Incorporating EDS Graph for AMR Parsing

Proceedings of the 10th Conference on Lexical and Computational Semantics, pages 202–211 August 5–6, 2021, Bangkok, Thailand (online) ©2021 Association for Computational Linguistics 202 Incorporating EDS Graph for AMR Parsing Ziyi SHOU and Fangzhen LIN Department of Computer Science The Hong Kong University of Science and Technology {zshou,flin}@cse.ust.hk Abstract AMR (Abstract Meaning Representation) and EDS (Elementary Dependency Structures) are two popular meaning representations in NLP/NLU. AMR is more abstract and concep- tual, while EDS is more low level, closer to the lexical structures of the given sentences. It is thus not surprising that EDS parsing is eas- ier than AMR parsing. In this work, we con- sider using information from EDS parsing to help improve the performance of AMR pars- ing. We adopt a transition-based parser and propose to add EDS graphs as additional se- mantic features using a graph encoder com- posed of LSTM layer and GCN layer. Our experimental results show that the additional information from EDS parsing indeed gives a boost to the performance of the base AMR parser used in our experiments. 1 Introduction Semantic parsing has long been considered a diffi- cult task and an important step to natural language understanding. A number of meaning representa- tion formalisms have been proposed. Well-known ones include EDS (Elementary Dependency Struc- tures; Oepen and Lønning, 2006), UCCA (Univer- sal Conceptual Cognitive Annotation; Abend and Rappoport, 2013), and AMR (Abstract Meaning Representation; Banarescu et al., 2013). Among them, AMR is more abstract from surface tokens and tries to capture the meaning of a sentence us- ing concepts that may not appear in the sentence. If one views an AMR encoding as a graph, the AMR graph is always composed of fewer nodes than other meaning representations and some nodes in the AMR graph cannot be anchored to tokens or strings of tokens in the sentence. But EDS tries to build a meaning representation using lexical terms that are presented in the sentence, and nodes in their parse trees are anchored. In comparison, AMR has a much more fine-grained classification for the named entities, total of 124 entity types (Lin and Xue, 2019). Thus not surprisingly, AMR parsers do not perform as well as the ones for EDS. Currently the parsing accuracies for AMR are in low 80s, while they can be high 90s for EDS. In this paper, we propose to use EDS improve the performance of the AMR parser. Figure 1: AMR and EDS graph for ”Imports were at $50.38 billion, up 19%.”, #20011008 sentence from the WSJ Corpus, Penn Treebank (Marcus et al., 1993). Take node #3 in AMR as an example. ”percentage- entity” is the node label, ”value” is the property of this node, and ”19” is the specific value. For node #10 in EDS, ”<35:37>” indicates the span of the correspond- ing surface string; ”card” is the node label, ”CARG” which means ”constant argument” is the property, and ”19” is the value. To see how information from EDS parsing can be of use to AMR, …

Content truncated to 500 words