Resources for building applications with Dependency Minimal Recursion Semantics

Resources for Building Applications with Dependency Minimal Recursion Semantics Ann Copestake∗, Guy Emerson∗, Michael Wayne Goodman†, Matic Horvat∗, Alexander Kuhnle∗, Ewa Muszy´nska∗ ∗ University of Cambridge, Cambridge, UK {aac10, gete2, mh693, aok25, emm68}@cam.ac.uk † University of Washington Seattle, WA, USA goodmami@uw.edu Abstract We describe resources aimed at increasing the usability of the semantic representations utilized within the DELPH-IN (Deep Linguistic Processing with HPSG) consortium. We concentrate in particular on the Dependency Minimal Recursion Semantics (DMRS) formalism, a graph-based representation designed for compositional semantic representation with deep grammars. Our main focus is on English, and specifically English Resource Semantics (ERS) as used in the English Resource Grammar. We first give an introduction to ERS and DMRS and a brief overview of some existing resources and then describe in detail a new repository which has been developed to simplify the use of ERS/DMRS. We explain a number of operations on DMRS graphs which our repository supports, with sketches of the algorithms, and illustrate how these operations can be exploited in application building. We believe that this work will aid researchers to exploit the rich and effective but complex DELPH-IN resources. Keywords: English Resource Semantics, Operations on Semantics, Tools for Semantics 1. Introduction DELPH-IN (http://www.delph-in.net) is an in- ternational collaboration between researchers with an in- terest in ‘deep’, linguistically-motivated, language process- ing. Several broad-coverage computational grammars have been produced, together with experimental grammars for languages in all the main language families. These gram- mars share a framework for compositional meaning repre- sentation known as Minimal Recursion Semantics (MRS; Copestake et al. (2005)), which has several variants, col- lectively referred to as *MRS. Currently, the most used and widest coverage DELPH-IN grammar is the English Re- source Grammar (ERG; Flickinger (2000)). *MRS from the ERG (henceforth English Resource Seman- tics: ERS) and other DELPH-IN grammars has been suc- cessfully used in a number of applications. These include machine translation (e.g., Bond et al. (2011)), informa- tion extraction and question answering (e.g., Copestake et al. (2006), Frank et al. (2005), MacKinlay et al. (2009), MacKinlay et al. (2012)), extraction of ontological rela- tionships (e.g., Herbelot and Copestake (2006)), question generation (e.g., Yao and Zhang (2010), Yao et al. (2012)), entailment recognition (e.g., Lien and Kouylekov (2014)), detection of scope of negation (e.g., Packard et al. (2014)) and natural language interfaces (e.g., Packard (2014)). In several cases (including Packard et al. (2014), Lien and Kouylekov (2014), Packard (2014), Yao and Zhang (2010), MacKinlay et al. (2009)) an ERS-based system has outper- formed other approaches in a competitive task. While an ERG-based system is not suitable for applications which require rapid analysis of very large quantities of text, pars- ing accuracy is competitive with other approaches and is generally good on new domains without any need for sub- stantial training or adaptation (MacKinlay et al. (2010), Dridan and Oepen (2013)). ERS is especially suitable in applications which require high precision in semantic tasks and where some lack of recall due to parse failure rates is tolerable. Furthermore, …

Content truncated to 500 words