Space Representations Jing Peng Montclair State University pengj@mail.montclair.edu Anna Feldman Montclair State University feldmana@mail.montclair.edu Hamza Jazmati Montclair State University jazmatih1@mail.montclair.edu Abstract We describe an algorithm for automatic classification of idiomatic and literal ex- pressions. Our starting point is that idioms and literal expressions occur in different contexts. Idioms tend to violate cohesive ties in local contexts, while literals are ex- pected to fit in. Our goal is to capture this intuition using a vector representation of words. We propose two approaches: (1) Compute inner product of context word vectors with the vector representing a tar- get expression. Since literal vectors pre- dict well local contexts, their inner prod- uct with contexts should be larger than idiomatic ones, thereby telling apart lit- erals from idioms; and (2) Compute lit- eral and idiomatic scatter (covariance) ma- trices from local contexts in word vec- tor space. Since the scatter matrices rep- resent context distributions, we can then measure the difference between the dis- tributions using the Frobenius norm. We provide experimental results validating the proposed techniques. 1 Introduction Despite the common belief that idioms are always idioms, potentially idiomatic expressions, such as hit the sack can appear in literal contexts. Fazly et al. (2009)’s analysis of 60 idioms from the British National Corpus (BNC) has shown that close to half of these also have a clear literal meaning; and of those with a literal meaning, on average around 40% of their usages are literal. Therefore, idioms present great challenges for many Natural Lan- guage Processing (NLP) applications. Most cur- rent translation systems rely on large repositories of idioms. In this paper we describe an algorithm for automatic classification of idiomatic and literal expressions. Similarly to Peng et al. (2014), we treat idioms as semantic outliers. Our assumption is that the context word distribution for a literal ex- pression will be different from the distribution for an idiomatic one. We capture the distribution in terms of covariance matrix in vector space. 2 Previous Work Previous approaches to idiom detection can be classified into two groups: 1) type-based extrac- tion, i.e., detecting idioms at the type level; 2) token-based detection, i.e., detecting idioms in context. Type-based extraction is based on the idea that idiomatic expressions exhibit certain lin- guistic properties such as non-compositionality that can distinguish them from literal expressions (Sag et al., 2002; Fazly et al., 2009). While many idioms do have these properties, many idioms fall on the continuum from being compositional to being partly unanalyzable to completely non- compositional (Cook et al., 2007). Katz and Gies- brech (2006), Birke and Sarkar (2006), Fazly et al. (2009), Sporleder and Li (2009), Li and Sporleder (2010), among others, notice that type-based ap- proaches do not work on expressions that can be interpreted idiomatically or literally depending on the context and thus, an approach that considers tokens in context is more appropriate for …

Content truncated to 500 words