Compositional Semantic Parsing across Graphbanks

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4576–4585 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 4576 Compositional Semantic Parsing Across Graphbanks Matthias Lindemann∗ and Jonas Groschwitz∗ and Alexander Koller Department of Language Science and Technology Saarland University {mlinde|jonasg|koller}@coli.uni-saarland.de Abstract Most semantic parsers that map sentences to graph-based meaning representations are hand- designed for specific graphbanks. We present a compositional neural semantic parser which achieves, for the first time, competitive accura- cies across a diverse range of graphbanks. In- corporating BERT embeddings and multi-task learning improves the accuracy further, setting new states of the art on DM, PAS, PSD, AMR 2015 and EDS. 1 Introduction Over the past few years, a wide variety of semantic graphbanks have become available. Although these corpora all pair natural-language sentences with graph-based semantic representations, they differ greatly in the design of these graphs (Kuhlmann and Oepen, 2016). Some, in particular the DM, PAS, and PSD corpora of the SemEval shared task on Semantic Dependency Parsing (Oepen et al., 2015), use the tokens of the sentence as nodes and connect them with semantic relations. By contrast, the AMRBank (Banarescu et al., 2013) represents the meaning of each word using a nontrivial con- cept graph; the EDS graphbank (Flickinger et al., 2017) encodes MRS representations (Copestake et al., 2005) as graphs with a many-to-many re- lation between tokens and nodes. In EDS, graph nodes are explicitly aligned with the tokens; in AMR, the alignments are implicit. The graphbanks also exhibit structural differences in their modeling of e.g. coordination or copula. Because of these differences in annotation schemes, the best performing semantic parsers are typically designed for one or very few specific graphbanks. For instance, the currently best sys- tem for DM, PAS, and PSD (Dozat and Manning, ∗Equal contribution 2018) assumes dependency graphs and cannot be directly applied to EDS or AMR. Conversely, top AMR parsers (Lyu and Titov, 2018) invest heavily into identifying AMR-specific alignments and con- cepts, which may not be useful in other graphbanks. Hershcovich et al. (2018) parse across different se- mantic graphbanks (UCCA, DM, AMR), but focus on UCCA and do poorly on DM. The system of Buys and Blunsom (2017) set a state of the art on EDS at the time, but does poorly on AMR. In this paper, we present a single semantic parser that does very well across all of DM, PAS, PSD, EDS and AMR (2015 and 2017). Our system is based on the compositional neural AMR parser of Groschwitz et al. (2018), which represents each graph with its compositional tree structure and learns to predict it through neural dependency pars- ing and supertagging. We show how to heuristi- cally compute the latent compositional structures of the graphs of DM, PAS, PSD, and EDS. This base parser already performs near the state of the art across all six graphbanks. We improve it fur- ther by using pretrained BERT embeddings (Devlin et al., 2019) and multi-task learning. …

Content truncated to 500 words