A Distant Supervision Approach to Semantic Role Labeling

Proceedings of the Fourth Joint Conference on Lexical and Computational Semantics (*SEM 2015), pages 239–248, Denver, Colorado, June 4–5, 2015. A Distant Supervision Approach to Semantic Role Labeling Peter Exner Marcus Klang Lund University Department of Computer Science Lund, Sweden {Peter.Exner, Marcus.Klang, Pierre.Nugues}@cs.lth.se Pierre Nugues Abstract Semantic role labeling has become a key mod- ule for many language processing applica- tions such as question answering, information extraction, sentiment analysis, and machine translation. To build an unrestricted semantic role labeler, the first step is to develop a com- prehensive proposition bank. However, creat- ing such a bank is a costly enterprise, which has only been achieved for a handful of lan- guages. In this paper, we describe a technique to build proposition banks for new languages using distant supervision. Starting from PropBank in English and loosely parallel corpora such as versions of Wikipedia in different languages, we carried out a mapping of semantic propo- sitions we extracted from English to syntactic structures in Swedish using named entities. We trained a semantic parser on the generated Swedish propositions and we report the results we obtained. Using the CoNLL 2009 evalua- tion script, we could reach the scores of 52.25 for labeled propositions and 62.44 for the un- labeled ones. We believe our approach can be applied to train semantic role labelers for other resource-scarce languages. 1 Introduction Semantic role labeling has become a key module for many language processing applications and its im- portance is growing in fields like question answer- ing (Shen and Lapata, 2007), information extraction (Christensen et al., 2010), sentiment analysis (Jo- hansson and Moschitti, 2011), and machine trans- lation (Liu and Gildea, 2010; Wu et al., 2011). To build an unrestricted semantic role labeler, the first step is to develop a comprehensive proposition bank. However, building proposition banks is a costly en- terprise and as a consequence of that, they only exist for a handful of languages such as English, Chinese, German, or Spanish. In this paper, we describe a technique to create proposition banks for new languages using distant supervision. Our approach builds on the transfer of semantic information through named entities. Start- ing from an existing proposition bank, PropBank in English (Palmer et al., 2005), and loosely parallel corpora such as versions of Wikipedia in different languages, we carried out a mapping of the semantic propositions we extracted from English to syntactic structures in the target language. We parsed the English edition of Wikipedia up to the predicate–argument structures using a se- mantic role labeler (Bj¨orkelund et al., 2010a) and the Swedish Wikipedia using a dependency parser (Nivre et al., 2006). We extracted all the named entities we found in the propositions and we dis- ambiguated them using the Wikidata nomenclature1. Using recurring entities, we aligned sentences in the two languages; we transferred the semantic annota- tion from English sentences to Swedish sentences; and we could identify 2,333 predicate–argument frames in Swedish. Finally, we used the resulting corpus to train a semantic role labeler for Swedish …

Content truncated to 500 words