Broad-Coverage Semantic Parsing as Transduction

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 3786–3798, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 3786 Broad-Coverage Semantic Parsing as Transduction Sheng Zhang Xutai Ma Kevin Duh Johns Hopkins UniversityBenjamin Van Durme {zsheng2, xutai ma}@jhu.edu {kevinduh, vandurme}@cs.jhu.edu Abstract We unify different broad-coverage semantic parsing tasks under a transduction paradigm, and propose an attention-based neural frame- work that incrementally builds a meaning rep- resentation via a sequence of semantic rela- tions. By leveraging multiple attention mecha- nisms, the transducer can be effectively trained without relying on a pre-trained aligner. Ex- periments conducted on three separate broad- coverage semantic parsing tasks – AMR, SDP and UCCA – demonstrate that our attention- based neural transducer improves the state of the art on both AMR and UCCA, and is com- petitive with the state of the art on SDP. 1 Introduction Broad-coverage semantic parsing aims at map- ping any natural language text, regardless of its domain, genre, or even the language itself, into a general-purpose meaning representation. As a long-standing topic of interest in computational linguistics, broad-coverage semantic parsing has targeted a number of meaning representation frameworks, including CCG (Steedman, 1996, 2001), DRS (Kamp and Reyle, 1993; Bos, 2008), AMR (Banarescu et al., 2013), UCCA (Abend and Rappoport, 2013), SDP (Oepen et al., 2014, 2015), and UDS (White et al., 2016).1 Each of these frameworks has their specific formal and linguistic assumptions. Such framework-specific “balkanization” results in a variety of framework- specific parsing approaches, and the state-of-the- art semantic parser for one framework is not al- ways applicable to another. For instance, the state- of-the-art approaches to SDP parsing (Dozat and 1Abbreviations respectively denote: Combinatory Cat- egorical Grammar, Discourse Representation Theory, Ab- stract Meaning Representation, Universal Conceptual Cog- nitive Annotation, Semantic Dependency Parsing, and Uni- versal Decompositional Semantics. Manning, 2018; Peng et al., 2017a) are not directly transferable to AMR and UCCA because of the lack of explicit alignments between tokens in the sentence and nodes in the semantic graph. While transition-based approaches are adapt- able to different broad-coverage semantic parsing tasks (Wang et al., 2018; Hershcovich et al., 2018; Damonte et al., 2017), when it comes to represen- tations such as AMR whose nodes are unanchored to tokens in the sentence, a pre-trained aligner has to be used to produce the reference transi- tion sequences (Wang et al., 2015; Damonte et al., 2017; Peng et al., 2017b). In contrast, there are attempts to develop attention-based approaches in a graph-based parsing paradigm (Dozat and Man- ning, 2018; Zhang et al., 2019), but they lack pars- ing incrementality, which is advocated in terms of computational efficiency and cognitive model- ing (Nivre, 2004; Huang and Sagae, 2010). In this paper, we approach different broad- coverage semantic parsing tasks under a uni- fied framework of transduction. We propose an attention-based neural transducer that extends the two-stage semantic parser of Zhang et al. (2019) to directly transduce input …

Content truncated to 500 words