Implicit Semantic Roles in a Multilingual Setting

Proceedings of the Fifth Joint Conference on Lexical and Computational Semantics (*SEM 2016), pages 45–54, Berlin, Germany, August 11-12, 2016. Implicit Semantic Roles in a Multilingual Setting Jennifer Sikos Yannick Versley Anette Frank Department of Computational Linguistics Heidelberg University, Germany Leibniz Science Campus “Empirical Linguistics and Computational Language Modeling” {sikos, versley, frank}@cl.uni-heidelberg.de Abstract Extending semantic role labeling (SRL) to detect and recover non-local arguments continues to be a challenge. Our work is the first to address the detection of im- plicit roles from a multilingual perspec- tive. We map predicate-argument struc- tures across English and German sen- tences, and we develop a classifier that dis- tinguishes implicit arguments from other translation shifts. Using a combination of alignment statistics and linguistic fea- tures, we achieve a precision of 0.68 de- spite a limited training set, which is a sig- nificant gain over the majority baseline. Our approach does not rely on pre-existing knowledge bases and is extendible to any language pair with parallel data and de- pendency parses. 1 Introduction Understanding events and their participants is a core NLP task, and SRL is the standard approach for identification and labeling of these events in text. SRL systems (T¨ackstr¨om et al., 2015; Roth and Woodsend, 2014) have benefited NLP appli- cations, and many approaches have been proposed to transfer semantic roles from English to other languages without further reliance on manual an- notation (Kozhevnikov and Titov, 2013; Pad´o and Lapata, 2009). However, event structures – both predicates and their arguments – are known to shift in the translation process, and this poor correspon- dence presents a bottleneck for the transference of semantic roles across languages. In some cases, the semantic content of an entire argument can be missing from the scope of its translated predicate. Arguments that are omitted are often treated as noise in state-of-the-art projection models; how- ever, our work views them as a valuable source of data - such arguments serve as naturally occurring training data for implicit role detection. We tar- get arguments that have been dislocated from their predicates, or are dropped entirely, in translated sentences. These non-isomorphic event structures can not only be leveraged as new training data for implicit role detection, but analyzing the shifts that trigger these implicit roles can guide improve- ments to systems that perform cross-lingual se- mantic role projection. Implicit Roles If a predicate is known to have multiple semantic arguments, only a subset might be expressed within the local boundary of its clause or sentence. SRL models typically restrict their search for semantic arguments to this local domain and are not designed to recover arguments situated in the broader discourse context. Non- local role linking extends the SRL task by recov- ering the semantic arguments not instantiated in the local scope of the predicate. One complicating factor is that these implicit arguments can either be found in the context, and thereby are recover- able, or they could be existentially interpreted and might not correspond to any referent in the text …

Content truncated to 500 words