Made for Each Other: Broad-Coverage Semantic Structures Meet Preposition Supersenses

Proceedings of the 23rd Conference on Computational Natural Language Learning, pages 174–185 Hong Kong, China, November 3-4, 2019. c⃝2019 Association for Computational Linguistics 174 Made for Each Other: Broad-coverage Semantic Structures Meet Preposition Supersenses Jakob Prange Nathan Schneider Georgetown University jakob@cs.georgetown.edu nathan.schneider@georgetown.edu Omri Abend The Hebrew University of Jerusalem oabend@cs.huji.ac.il Abstract Universal Conceptual Cognitive Annotation (UCCA; Abend and Rappoport, 2013) is a typologically-informed, broad-coverage se- mantic annotation scheme that describes coarse-grained predicate-argument structure but currently lacks semantic roles. We argue that lexicon-free annotation of the semantic roles marked by prepositions, as formulated by Schneider et al. (2018), is complementary and suitable for integration within UCCA. We show empirically for English that the schemes, though annotated independently, are compati- ble and can be combined in a single semantic graph. A comparison of several approaches to parsing the integrated representation lays the groundwork for future research on this task. 1 Introduction A common thread in many approaches to mean- ing representation is the idea that abstract struc- tures can describe semantic invariants that hold across paraphrasing or translation: for example, semantic dependency relations capturing predicate- argument structures or other types of semantic relations that can be annotated within sentences (e.g., Böhmová et al., 2003; Oepen et al., 2015; Banarescu et al., 2013). These annotation schemes can be distinguished by various design principles such as language-specificity; the level of granu- larity of meaning elements; the reliance on mor- phosyntactic criteria to define the units of semantic annotation; the extent to which human annotators specify semantics from scratch; and many others (Abend and Rappoport, 2017). In this work, we seize an opportunity to unite two previously unrelated—yet complementary— meaning representations in NLP. On the one hand, Universal Conceptual Cognitive Annotation (UCCA; Abend and Rappoport, 2013) provides I went to ohm after reading some of the reviews H A Process A∣GOAL R Center Linker H∣EXPLANATION Process A Q∣QUANTITY R Fxn Center A Figure 1: Semantic parse illustrating the integrated rep- resentation proposed here. Solid edges are the UCCA parse’s primary edges, and the dotted edge is a re- mote edge. Dashed arrows show how SNACS labels (green small caps) have been mapped onto edges of the UCCA structure from the prepositions on which they were originally annotated. The following UCCA cat- egories are abbreviated: A = Participant, R = Relator, H = Parallel scene, Q = Quantifier, Fxn = Function. a skeletal structure of semantic units and rela- tions, with typologically-based criteria for mark- ing predicate-argument structures, based on “Basic Linguistic Theory”, an established framework for typological description (Dixon, 2010/2012). On the other hand, a recent approach to annotation of English prepositions and possessives (SNACS; Schneider et al., 2018) provides an inventory of labels that characterize semantic relations. UCCA and SNACS follow similar design principles: they are both language-neutral, with general-purpose coarse-grained labels rather than lexically-specific senses or roles; and they are both designed for direct semantic annotation, without requiring a syn- tactic parse as a foundation. The philosophy is that these properties will facilitate …

Content truncated to 500 words