The ACoLi Dictionary Graph

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 3281–3290 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 3281 The ACoLi Dictionary Graph Christian Chiarcos, Christian F¨ath, Maxim Ionov Applied Computational Linguistics Lab (ACoLi) Goethe University Frankfurt, Germany {chiarcos, faeth, ionov}@em.uni-frankfurt.de Abstract In this paper, we report the release of the ACoLi Dictionary Graph, a large-scale collection of multilingual open source dictionaries available in two machine-readable formats: a graph representation in RDF, using the OntoLex-Lemon vocabulary, and a simple tabular data format to facilitate their use in NLP tasks, such as translation inference across dictionaries. We describe the mapping and harmonization of the underlying data structures into a unified representation, its serialization in RDF and TSV, and the release of a massive and coherent amount of lexical data under open licenses. Keywords: lexical resources, OntoLex-Lemon, RDF 1. Motivation Since the confusion of tongues, multilinguality has been both a challenge for and a treasure of mankind, with a number of challenges in the modern age, in particular, the technical challenge to overcome language barriers, and the socio-cultural challenge to preserve linguistic and cultural identities under the pressure of globalization. For both as- pects, the publication of lexical data in machine-readable forms is of utmost importance, as this enables technical so- lutions to provide services across language barriers, in par- ticular for speakers of languages that are underresourced with respect to NLP tools and language resources. A challenge for the vast amount of lexical information available in digital form, is, however, that it comes in var- ious forms. Several widely-used standards for the repre- sentation of lexical information in interoperable form do exist for a long time, e.g., in the context of the Text En- coding Initiative (Burnard and Sperberg-McQueen, 2012), the Lexical Markup Framework (Francopoulo et al., 2006) or tool-specific formats such as the StarDict format,1 but they are limited with respect to their interoperability with each other,2 and – more importantly – they are focus- ing on the standardization of dictionaries as independent, machine-readable entities, whereas many technical appli- cations require an additional focus on the capability of in- tegrating information across different lexical-conceptional resources. Both problems have been driving the more re- cent development of the OntoLex-Lemon vocabulary (Mc- Crae et al., 2017) into the most important data model for the publication of lexical resources on the web of data, and the publication of lexical-conceptual data as Linguistic Linked (Open) Data. OntoLex-Lemon provides a coherent machine-readable representation that is designed to facili- tate its integration across dictionaries and with sources of conceptual knowledge such as ontologies and knowledge bases. With this paper, we announce the release of a massive 1http://stardict.sourceforge.net/ 2Aside from using different formats, the current lack of inter- operability between existing representation formalisms is because of conceptual differences, e.g., the varying level of detail they pro- vide in a machine-readable fashion. collection of multilingual dictionaries in RDF, using the OntoLex-Lemon vocabulary. To facilitate the usage of this data …

Content truncated to 500 words